From f5bd8657296468624289fb896a0449c75f1579f7 Mon Sep 17 00:00:00 2001 From: hieu-jan <150573299+hieu-jan@users.noreply.github.com> Date: Sat, 9 Mar 2024 07:48:49 +0900 Subject: [PATCH 01/37] docs: update slogan --- docs/docs/about/about.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/docs/about/about.md b/docs/docs/about/about.md index 0f1651c1a..a047ab910 100644 --- a/docs/docs/about/about.md +++ b/docs/docs/about/about.md @@ -18,7 +18,7 @@ keywords: ] --- -Jan turns computers into a thinking machine to change how you use computers. +Jan turns computers into thinking machines to change how we use them. Jan is created and maintained by Jan Labs, a robotics company. With Jan, you can: From 3e045148b15759b8c83f73a0781d16db4818645c Mon Sep 17 00:00:00 2001 From: hieu-jan <150573299+hieu-jan@users.noreply.github.com> Date: Sun, 10 Mar 2024 17:27:58 +0900 Subject: [PATCH 02/37] docs: update website-docs content --- docs/README.md | 60 +++++++------ docs/docs/.gitignore | 1 + .../how-we-work/website-docs/website-docs.md | 87 ++++++++++++++++++- 3 files changed, 119 insertions(+), 29 deletions(-) create mode 100644 docs/docs/.gitignore diff --git a/docs/README.md b/docs/README.md index b5d31702f..2b27b354b 100644 --- a/docs/README.md +++ b/docs/README.md @@ -1,18 +1,18 @@ -# Website +This website is built using [Docusaurus 3.0](https://docusaurus.io/), a modern static website generator. -This website is built using [Docusaurus 2](https://docusaurus.io/), a modern static website generator. - -## Information Architecture +### Information Architecture We try to **keep routes consistent** to maintain SEO. -- `/guides`: Guides on how to use the Jan application, with GIFs. For end users who are directly using Jan. Always assume users are not technical. +- **`/guides/`**: Guides on how to use the Jan application. For end users who are directly using Jan. -- `/developer`: Developer docs on how to extend Jan. These pages are about what people can build with our software. We must hide the complexity of HOW the app is built, but explain just enough of the high level architecture so devs know enough to build on top of it. +- **`/developer/`**: Developer docs on how to extend Jan. These pages are about what people can build with our software. -- `/api-reference`: Reference documentation, written in Swagger/OpenAPI format. +- **`/api-reference/`**: Reference documentation for the Jan API server, written in Swagger/OpenAPI format. -- `/docs`: Engineering specs and product specs, i.e. HOW the app is built. Mostly for internal reference and for our core contributors who are building the SDK itself. +- **`/changelog/`**: A list of changes made to the Jan application with each release. + +- **`/blog/`**: A blog for the Jan application. ### Sidebar Autogeneration @@ -20,34 +20,36 @@ The order of each page is either explicitly defined in `sidebar.js` or follows t Important slugs are hardcoded at the document level (and shouldn't be rerouted): -```md +``` --- title: Overview slug: /docs --- ``` -## Contributing +## How to Contribute -### Installation +Refer to the [Contributing Guide](https://github.com/janhq/jan/blob/dev/CONTRIBUTING.md) for more comprehensive information on how to contribute to the Jan project. -``` -$ yarn -``` +### Pre-requisites and Installation -### Local Development +- [Node.js](https://nodejs.org/en/) (version 20.0.0 or higher) +- [yarn](https://yarnpkg.com/) (version 1.22.0 or higher) -``` -$ cp .env.example .env -$ yarn start +#### Installation + +```bash +cd jan/docs +yarn install +yarn start ``` This command starts a local development server and opens up a browser window. Most changes are reflected live without having to restart the server. -### Build +#### Build -``` -$ yarn build +```bash +yarn build ``` This command generates static content into the `build` directory and can be served using any static contents hosting service. @@ -56,25 +58,27 @@ This command generates static content into the `build` directory and can be serv Using SSH: -``` -$ USE_SSH=true yarn deploy +```bash +USE_SSH=true yarn deploy ``` Not using SSH: -``` -$ GIT_USER= yarn deploy +```bash +GIT_USER= yarn deploy ``` If you are using GitHub pages for hosting, this command is a convenient way to build the website and push to the `gh-pages` branch. ### Preview URL, Pre-release and Publishing Documentation -When a PR is created, the preview URL will be automatically commented on the PR. +- When a pull request is created, the preview URL will be automatically commented on the pull request. -The documentation will then be published to [https://jan.ai/](https://jan.ai/) when the PR is merged to `main`. +- The documentation will then be published to [https://dev.jan.ai/](https://dev.jan.ai/) when the pull request is merged to `dev`. + +- Our open-source maintainers will sync the updated content from `dev` to `docs` branch, which will then be published to [https://jan.ai/](https://jan.ai/). ### Additional Plugins - @docusaurus/theme-live-codeblock -- [Redocusaurus](https://redocusaurus.vercel.app/): manually upload swagger files at `/openapi/OpenAPISpec.json` +- [Redocusaurus](https://redocusaurus.vercel.app/): manually upload swagger files at `/openapi/jan.yaml` to update the API reference documentation. diff --git a/docs/docs/.gitignore b/docs/docs/.gitignore new file mode 100644 index 000000000..ff2a8c5fd --- /dev/null +++ b/docs/docs/.gitignore @@ -0,0 +1 @@ +guides/changelogs \ No newline at end of file diff --git a/docs/docs/how-we-work/website-docs/website-docs.md b/docs/docs/how-we-work/website-docs/website-docs.md index a152c6dc9..19fdc1676 100644 --- a/docs/docs/how-we-work/website-docs/website-docs.md +++ b/docs/docs/how-we-work/website-docs/website-docs.md @@ -1,3 +1,88 @@ --- title: Website & Docs ---- \ No newline at end of file +--- + +This website is built using [Docusaurus 3.0](https://docusaurus.io/), a modern static website generator. + +### Information Architecture + +We try to **keep routes consistent** to maintain SEO. + +- **`/guides/`**: Guides on how to use the Jan application. For end users who are directly using Jan. + +- **`/developer/`**: Developer docs on how to extend Jan. These pages are about what people can build with our software. + +- **`/api-reference/`**: Reference documentation for the Jan API server, written in Swagger/OpenAPI format. + +- **`/changelog/`**: A list of changes made to the Jan application with each release. + +- **`/blog/`**: A blog for the Jan application. + +### Sidebar Autogeneration + +The order of each page is either explicitly defined in `sidebar.js` or follows the [Docusaurus autogenerated](https://docusaurus.io/docs/next/sidebar/autogenerated) naming format, `##-path-name.md`. + +Important slugs are hardcoded at the document level (and shouldn't be rerouted): + +``` +--- +title: Overview +slug: /docs +--- +``` + +## How to Contribute + +Refer to the [Contributing Guide](https://github.com/janhq/jan/blob/dev/CONTRIBUTING.md) for more comprehensive information on how to contribute to the Jan project. + +### Pre-requisites and Installation + +- [Node.js](https://nodejs.org/en/) (version 20.0.0 or higher) +- [yarn](https://yarnpkg.com/) (version 1.22.0 or higher) + +#### Installation + +```bash +cd jan/docs +yarn install +yarn start +``` + +This command starts a local development server and opens up a browser window. Most changes are reflected live without having to restart the server. + +#### Build + +```bash +yarn build +``` + +This command generates static content into the `build` directory and can be served using any static contents hosting service. + +### Deployment + +Using SSH: + +```bash +USE_SSH=true yarn deploy +``` + +Not using SSH: + +```bash +GIT_USER= yarn deploy +``` + +If you are using GitHub pages for hosting, this command is a convenient way to build the website and push to the `gh-pages` branch. + +### Preview URL, Pre-release and Publishing Documentation + +- When a pull request is created, the preview URL will be automatically commented on the pull request. + +- The documentation will then be published to [https://dev.jan.ai/](https://dev.jan.ai/) when the pull request is merged to `dev`. + +- Our open-source maintainers will sync the updated content from `dev` to `docs` branch, which will then be published to [https://jan.ai/](https://jan.ai/). + +### Additional Plugins + +- @docusaurus/theme-live-codeblock +- [Redocusaurus](https://redocusaurus.vercel.app/): manually upload swagger files at `/openapi/jan.yaml` to update the API reference documentation. From 7cd2a76828b6ce93424d967a32f31a065d4e2b97 Mon Sep 17 00:00:00 2001 From: hieu-jan <150573299+hieu-jan@users.noreply.github.com> Date: Sun, 10 Mar 2024 17:31:26 +0900 Subject: [PATCH 03/37] docs: update README.md --- docs/README.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/docs/README.md b/docs/README.md index 2b27b354b..02330bc80 100644 --- a/docs/README.md +++ b/docs/README.md @@ -1,3 +1,5 @@ +# Website & Docs + This website is built using [Docusaurus 3.0](https://docusaurus.io/), a modern static website generator. ### Information Architecture From 6c9fa258138a1d045851c0e2c965e9d71c4b3d0f Mon Sep 17 00:00:00 2001 From: Henry <150573299+hieu-jan@users.noreply.github.com> Date: Sun, 10 Mar 2024 17:45:06 +0900 Subject: [PATCH 04/37] delete docs/docs/.gitignore --- docs/docs/.gitignore | 1 - 1 file changed, 1 deletion(-) delete mode 100644 docs/docs/.gitignore diff --git a/docs/docs/.gitignore b/docs/docs/.gitignore deleted file mode 100644 index ff2a8c5fd..000000000 --- a/docs/docs/.gitignore +++ /dev/null @@ -1 +0,0 @@ -guides/changelogs \ No newline at end of file From 7026d2d594d593b6d7efd991f2ba827c4a12e767 Mon Sep 17 00:00:00 2001 From: hieu-jan <150573299+hieu-jan@users.noreply.github.com> Date: Sun, 10 Mar 2024 18:19:19 +0900 Subject: [PATCH 05/37] feat: update changelog configuration --- docs/plugins/changelog-plugin/fetchData.js | 21 +++------------------ 1 file changed, 3 insertions(+), 18 deletions(-) diff --git a/docs/plugins/changelog-plugin/fetchData.js b/docs/plugins/changelog-plugin/fetchData.js index dce30c147..351ab3932 100644 --- a/docs/plugins/changelog-plugin/fetchData.js +++ b/docs/plugins/changelog-plugin/fetchData.js @@ -6,14 +6,13 @@ async function fetchData(siteConfig) { const repo = siteConfig.projectName; const apiUrl = `https://api.github.com/repos/${owner}/${repo}/releases`; - const outputDirectory = path.join(__dirname, '../../docs/guides/changelogs'); + const outputDirectory = path.join(__dirname, '../../docs/releases/changelog'); if (!fs.existsSync(outputDirectory)) { fs.mkdirSync(outputDirectory); } let counter = 1; - const categoryFilePath = path.join(outputDirectory, '_category_.json'); const cacheFilePath = path.join(outputDirectory, 'cache.json'); let cachedData = {}; @@ -83,7 +82,7 @@ async function fetchData(siteConfig) { const changes = release.body; - let markdownContent = `---\nsidebar_position: ${counter}\n---\n# ${version}\n\nFor more details, [GitHub Issues](${releaseUrl})\n\nHighlighted Issue: ${issueLink}\n\n${changes}\n`; + let markdownContent = `---\nsidebar_position: ${counter}\nslug: /changelog/changelog-${version}\n---\n# ${version}\n\nFor more details, [GitHub Issues](${releaseUrl})\n\nHighlighted Issue: ${issueLink}\n\n${changes}\n`; // Write to a separate markdown file for each version const outputFilePath = path.join(outputDirectory, `changelog-${version}.mdx`); @@ -93,20 +92,6 @@ async function fetchData(siteConfig) { counter++; } - - // Create _category_.json file - const categoryContent = { - label: 'Changelogs', - position: 5, - link: { - type: 'generated-index', - description: 'Changelog for Jan', - }, - }; - - fs.writeFileSync(categoryFilePath, JSON.stringify(categoryContent, null, 2), 'utf-8'); - - console.log(`_category_.json has been created at: ${categoryFilePath}`); } -module.exports = fetchData; +module.exports = fetchData; \ No newline at end of file From 90d208a22827a01d016f1fec8f507cc87ad46744 Mon Sep 17 00:00:00 2001 From: hieu-jan <150573299+hieu-jan@users.noreply.github.com> Date: Sun, 10 Mar 2024 18:28:53 +0900 Subject: [PATCH 06/37] feat: fetching changelogs --- .../releases/changelog/changelog-v0.2.0.mdx | 72 ++-- .../releases/changelog/changelog-v0.2.1.mdx | 164 ++++---- .../releases/changelog/changelog-v0.2.2.mdx | 64 +-- .../releases/changelog/changelog-v0.2.3.mdx | 38 +- .../releases/changelog/changelog-v0.3.0.mdx | 38 +- .../releases/changelog/changelog-v0.3.1.mdx | 134 +++---- .../releases/changelog/changelog-v0.3.2.mdx | 70 ++-- .../releases/changelog/changelog-v0.3.3.mdx | 180 ++++----- .../releases/changelog/changelog-v0.4.0.mdx | 128 +++--- .../releases/changelog/changelog-v0.4.1.mdx | 102 ++--- .../releases/changelog/changelog-v0.4.2.mdx | 78 ++-- .../releases/changelog/changelog-v0.4.3.mdx | 116 +++--- .../releases/changelog/changelog-v0.4.4.mdx | 374 +++++++++--------- .../releases/changelog/changelog-v0.4.5.mdx | 184 ++++----- .../releases/changelog/changelog-v0.4.6.mdx | 106 ++--- .../releases/changelog/changelog-v0.4.7.mdx | 214 +++++----- 16 files changed, 1031 insertions(+), 1031 deletions(-) diff --git a/docs/docs/releases/changelog/changelog-v0.2.0.mdx b/docs/docs/releases/changelog/changelog-v0.2.0.mdx index 2115095ca..5e2225cf3 100644 --- a/docs/docs/releases/changelog/changelog-v0.2.0.mdx +++ b/docs/docs/releases/changelog/changelog-v0.2.0.mdx @@ -8,40 +8,40 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.2 Highlighted Issue: [Issue #342: feat: Add Jan Hacker House event page to Docs](https://github.com/janhq/jan/pull/342) -## Changes - -- feat: Add Jan Hacker House event page to Docs @dan-jan (#342) -- feat: Hide incomplete Hardware section from Docs site @dan-jan (#341) -- style: better chatbox ui @0xSage (#338) -- feat: allowing user to fetch models from github @namchuai (#319) -- fixes: #247 - inference plugin should check nitro service available @louis-jan (#313) -- Fix icon error for linux app @hiento09 (#316) -- docs: initial hardware content @Its-Alamin-H (#240) -- fixes #277 - bug: memory utilization always at 99% @louis-jan (#309) -- Docusaurus parser string from githubapi to get latest release @hiento09 (#312) -- Footer background, CTA \& Highlight colors @drakehere (#288) -- Fix CI Test run failed on ubuntu and change release file app name @hiento09 (#307) -- Add docusaurus test build pipeline @hiento09 (#302) -- fix: #271 Cannot read properties of undefined (reading 'map') @louis-jan (#300) -- Fix Docusaurus server side render error @hiento09 (#301) -- fix #283: small ui fixes @namchuai (#299) - -## 🐛 Bug Fixes - -- Fix #290: Add description in package.json and rename to jan @hiento09 (#333) - -## 🧰 Maintenance - -- Add Documentation category to release note template @hiento09 (#332) -- Chore/release note template @hiento09 (#323) -- Add release note template @hiento09 (#322) - -## 📖 Documentaion - -- Add auto update app download url on jan.ai @hiento09 (#311) -- docs: update per v0.1.3 @0xSage (#280) - -## Contributor - -@0xSage, @Its-Alamin-H, @dan-jan, @drakehere, @hiento09, @hientominh, @louis-jan, @namchuai, Hien To and James +## Changes + +- feat: Add Jan Hacker House event page to Docs @dan-jan (#342) +- feat: Hide incomplete Hardware section from Docs site @dan-jan (#341) +- style: better chatbox ui @0xSage (#338) +- feat: allowing user to fetch models from github @namchuai (#319) +- fixes: #247 - inference plugin should check nitro service available @louis-jan (#313) +- Fix icon error for linux app @hiento09 (#316) +- docs: initial hardware content @Its-Alamin-H (#240) +- fixes #277 - bug: memory utilization always at 99% @louis-jan (#309) +- Docusaurus parser string from githubapi to get latest release @hiento09 (#312) +- Footer background, CTA \& Highlight colors @drakehere (#288) +- Fix CI Test run failed on ubuntu and change release file app name @hiento09 (#307) +- Add docusaurus test build pipeline @hiento09 (#302) +- fix: #271 Cannot read properties of undefined (reading 'map') @louis-jan (#300) +- Fix Docusaurus server side render error @hiento09 (#301) +- fix #283: small ui fixes @namchuai (#299) + +## 🐛 Bug Fixes + +- Fix #290: Add description in package.json and rename to jan @hiento09 (#333) + +## 🧰 Maintenance + +- Add Documentation category to release note template @hiento09 (#332) +- Chore/release note template @hiento09 (#323) +- Add release note template @hiento09 (#322) + +## 📖 Documentaion + +- Add auto update app download url on jan.ai @hiento09 (#311) +- docs: update per v0.1.3 @0xSage (#280) + +## Contributor + +@0xSage, @Its-Alamin-H, @dan-jan, @drakehere, @hiento09, @hientominh, @louis-jan, @namchuai, Hien To and James diff --git a/docs/docs/releases/changelog/changelog-v0.2.1.mdx b/docs/docs/releases/changelog/changelog-v0.2.1.mdx index 5afc5999b..97e1052a3 100644 --- a/docs/docs/releases/changelog/changelog-v0.2.1.mdx +++ b/docs/docs/releases/changelog/changelog-v0.2.1.mdx @@ -8,86 +8,86 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.2 Highlighted Issue: [Issue #446: fix: model is started but the indicator is not stopped loading](https://github.com/janhq/jan/pull/446) -## Changes - -- fix: model is started but the indicator is not stopped loading @louis-jan (#446) -- fix: bring back install plugin manually function @louis-jan (#448) -- fix: duplicated messages when user switch between conversations @namchuai (#441) -- chore: added loader starting and stopping model @urmauur (#438) -- chore: Change license to AGPL @dan-jan (#442) -- fix: plugin \& model catalog import cache are not cleared properly @louis-jan (#437) -- fix error codesign @hiento09 (#439) -- fix: app version and cleanup unused code @urmauur (#434) -- chore: update core service - get plugin manifest @louis-jan (#432) -- ui: interface revamp @urmauur (#429) -- fix: scroll on explore models does not work @namchuai (#427) -- feat: adding create bot functionality @namchuai (#368) -- chore: install or update a plugin should not interrupt dev process @louis-jan (#420) -- chore: Update nitro 0.1.2 windows/ linux @vuonghoainam (#421) -- chore: update core service enums @louis-jan (#414) -- feat: chat with documents plugin @louis-jan (#417) -- misc: setup prettier @urmauur (#412) -- adr: 007 - Jan Plugin Catalog @louis-jan (#408) -- adr: 006 - Jan Core Module @louis-jan (#404) -- feat: Support for nitro release 0.1.2 @vuonghoainam (#409) -- feat: explore plugins from the npm repository and install them remotely @louis-jan (#399) -- feat: fix event description @dan-jan (#400) -- fix: high cpu usage @louis-jan (#401) -- docs: model installation ADR @0xSage (#390) -- chore: update core events module @louis-jan (#394) -- feat: Update Social OG Image and Meta Description @dan-jan (#387) -- misc: UI home @urmauur (#392) -- Update hcmc-oct23.md @0xSage (#389) -- chore: remove deprecated extension functions @louis-jan (#388) -- Fix bugs image overlap dropdown button download @urmauur (#384) -- chore: resolve fetch models api limit rate @louis-jan (#383) -- chore: update convo summary @louis-jan (#378) -- Update interface landing page @urmauur (#381) -- Add simple copywriting changes @dan-jan (#382) -- chore: update core services and module export @louis-jan (#376) -- chore: #371 - reference to plugin name and module path as variables @louis-jan (#372) -- feat: Edit event details, hide all unnecessary website sections @dan-jan (#369) -- docs: UI Service ADR @0xSage (#318) -- Feat/issue 255 adr 001 jand cloud native @nam-john-ho (#262) -- Move plugins folder from electron to root folder @hiento09 (#366) -- feature: core plugin support events \& preferences modules @louis-jan (#365) -- Fix/250 @namchuai (#349) -- Change License and update README @dan-jan (#356) -- Jan 339 @dan-jan (#348) -- feat: Jan 339 @dan-jan (#347) -- Add social og:image @dan-jan (#346) -- feat(ard): Add adr 002 @vuonghoainam (#261) - -## 🚀 Features - -- #357 plugin \& app can subscribe and emit events @louis-jan (#358) -- feature: @janhq/plugin-core module \& usage @louis-jan (#321) - -## 🐛 Bug Fixes - -- Change to load nitron on windows and linux from bash/shell script @hiento09 (#451) -- Fix data-plugin install failed on mac silicon from npm @hiento09 (#413) -- Correct version of plugins @hiento09 (#374) - -## 🧰 Maintenance - -- upgrade leveldown to newest version @hiento09 (#447) -- Update auto-sign plugin by search file permission 664 @hiento09 (#445) -- Change codesign plugin folder in ci @hiento09 (#440) -- Add continue on error for import cert @hiento09 (#436) -- Update code siging for new data plugin @hiento09 (#433) -- Add readme inference plugin @hiento09 (#426) -- Add username to remote origin @hiento09 (#425) -- Add auto create PR to plugin-catalog when a new version of plugin is … @hiento09 (#416) -- Fix data-plugin install failed on mac silicon from npm @hiento09 (#413) -- Chore/remove package lock @hiento09 (#398) -- Refactor cicd @hiento09 (#397) -- Correct version of plugins @hiento09 (#374) -- Rename plugin-core to core @hiento09 (#370) -- Fix error check change in plugins folder @hiento09 (#367) -- chore: jan.ai nits @0xSage (#354) - -## Contributor - -@0xSage, @dan-jan, @hiento09, @jan-service-account, @louis-jan, @nam-john-ho, @namchuai, @tikikun, @urmauur, @vuonghoainam and Hien To +## Changes + +- fix: model is started but the indicator is not stopped loading @louis-jan (#446) +- fix: bring back install plugin manually function @louis-jan (#448) +- fix: duplicated messages when user switch between conversations @namchuai (#441) +- chore: added loader starting and stopping model @urmauur (#438) +- chore: Change license to AGPL @dan-jan (#442) +- fix: plugin \& model catalog import cache are not cleared properly @louis-jan (#437) +- fix error codesign @hiento09 (#439) +- fix: app version and cleanup unused code @urmauur (#434) +- chore: update core service - get plugin manifest @louis-jan (#432) +- ui: interface revamp @urmauur (#429) +- fix: scroll on explore models does not work @namchuai (#427) +- feat: adding create bot functionality @namchuai (#368) +- chore: install or update a plugin should not interrupt dev process @louis-jan (#420) +- chore: Update nitro 0.1.2 windows/ linux @vuonghoainam (#421) +- chore: update core service enums @louis-jan (#414) +- feat: chat with documents plugin @louis-jan (#417) +- misc: setup prettier @urmauur (#412) +- adr: 007 - Jan Plugin Catalog @louis-jan (#408) +- adr: 006 - Jan Core Module @louis-jan (#404) +- feat: Support for nitro release 0.1.2 @vuonghoainam (#409) +- feat: explore plugins from the npm repository and install them remotely @louis-jan (#399) +- feat: fix event description @dan-jan (#400) +- fix: high cpu usage @louis-jan (#401) +- docs: model installation ADR @0xSage (#390) +- chore: update core events module @louis-jan (#394) +- feat: Update Social OG Image and Meta Description @dan-jan (#387) +- misc: UI home @urmauur (#392) +- Update hcmc-oct23.md @0xSage (#389) +- chore: remove deprecated extension functions @louis-jan (#388) +- Fix bugs image overlap dropdown button download @urmauur (#384) +- chore: resolve fetch models api limit rate @louis-jan (#383) +- chore: update convo summary @louis-jan (#378) +- Update interface landing page @urmauur (#381) +- Add simple copywriting changes @dan-jan (#382) +- chore: update core services and module export @louis-jan (#376) +- chore: #371 - reference to plugin name and module path as variables @louis-jan (#372) +- feat: Edit event details, hide all unnecessary website sections @dan-jan (#369) +- docs: UI Service ADR @0xSage (#318) +- Feat/issue 255 adr 001 jand cloud native @nam-john-ho (#262) +- Move plugins folder from electron to root folder @hiento09 (#366) +- feature: core plugin support events \& preferences modules @louis-jan (#365) +- Fix/250 @namchuai (#349) +- Change License and update README @dan-jan (#356) +- Jan 339 @dan-jan (#348) +- feat: Jan 339 @dan-jan (#347) +- Add social og:image @dan-jan (#346) +- feat(ard): Add adr 002 @vuonghoainam (#261) + +## 🚀 Features + +- #357 plugin \& app can subscribe and emit events @louis-jan (#358) +- feature: @janhq/plugin-core module \& usage @louis-jan (#321) + +## 🐛 Bug Fixes + +- Change to load nitron on windows and linux from bash/shell script @hiento09 (#451) +- Fix data-plugin install failed on mac silicon from npm @hiento09 (#413) +- Correct version of plugins @hiento09 (#374) + +## 🧰 Maintenance + +- upgrade leveldown to newest version @hiento09 (#447) +- Update auto-sign plugin by search file permission 664 @hiento09 (#445) +- Change codesign plugin folder in ci @hiento09 (#440) +- Add continue on error for import cert @hiento09 (#436) +- Update code siging for new data plugin @hiento09 (#433) +- Add readme inference plugin @hiento09 (#426) +- Add username to remote origin @hiento09 (#425) +- Add auto create PR to plugin-catalog when a new version of plugin is … @hiento09 (#416) +- Fix data-plugin install failed on mac silicon from npm @hiento09 (#413) +- Chore/remove package lock @hiento09 (#398) +- Refactor cicd @hiento09 (#397) +- Correct version of plugins @hiento09 (#374) +- Rename plugin-core to core @hiento09 (#370) +- Fix error check change in plugins folder @hiento09 (#367) +- chore: jan.ai nits @0xSage (#354) + +## Contributor + +@0xSage, @dan-jan, @hiento09, @jan-service-account, @louis-jan, @nam-john-ho, @namchuai, @tikikun, @urmauur, @vuonghoainam and Hien To diff --git a/docs/docs/releases/changelog/changelog-v0.2.2.mdx b/docs/docs/releases/changelog/changelog-v0.2.2.mdx index 54a317bff..54d1a8bbf 100644 --- a/docs/docs/releases/changelog/changelog-v0.2.2.mdx +++ b/docs/docs/releases/changelog/changelog-v0.2.2.mdx @@ -8,36 +8,36 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.2 Highlighted Issue: [Issue #469: chore: plugin and app version dependency](https://github.com/janhq/jan/pull/469) -## Changes - -- chore: plugin and app version dependency @louis-jan (#469) -- bug: showing a modal when user start conf but model not active @urmauur (#466) -- fix: duplicated modal and loading state @louis-jan (#465) -- bug: fix overflow scroll horizontal message @urmauur (#464) -- bug: avoid chat body scroll horizontal @urmauur (#462) -- bug: fix logic plugin update plugin and show installed version @urmauur (#459) -- bug: chat view drops enumeration @urmauur (#456) -- fix: allow switching models when switch between conversations @namchuai (#458) -- fix: CI run fails on windows @louis-jan (#463) -- fix: failed to build electron app @louis-jan (#461) -- fix: correct app version display @louis-jan (#452) -- fix: enable link color blue on docusaurus markdown @urmauur (#449) - -## 🚀 Features - -- feat: Add ADR-008 for extensible Jan @vuonghoainam (#431) - -## 🐛 Bug Fixes - -- data-plugin force leveldown to 6.1.1 @hiento09 (#453) - -## 🧰 Maintenance - -- Use electron-rebuild to build leveldown@5.6.0 for darwin arm64 @hiento09 (#455) -- data-plugin force leveldown back to 5.6.0 and rebuild for darwin arm64 @hiento09 (#454) -- data-plugin force leveldown to 6.1.1 @hiento09 (#453) - -## Contributor - -@hiento09, @jan-service-account, @louis-jan, @namchuai, @urmauur and @vuonghoainam +## Changes + +- chore: plugin and app version dependency @louis-jan (#469) +- bug: showing a modal when user start conf but model not active @urmauur (#466) +- fix: duplicated modal and loading state @louis-jan (#465) +- bug: fix overflow scroll horizontal message @urmauur (#464) +- bug: avoid chat body scroll horizontal @urmauur (#462) +- bug: fix logic plugin update plugin and show installed version @urmauur (#459) +- bug: chat view drops enumeration @urmauur (#456) +- fix: allow switching models when switch between conversations @namchuai (#458) +- fix: CI run fails on windows @louis-jan (#463) +- fix: failed to build electron app @louis-jan (#461) +- fix: correct app version display @louis-jan (#452) +- fix: enable link color blue on docusaurus markdown @urmauur (#449) + +## 🚀 Features + +- feat: Add ADR-008 for extensible Jan @vuonghoainam (#431) + +## 🐛 Bug Fixes + +- data-plugin force leveldown to 6.1.1 @hiento09 (#453) + +## 🧰 Maintenance + +- Use electron-rebuild to build leveldown@5.6.0 for darwin arm64 @hiento09 (#455) +- data-plugin force leveldown back to 5.6.0 and rebuild for darwin arm64 @hiento09 (#454) +- data-plugin force leveldown to 6.1.1 @hiento09 (#453) + +## Contributor + +@hiento09, @jan-service-account, @louis-jan, @namchuai, @urmauur and @vuonghoainam diff --git a/docs/docs/releases/changelog/changelog-v0.2.3.mdx b/docs/docs/releases/changelog/changelog-v0.2.3.mdx index 20ae6e077..72dfbbd8a 100644 --- a/docs/docs/releases/changelog/changelog-v0.2.3.mdx +++ b/docs/docs/releases/changelog/changelog-v0.2.3.mdx @@ -8,23 +8,23 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.2 Highlighted Issue: [Issue #482: fix: hide preferences section if empty](https://github.com/janhq/jan/pull/482) -## Changes - -- fix: hide preferences section if empty @louis-jan (#482) -- chore: fix conversation summary @louis-jan (#480) -- chore: missing create conversation button when there is no conversation is selected @louis-jan (#478) -- fix: download now change state immediately @namchuai (#475) -- chore: add required app version to edge release plugin @louis-jan (#471) - -## 🐛 Bug Fixes - -- add rebuild for mac x64 @hiento09 (#473) - -## 🧰 Maintenance - -- Add build deps for data-plugin in CI @hiento09 (#472) - -## Contributor - -@hiento09, @hientominh, @jan-service-account, @louis-jan and @namchuai +## Changes + +- fix: hide preferences section if empty @louis-jan (#482) +- chore: fix conversation summary @louis-jan (#480) +- chore: missing create conversation button when there is no conversation is selected @louis-jan (#478) +- fix: download now change state immediately @namchuai (#475) +- chore: add required app version to edge release plugin @louis-jan (#471) + +## 🐛 Bug Fixes + +- add rebuild for mac x64 @hiento09 (#473) + +## 🧰 Maintenance + +- Add build deps for data-plugin in CI @hiento09 (#472) + +## Contributor + +@hiento09, @hientominh, @jan-service-account, @louis-jan and @namchuai diff --git a/docs/docs/releases/changelog/changelog-v0.3.0.mdx b/docs/docs/releases/changelog/changelog-v0.3.0.mdx index 5c0d94400..603cc8f29 100644 --- a/docs/docs/releases/changelog/changelog-v0.3.0.mdx +++ b/docs/docs/releases/changelog/changelog-v0.3.0.mdx @@ -8,23 +8,23 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.3 Highlighted Issue: [Issue #482: fix: hide preferences section if empty](https://github.com/janhq/jan/pull/482) -## Changes - -- fix: hide preferences section if empty @louis-jan (#482) -- chore: fix conversation summary @louis-jan (#480) -- chore: missing create conversation button when there is no conversation is selected @louis-jan (#478) -- fix: download now change state immediately @namchuai (#475) -- chore: add required app version to edge release plugin @louis-jan (#471) - -## 🐛 Bug Fixes - -- add rebuild for mac x64 @hiento09 (#473) - -## 🧰 Maintenance - -- Add build deps for data-plugin in CI @hiento09 (#472) - -## Contributor - -@hiento09, @hientominh, @jan-service-account, @louis-jan and @namchuai +## Changes + +- fix: hide preferences section if empty @louis-jan (#482) +- chore: fix conversation summary @louis-jan (#480) +- chore: missing create conversation button when there is no conversation is selected @louis-jan (#478) +- fix: download now change state immediately @namchuai (#475) +- chore: add required app version to edge release plugin @louis-jan (#471) + +## 🐛 Bug Fixes + +- add rebuild for mac x64 @hiento09 (#473) + +## 🧰 Maintenance + +- Add build deps for data-plugin in CI @hiento09 (#472) + +## Contributor + +@hiento09, @hientominh, @jan-service-account, @louis-jan and @namchuai diff --git a/docs/docs/releases/changelog/changelog-v0.3.1.mdx b/docs/docs/releases/changelog/changelog-v0.3.1.mdx index c636db7c8..531882196 100644 --- a/docs/docs/releases/changelog/changelog-v0.3.1.mdx +++ b/docs/docs/releases/changelog/changelog-v0.3.1.mdx @@ -8,71 +8,71 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.3 Highlighted Issue: [Issue #580: fix: preformatted text indents the first line strangely](https://github.com/janhq/jan/pull/580) -## Changes - -- fix: preformatted text indents the first line strangely @louis-jan (#580) -- fix: failed to package app since core and uikit are not being built @louis-jan (#575) -- cleanup: remove component folder and cleanup conversation screen @urmauur (#574) -- bug: update convo state when user change model @urmauur (#571) -- fix(#566): jan cannot retrieve the conversations @namchuai (#570) -- bug: Toast messages shows [object object] @urmauur (#569) -- ui: improve state of welcome screen @urmauur (#563) -- chore: fixed an issue where app does not yield message result @louis-jan (#561) -- Update readme @urmauur (#560) -- ui: standalone UIKit and refactor @urmauur (#557) -- Small description changes @dan-jan (#558) -- add 'change download button based on OS' feature @Vikram-2004 (#551) -- feat: revamp plugin architecture @louis-jan (#535) -- Fix mobile padding @imtuyethan (#550) -- chore: Update Readme @dan-jan (#549) -- Update Homepage and README with 1-line pitch @dan-jan (#548) -- docs: Add About, Events, Blog @dan-jan (#546) -- Ashley/update website content @imtuyethan (#545) -- Add guides @hahuyhoang411 (#488) -- Structure Docs @dan-jan (#536) -- Update README.md @imtuyethan (#533) -- Chore: Setup "Jan Improvements Proposal" workflow @dan-jan (#534) -- Update website tag line @imtuyethan (#527) -- fix: #396 - allow user to cancel a model download @louis-jan (#530) -- fix: #479 - Toggle plugin is now experimental feature @louis-jan (#531) -- chore: disable app update on test @louis-jan (#521) -- bug: chat UI is not consistent @urmauur (#520) -- refactor: plugin manager and execution as ts @louis-jan (#504) -- fix: app toolbar is gone on windows @louis-jan (#503) -- Chore: refactor code, hide plugin menu in web @ghost (#502) -- fix: dest.end is not a function @louis-jan (#501) -- #255: Jan cloud native @ghost (#320) -- bug: download new version should show in status bar @urmauur (#500) -- feat: add New Conversation button on the conversation sidebar @urmauur (#499) -- chore: update plugin readme @louis-jan (#497) -- chore: update plugins license @louis-jan (#496) -- #255: Read plugins manifest from CDN @ghost (#495) -- chore: update plugin sdk - add appDataPath @louis-jan (#492) -- chore: enable back bot function for edge-release @louis-jan (#474) -- chore: attempt to kill Nitro subprocesses @louis-jan (#484) -- docs: new dev hub @0xSage (#450) - -## 🚀 Features - -- feat: Experimental Feature Toggle @louis-jan (#525) - -## 🐛 Bug Fixes - -- Add rebuild leveldown for arm on mac intel @hiento09 (#487) - -## 🧰 Maintenance - -- Bump nitro version from 0.1.4 to 0.1.6 @hiento09 (#581) -- Add set yarn network timeout for uikit @hiento09 (#579) -- Fix error CI e2e run failed on windows @hiento09 (#578) -- Fix build plugins macos codesiging error @hiento09 (#576) -- Add install nitro mac intel inference plugin build locally @hiento09 (#542) -- Bump nitro version to 0.1.4 @hiento09 (#532) -- Chore/update yarn dev script @hiento09 (#529) -- Inference Plugin pull nitro binary from release @hiento09 (#524) -- Correct version and license @hiento09 (#498) - -## Contributor - -@0xSage, @Vikram-2004, @dan-jan, @hahuyhoang411, @hiento09, @imtuyethan, @jan-service-account, @louis-jan, @namchuai, @tikikun, @urmauur, Han, James, John and nam-john-ho +## Changes + +- fix: preformatted text indents the first line strangely @louis-jan (#580) +- fix: failed to package app since core and uikit are not being built @louis-jan (#575) +- cleanup: remove component folder and cleanup conversation screen @urmauur (#574) +- bug: update convo state when user change model @urmauur (#571) +- fix(#566): jan cannot retrieve the conversations @namchuai (#570) +- bug: Toast messages shows [object object] @urmauur (#569) +- ui: improve state of welcome screen @urmauur (#563) +- chore: fixed an issue where app does not yield message result @louis-jan (#561) +- Update readme @urmauur (#560) +- ui: standalone UIKit and refactor @urmauur (#557) +- Small description changes @dan-jan (#558) +- add 'change download button based on OS' feature @Vikram-2004 (#551) +- feat: revamp plugin architecture @louis-jan (#535) +- Fix mobile padding @imtuyethan (#550) +- chore: Update Readme @dan-jan (#549) +- Update Homepage and README with 1-line pitch @dan-jan (#548) +- docs: Add About, Events, Blog @dan-jan (#546) +- Ashley/update website content @imtuyethan (#545) +- Add guides @hahuyhoang411 (#488) +- Structure Docs @dan-jan (#536) +- Update README.md @imtuyethan (#533) +- Chore: Setup "Jan Improvements Proposal" workflow @dan-jan (#534) +- Update website tag line @imtuyethan (#527) +- fix: #396 - allow user to cancel a model download @louis-jan (#530) +- fix: #479 - Toggle plugin is now experimental feature @louis-jan (#531) +- chore: disable app update on test @louis-jan (#521) +- bug: chat UI is not consistent @urmauur (#520) +- refactor: plugin manager and execution as ts @louis-jan (#504) +- fix: app toolbar is gone on windows @louis-jan (#503) +- Chore: refactor code, hide plugin menu in web @ghost (#502) +- fix: dest.end is not a function @louis-jan (#501) +- #255: Jan cloud native @ghost (#320) +- bug: download new version should show in status bar @urmauur (#500) +- feat: add New Conversation button on the conversation sidebar @urmauur (#499) +- chore: update plugin readme @louis-jan (#497) +- chore: update plugins license @louis-jan (#496) +- #255: Read plugins manifest from CDN @ghost (#495) +- chore: update plugin sdk - add appDataPath @louis-jan (#492) +- chore: enable back bot function for edge-release @louis-jan (#474) +- chore: attempt to kill Nitro subprocesses @louis-jan (#484) +- docs: new dev hub @0xSage (#450) + +## 🚀 Features + +- feat: Experimental Feature Toggle @louis-jan (#525) + +## 🐛 Bug Fixes + +- Add rebuild leveldown for arm on mac intel @hiento09 (#487) + +## 🧰 Maintenance + +- Bump nitro version from 0.1.4 to 0.1.6 @hiento09 (#581) +- Add set yarn network timeout for uikit @hiento09 (#579) +- Fix error CI e2e run failed on windows @hiento09 (#578) +- Fix build plugins macos codesiging error @hiento09 (#576) +- Add install nitro mac intel inference plugin build locally @hiento09 (#542) +- Bump nitro version to 0.1.4 @hiento09 (#532) +- Chore/update yarn dev script @hiento09 (#529) +- Inference Plugin pull nitro binary from release @hiento09 (#524) +- Correct version and license @hiento09 (#498) + +## Contributor + +@0xSage, @Vikram-2004, @dan-jan, @hahuyhoang411, @hiento09, @imtuyethan, @jan-service-account, @louis-jan, @namchuai, @tikikun, @urmauur, Han, James, John and nam-john-ho diff --git a/docs/docs/releases/changelog/changelog-v0.3.2.mdx b/docs/docs/releases/changelog/changelog-v0.3.2.mdx index 01e403ef9..b62a878d1 100644 --- a/docs/docs/releases/changelog/changelog-v0.3.2.mdx +++ b/docs/docs/releases/changelog/changelog-v0.3.2.mdx @@ -8,39 +8,39 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.3 Highlighted Issue: [Issue #612: fix: disabled required env](https://github.com/janhq/jan/pull/612) -## Changes - -- fix: disabled required env @urmauur (#612) -- Install Posthog snippet @imtuyethan (#573) -- web: google tag manager @urmauur (#562) -- docs: fix syntax highlighting @0xSage (#602) -- chore: remove past event @0xSage (#600) -- docs: new docs @0xSage (#599) -- [chore]: Update docs @dan-jan (#597) - -## 🚀 Features - -- refactor: main electron with managers and handlers @louis-jan (#610) - -## 🐛 Bug Fixes - -- Fix: Failed to load model - unload model nitro @louis-jan (#616) -- Restore cpx nitro step in yarn script @hiento09 (#617) -- fix(#591): prevent duplicate message id issue @namchuai (#595) -- bug: cancelling a model download should be delete the model file on user data @urmauur (#613) -- bug: fix weird padding vertical snippet code @urmauur (#608) -- bug: Fix button download detect intel or apple silicon @urmauur (#609) -- bug: enable delete conversation after deleted model @urmauur (#594) -- bug: download modal should truncate model name @urmauur (#592) -- bug: support multiple line input chat using Textarea instead @urmauur (#593) - -## 🧰 Maintenance - -- refactor: main electron with managers and handlers @louis-jan (#610) -- Chore/refactor yarn script @hiento09 (#615) -- fix: line height and update typography component @urmauur (#611) - -## Contributor - -@0xSage, @dan-jan, @hiento09, @imtuyethan, @jan-service-account, @louis-jan, @namchuai and @urmauur +## Changes + +- fix: disabled required env @urmauur (#612) +- Install Posthog snippet @imtuyethan (#573) +- web: google tag manager @urmauur (#562) +- docs: fix syntax highlighting @0xSage (#602) +- chore: remove past event @0xSage (#600) +- docs: new docs @0xSage (#599) +- [chore]: Update docs @dan-jan (#597) + +## 🚀 Features + +- refactor: main electron with managers and handlers @louis-jan (#610) + +## 🐛 Bug Fixes + +- Fix: Failed to load model - unload model nitro @louis-jan (#616) +- Restore cpx nitro step in yarn script @hiento09 (#617) +- fix(#591): prevent duplicate message id issue @namchuai (#595) +- bug: cancelling a model download should be delete the model file on user data @urmauur (#613) +- bug: fix weird padding vertical snippet code @urmauur (#608) +- bug: Fix button download detect intel or apple silicon @urmauur (#609) +- bug: enable delete conversation after deleted model @urmauur (#594) +- bug: download modal should truncate model name @urmauur (#592) +- bug: support multiple line input chat using Textarea instead @urmauur (#593) + +## 🧰 Maintenance + +- refactor: main electron with managers and handlers @louis-jan (#610) +- Chore/refactor yarn script @hiento09 (#615) +- fix: line height and update typography component @urmauur (#611) + +## Contributor + +@0xSage, @dan-jan, @hiento09, @imtuyethan, @jan-service-account, @louis-jan, @namchuai and @urmauur diff --git a/docs/docs/releases/changelog/changelog-v0.3.3.mdx b/docs/docs/releases/changelog/changelog-v0.3.3.mdx index 9d8e75076..66a871186 100644 --- a/docs/docs/releases/changelog/changelog-v0.3.3.mdx +++ b/docs/docs/releases/changelog/changelog-v0.3.3.mdx @@ -8,94 +8,94 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.3 Highlighted Issue: [Issue #719: docs: cleanup](https://github.com/janhq/jan/pull/719) -## Changes - -- docs: cleanup @0xSage (#719) -- docs: threads and messages @0xSage (#681) -- Updating Onboarding Kit @Diane0111 (#675) -- Update issue templates @0xSage (#685) -- docs: polish models spec @0xSage (#680) -- Feature: Preview URL for each PR and add pre-release.jan.ai as staging of Jan Docs @hiento09 (#669) -- Migrate Model definitions to Swagger/OpenAPI @dan-jan (#659) -- [docs] Add Introduction and refactor Models Spec @dan-jan (#657) -- docs: Add model methods to swagger @0xSage (#660) -- Models Spec: Delete broken Markdown links @dan-jan (#648) -- docs: assistants and threads specs @0xSage (#646) - -## 🚀 Features - -- improvement: styling message action toolbar @urmauur (#737) -- experimental: allow user to give instruction for the conversation @louis-jan (#714) -- docs/enable-seo-docusaurus @hieu-jan (#725) -- Add windows code sign to CI @hiento09 (#712) -- docs: update installation guide @hieu-jan (#664) -- chore: Update based on team discussion on Nov 20 @vuonghoainam (#673) -- docs: add OpenAI swagger file @hieu-jan (#623) -- Update landing page Jan @urmauur (#638) - -## 🐛 Bug Fixes - -- chore: open app data should lead user to jan root @louis-jan (#749) -- fix: cancel download does not work @louis-jan (#746) -- fix: error when switching between threads @louis-jan (#736) -- chore: app raises error when attempting to start a model that is already starting @louis-jan (#721) -- bug: fix filter list menu from command base on search type and make a symbol base on OS @urmauur (#723) -- bug: fix clickable small download button on chat screen @urmauur (#722) -- fix: incorrect update progress bar visibility check @louis-jan (#713) -- fix: app shows wrong performance tag, all say not enough ram on windows @louis-jan (#699) -- bug: fix padding quotations and numbering list @urmauur (#695) -- fix: local npm module update does not reflect web app @louis-jan (#677) -- [bug] fix markdown todo items shifted to the left and remove the dots @urmauur (#694) -- bug: fix footer and section spacing landing page @urmauur (#683) -- bug: fix anchor link sidebar openapi @urmauur (#668) -- refactor: remove unused hooks and resolve no-explicit-any @louis-jan (#647) -- bug: fix titles should have spaces in between @urmauur (#652) -- bug: fix compatibility content not fully display @urmauur (#653) - -## 🧰 Maintenance - -- chore: fix app grammar @0xSage (#750) -- chore: bumb nitro version @louis-jan (#740) -- chore: fs module should not cover app logic @louis-jan (#720) -- API Reference for Models, Messages, Threads @hahuyhoang411 (#679) -- docs: upgrade mdx-js package @hieu-jan (#705) -- [docs] Update Docusaurus to 3.0 and fix closing tag issue in Handbook @dan-jan (#704) -- Fix error docs pipeline run failed @hiento09 (#702) -- Revert docs CICD trigger on push to main instead of tag-based @hiento09 (#698) -- fix: local npm module update does not reflect web app @louis-jan (#677) -- Chore: refactor to makefile @hiento09 (#691) -- Add Instruction to publish docs @hiento09 (#687) -- chore/add-mermaid @hieu-jan (#672) -- chore/update package docs @hieu-jan (#670) -- Enhance Cross-Platform Argument Handling for Nitro Startup Scripts @hiento09 (#674) -- refactor: remove unused hooks and resolve no-explicit-any @louis-jan (#647) -- docs: add OpenAI swagger file @hieu-jan (#623) -- Preliminary Restructure of Docs @dan-jan (#655) -- Model specs @vuonghoainam (#641) -- refactor: refactor app entities @louis-jan (#626) -- refactor: move file to jan root @namchuai (#598) -- Add run-script-os @linhtran174 (#620) -- Refactor Jan Documentation @dan-jan (#625) - -## 📖 Documentaion - -- docs: update specs/product @0xSage (#744) -- docs/enable-seo-docusaurus @hieu-jan (#725) -- docs: assistant spec @vuonghoainam (#707) -- docs: Refactor Jan Site Structure @dan-jan (#706) -- docs/improve install docs @hieu-jan (#708) -- API Reference for Models, Messages, Threads @hahuyhoang411 (#679) -- [docs] Update Docusaurus to 3.0 and fix closing tag issue in Handbook @dan-jan (#704) -- docs: update installation guide @hieu-jan (#664) -- chore: Update based on team discussion on Nov 20 @vuonghoainam (#673) -- docs: add OpenAI swagger file @hieu-jan (#623) -- Preliminary Restructure of Docs @dan-jan (#655) -- Fix: specs revision @vuonghoainam (#649) -- Model specs @vuonghoainam (#641) -- Update README.md @imtuyethan (#629) -- Refactor Jan Documentation @dan-jan (#625) - -## Contributor - -@0xSage, @Diane0111, @dan-jan, @hahuyhoang411, @hiento09, @hieu-jan, @imtuyethan, @linhtran174, @louis-jan, @namchuai, @urmauur, @vuonghoainam and Le Tra Mi +## Changes + +- docs: cleanup @0xSage (#719) +- docs: threads and messages @0xSage (#681) +- Updating Onboarding Kit @Diane0111 (#675) +- Update issue templates @0xSage (#685) +- docs: polish models spec @0xSage (#680) +- Feature: Preview URL for each PR and add pre-release.jan.ai as staging of Jan Docs @hiento09 (#669) +- Migrate Model definitions to Swagger/OpenAPI @dan-jan (#659) +- [docs] Add Introduction and refactor Models Spec @dan-jan (#657) +- docs: Add model methods to swagger @0xSage (#660) +- Models Spec: Delete broken Markdown links @dan-jan (#648) +- docs: assistants and threads specs @0xSage (#646) + +## 🚀 Features + +- improvement: styling message action toolbar @urmauur (#737) +- experimental: allow user to give instruction for the conversation @louis-jan (#714) +- docs/enable-seo-docusaurus @hieu-jan (#725) +- Add windows code sign to CI @hiento09 (#712) +- docs: update installation guide @hieu-jan (#664) +- chore: Update based on team discussion on Nov 20 @vuonghoainam (#673) +- docs: add OpenAI swagger file @hieu-jan (#623) +- Update landing page Jan @urmauur (#638) + +## 🐛 Bug Fixes + +- chore: open app data should lead user to jan root @louis-jan (#749) +- fix: cancel download does not work @louis-jan (#746) +- fix: error when switching between threads @louis-jan (#736) +- chore: app raises error when attempting to start a model that is already starting @louis-jan (#721) +- bug: fix filter list menu from command base on search type and make a symbol base on OS @urmauur (#723) +- bug: fix clickable small download button on chat screen @urmauur (#722) +- fix: incorrect update progress bar visibility check @louis-jan (#713) +- fix: app shows wrong performance tag, all say not enough ram on windows @louis-jan (#699) +- bug: fix padding quotations and numbering list @urmauur (#695) +- fix: local npm module update does not reflect web app @louis-jan (#677) +- [bug] fix markdown todo items shifted to the left and remove the dots @urmauur (#694) +- bug: fix footer and section spacing landing page @urmauur (#683) +- bug: fix anchor link sidebar openapi @urmauur (#668) +- refactor: remove unused hooks and resolve no-explicit-any @louis-jan (#647) +- bug: fix titles should have spaces in between @urmauur (#652) +- bug: fix compatibility content not fully display @urmauur (#653) + +## 🧰 Maintenance + +- chore: fix app grammar @0xSage (#750) +- chore: bumb nitro version @louis-jan (#740) +- chore: fs module should not cover app logic @louis-jan (#720) +- API Reference for Models, Messages, Threads @hahuyhoang411 (#679) +- docs: upgrade mdx-js package @hieu-jan (#705) +- [docs] Update Docusaurus to 3.0 and fix closing tag issue in Handbook @dan-jan (#704) +- Fix error docs pipeline run failed @hiento09 (#702) +- Revert docs CICD trigger on push to main instead of tag-based @hiento09 (#698) +- fix: local npm module update does not reflect web app @louis-jan (#677) +- Chore: refactor to makefile @hiento09 (#691) +- Add Instruction to publish docs @hiento09 (#687) +- chore/add-mermaid @hieu-jan (#672) +- chore/update package docs @hieu-jan (#670) +- Enhance Cross-Platform Argument Handling for Nitro Startup Scripts @hiento09 (#674) +- refactor: remove unused hooks and resolve no-explicit-any @louis-jan (#647) +- docs: add OpenAI swagger file @hieu-jan (#623) +- Preliminary Restructure of Docs @dan-jan (#655) +- Model specs @vuonghoainam (#641) +- refactor: refactor app entities @louis-jan (#626) +- refactor: move file to jan root @namchuai (#598) +- Add run-script-os @linhtran174 (#620) +- Refactor Jan Documentation @dan-jan (#625) + +## 📖 Documentaion + +- docs: update specs/product @0xSage (#744) +- docs/enable-seo-docusaurus @hieu-jan (#725) +- docs: assistant spec @vuonghoainam (#707) +- docs: Refactor Jan Site Structure @dan-jan (#706) +- docs/improve install docs @hieu-jan (#708) +- API Reference for Models, Messages, Threads @hahuyhoang411 (#679) +- [docs] Update Docusaurus to 3.0 and fix closing tag issue in Handbook @dan-jan (#704) +- docs: update installation guide @hieu-jan (#664) +- chore: Update based on team discussion on Nov 20 @vuonghoainam (#673) +- docs: add OpenAI swagger file @hieu-jan (#623) +- Preliminary Restructure of Docs @dan-jan (#655) +- Fix: specs revision @vuonghoainam (#649) +- Model specs @vuonghoainam (#641) +- Update README.md @imtuyethan (#629) +- Refactor Jan Documentation @dan-jan (#625) + +## Contributor + +@0xSage, @Diane0111, @dan-jan, @hahuyhoang411, @hiento09, @hieu-jan, @imtuyethan, @linhtran174, @louis-jan, @namchuai, @urmauur, @vuonghoainam and Le Tra Mi diff --git a/docs/docs/releases/changelog/changelog-v0.4.0.mdx b/docs/docs/releases/changelog/changelog-v0.4.0.mdx index 3afa7849e..142d55a60 100644 --- a/docs/docs/releases/changelog/changelog-v0.4.0.mdx +++ b/docs/docs/releases/changelog/changelog-v0.4.0.mdx @@ -8,68 +8,68 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.4 Highlighted Issue: [Issue #878: bug: fix tag description showing a title and fix card right panel](https://github.com/janhq/jan/pull/878) -## Changes - -- bug: fix tag description showing a title and fix card right panel @urmauur (#878) -- fix/no-assistant-available-fresh-install @louis-jan (#876) -- Model.json update @hahuyhoang411 (#870) -- Hotfix desc for openhermes @hahuyhoang411 (#864) -- Openhermes update v1 @hahuyhoang411 (#863) -- update deepseek 1.3b @hahuyhoang411 (#858) -- Update tags @hahuyhoang411 (#857) -- Update model hub @hahuyhoang411 (#829) -- hotfix: fix typo @tikikun (#836) -- chore: pre-populate Jan's /models folder with model.jsons @hahuyhoang411 (#775) -- chore: clarification changes to the model settings and model parameters @tikikun (#742) - -## 🚀 Features - -- feat: revamp landing page @urmauur (#745) -- feat : add cover image model hub screen @urmauur (#872) -- feat: boilerplate for express server localhost 1337 @linhtran174 (#803) -- enhancement: revamp hub screen @urmauur (#825) -- feat: revamp thread screen @urmauur (#802) -- docs/update-api-reference @hieu-jan (#739) -- refactor: model plugin to follow new specs @namchuai (#682) - -## 🐛 Fixes - -- fix: Nitro interface update to prevent warning @vuonghoainam (#877) -- fix: delete message break the entire thread @louis-jan (#869) -- fix: can not download multiple models at once @louis-jan (#867) -- fix: production CI workflow does not populate models @louis-jan (#862) -- fix: update wrong main view state when use a model @namchuai (#861) -- fix: handle crash issue on hljs highlighting @louis-jan (#859) -- fix: empty assistant instruction by default @louis-jan (#855) -- bug: fix broken banner position hub screen @urmauur (#846) -- fix: not update active model when using resend button @namchuai (#834) -- Hotfix jan windows download nitro failed @hiento09 (#838) -- Switch to download nitro .tar.gz file instead of .zip file on windows @hiento09 (#832) -- fix/docusaurus-seo @hieu-jan (#818) -- fix: CI script - reorder copy models action @louis-jan (#819) -- fix: messages sync is not threadsafe @louis-jan (#784) -- Fix Makefile Indentation Issue @hiento09 (#788) - -## 🧰 Maintenance - -- chore: update model ranking @louis-jan (#874) -- Bump nitro version to 0.1.21 - nitro has windows codesign @hiento09 (#843) -- Hotfix jan windows download nitro failed @hiento09 (#838) -- 810 docs add modeljson and revamp models specs page @tikikun (#816) -- Add document for nightly build and update message for manual build @hiento09 (#831) -- chore: Bump nitro to 0.1.20 @vuonghoainam (#830) -- Refactor build:extension command @hiento09 (#822) -- feat: pre-populate Jan's /models folder @namchuai (#796) -- chore: fix pr auto labeling @0xSage (#812) -- chore: add gi automations @0xSage (#809) -- refactor: jan extensions @louis-jan (#799) -- Remove .zip in artifact name @hiento09 (#800) -- docs/update-api-reference @hieu-jan (#739) -- Add nightly build ci @hiento09 (#794) -- Fix Makefile Indentation Issue @hiento09 (#788) -- Switch from .zip to .tar.gz for nitro url inference plugin @hiento09 (#781) - -## Contributor - -@0xSage, @hahuyhoang411, @hiento09, @hieu-jan, @linhtran174, @louis-jan, @namchuai, @tikikun, @urmauur and @vuonghoainam +## Changes + +- bug: fix tag description showing a title and fix card right panel @urmauur (#878) +- fix/no-assistant-available-fresh-install @louis-jan (#876) +- Model.json update @hahuyhoang411 (#870) +- Hotfix desc for openhermes @hahuyhoang411 (#864) +- Openhermes update v1 @hahuyhoang411 (#863) +- update deepseek 1.3b @hahuyhoang411 (#858) +- Update tags @hahuyhoang411 (#857) +- Update model hub @hahuyhoang411 (#829) +- hotfix: fix typo @tikikun (#836) +- chore: pre-populate Jan's /models folder with model.jsons @hahuyhoang411 (#775) +- chore: clarification changes to the model settings and model parameters @tikikun (#742) + +## 🚀 Features + +- feat: revamp landing page @urmauur (#745) +- feat : add cover image model hub screen @urmauur (#872) +- feat: boilerplate for express server localhost 1337 @linhtran174 (#803) +- enhancement: revamp hub screen @urmauur (#825) +- feat: revamp thread screen @urmauur (#802) +- docs/update-api-reference @hieu-jan (#739) +- refactor: model plugin to follow new specs @namchuai (#682) + +## 🐛 Fixes + +- fix: Nitro interface update to prevent warning @vuonghoainam (#877) +- fix: delete message break the entire thread @louis-jan (#869) +- fix: can not download multiple models at once @louis-jan (#867) +- fix: production CI workflow does not populate models @louis-jan (#862) +- fix: update wrong main view state when use a model @namchuai (#861) +- fix: handle crash issue on hljs highlighting @louis-jan (#859) +- fix: empty assistant instruction by default @louis-jan (#855) +- bug: fix broken banner position hub screen @urmauur (#846) +- fix: not update active model when using resend button @namchuai (#834) +- Hotfix jan windows download nitro failed @hiento09 (#838) +- Switch to download nitro .tar.gz file instead of .zip file on windows @hiento09 (#832) +- fix/docusaurus-seo @hieu-jan (#818) +- fix: CI script - reorder copy models action @louis-jan (#819) +- fix: messages sync is not threadsafe @louis-jan (#784) +- Fix Makefile Indentation Issue @hiento09 (#788) + +## 🧰 Maintenance + +- chore: update model ranking @louis-jan (#874) +- Bump nitro version to 0.1.21 - nitro has windows codesign @hiento09 (#843) +- Hotfix jan windows download nitro failed @hiento09 (#838) +- 810 docs add modeljson and revamp models specs page @tikikun (#816) +- Add document for nightly build and update message for manual build @hiento09 (#831) +- chore: Bump nitro to 0.1.20 @vuonghoainam (#830) +- Refactor build:extension command @hiento09 (#822) +- feat: pre-populate Jan's /models folder @namchuai (#796) +- chore: fix pr auto labeling @0xSage (#812) +- chore: add gi automations @0xSage (#809) +- refactor: jan extensions @louis-jan (#799) +- Remove .zip in artifact name @hiento09 (#800) +- docs/update-api-reference @hieu-jan (#739) +- Add nightly build ci @hiento09 (#794) +- Fix Makefile Indentation Issue @hiento09 (#788) +- Switch from .zip to .tar.gz for nitro url inference plugin @hiento09 (#781) + +## Contributor + +@0xSage, @hahuyhoang411, @hiento09, @hieu-jan, @linhtran174, @louis-jan, @namchuai, @tikikun, @urmauur and @vuonghoainam diff --git a/docs/docs/releases/changelog/changelog-v0.4.1.mdx b/docs/docs/releases/changelog/changelog-v0.4.1.mdx index 10cf4308a..38be38b74 100644 --- a/docs/docs/releases/changelog/changelog-v0.4.1.mdx +++ b/docs/docs/releases/changelog/changelog-v0.4.1.mdx @@ -8,55 +8,55 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.4 Highlighted Issue: [Issue #903: Update README.md](https://github.com/janhq/jan/pull/903) -## Changes - -- Update README.md @imtuyethan (#903) - -## 🚀 Features - -- feat: Kill nitro process with API - nitro 0.1.27 @vuonghoainam (#975) -- feat: Inference Nitro with Prompt Template @hahuyhoang411 (#952) -- feat: Add NVIDIA triton trt-llm extension @vuonghoainam (#888) -- feat: Hotfit for Nitro loading on CPU with hyper-threading support @vuonghoainam (#931) -- feat: adding model params @namchuai (#886) -- feat: Multiple inference engines for nitro and openai @vuonghoainam (#814) -- docs: add json schema for engine and model parameters @tikikun (#840) -- feat: improve SEO keywords @hieu-jan (#894) -- enhancement: fix spacing landing page responsive @urmauur (#891) -- bug: added label coming soon for windows and linux @urmauur (#881) - -## 🐛 Fixes - -- fix: 963 can not run openai models on windows @louis-jan (#974) -- fix: Inference engine Nitro with Windows with/ without CUDA @vuonghoainam (#950) -- Fix error Jan app linux crash @hiento09 (#958) -- fix: windows bug - control buttons close,max,min hidden @linhtran174 (#949) -- bug: fix ui landing page @urmauur (#937) -- fix: model parameters for inference extensions @vuonghoainam (#935) -- [bug] Fix floating border outside card right panel @urmauur (#934) -- fix: import\_typescript.default.isTokenKind is not a function @louis-jan (#923) -- bug: fix syntax formatting @urmauur (#899) -- bug: update metadata title and desc @urmauur (#884) -- fix: download button text color is blending into the background @louis-jan (#883) - -## 🧰 Maintenance - -- chore: add desktop app analytics @louis-jan (#978) -- refactor: clean types and interfaces @0xSage (#966) -- docs: scaffold dev docs @0xSage (#856) -- chore: Bump nitro to 0.1.26 @vuonghoainam (#960) -- Update update-release-url.yml @hiento09 (#951) -- Fix update release url pipeline run failed @hiento09 (#947) -- chore: Bumpt nitro bin version to version 0.1.23 @vuonghoainam (#942) -- Fix update release url pipeline @hiento09 (#941) -- CI automatically update Update README with Nightly Build Information and stable download URL @hiento09 (#940) -- refactor: deprecate invokers - auto proxying apis - strict types @louis-jan (#924) -- docs: standardize yaml files @hieu-jan (#933) -- chore: universal module definition @louis-jan (#902) -- docs: add assistants api reference @hieu-jan (#801) -- docs: add json schema for engine and model parameters @tikikun (#840) - -## Contributor - -@0xSage, @hahuyhoang411, @hiento09, @hieu-jan, @imtuyethan, @jan-service-account, @linhtran174, @louis-jan, @namchuai, @tikikun, @urmauur and @vuonghoainam +## Changes + +- Update README.md @imtuyethan (#903) + +## 🚀 Features + +- feat: Kill nitro process with API - nitro 0.1.27 @vuonghoainam (#975) +- feat: Inference Nitro with Prompt Template @hahuyhoang411 (#952) +- feat: Add NVIDIA triton trt-llm extension @vuonghoainam (#888) +- feat: Hotfit for Nitro loading on CPU with hyper-threading support @vuonghoainam (#931) +- feat: adding model params @namchuai (#886) +- feat: Multiple inference engines for nitro and openai @vuonghoainam (#814) +- docs: add json schema for engine and model parameters @tikikun (#840) +- feat: improve SEO keywords @hieu-jan (#894) +- enhancement: fix spacing landing page responsive @urmauur (#891) +- bug: added label coming soon for windows and linux @urmauur (#881) + +## 🐛 Fixes + +- fix: 963 can not run openai models on windows @louis-jan (#974) +- fix: Inference engine Nitro with Windows with/ without CUDA @vuonghoainam (#950) +- Fix error Jan app linux crash @hiento09 (#958) +- fix: windows bug - control buttons close,max,min hidden @linhtran174 (#949) +- bug: fix ui landing page @urmauur (#937) +- fix: model parameters for inference extensions @vuonghoainam (#935) +- [bug] Fix floating border outside card right panel @urmauur (#934) +- fix: import\_typescript.default.isTokenKind is not a function @louis-jan (#923) +- bug: fix syntax formatting @urmauur (#899) +- bug: update metadata title and desc @urmauur (#884) +- fix: download button text color is blending into the background @louis-jan (#883) + +## 🧰 Maintenance + +- chore: add desktop app analytics @louis-jan (#978) +- refactor: clean types and interfaces @0xSage (#966) +- docs: scaffold dev docs @0xSage (#856) +- chore: Bump nitro to 0.1.26 @vuonghoainam (#960) +- Update update-release-url.yml @hiento09 (#951) +- Fix update release url pipeline run failed @hiento09 (#947) +- chore: Bumpt nitro bin version to version 0.1.23 @vuonghoainam (#942) +- Fix update release url pipeline @hiento09 (#941) +- CI automatically update Update README with Nightly Build Information and stable download URL @hiento09 (#940) +- refactor: deprecate invokers - auto proxying apis - strict types @louis-jan (#924) +- docs: standardize yaml files @hieu-jan (#933) +- chore: universal module definition @louis-jan (#902) +- docs: add assistants api reference @hieu-jan (#801) +- docs: add json schema for engine and model parameters @tikikun (#840) + +## Contributor + +@0xSage, @hahuyhoang411, @hiento09, @hieu-jan, @imtuyethan, @jan-service-account, @linhtran174, @louis-jan, @namchuai, @tikikun, @urmauur and @vuonghoainam diff --git a/docs/docs/releases/changelog/changelog-v0.4.2.mdx b/docs/docs/releases/changelog/changelog-v0.4.2.mdx index 4b64ca7f4..77b15a555 100644 --- a/docs/docs/releases/changelog/changelog-v0.4.2.mdx +++ b/docs/docs/releases/changelog/changelog-v0.4.2.mdx @@ -8,43 +8,43 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.4 Highlighted Issue: [Issue #1033: Hotfix model hub](https://github.com/janhq/jan/pull/1033) -## Changes - -- Hotfix model hub @hahuyhoang411 (#1033) -- Update Model.json @hahuyhoang411 (#1005) - -## 🚀 Features - -- feat: app theme depend on local storage instead native theme electron @urmauur (#1014) -- feat: move stop inference button into the send button @urmauur (#1011) -- feat: loader when starting model @urmauur (#945) -- fix: enable download app linux @urmauur (#993) -- fix: remove coming soon windows @urmauur (#986) - -## 🐛 Fixes - -- fix: migrate new models @louis-jan (#1034) -- fix: add input for api key remote model @urmauur (#1031) -- fix bug #1013, enable posthog for release app version only @hiento09 (#1019) -- fix: delete first message then regenerate again will break the thread @louis-jan (#1015) -- fix: #995 - Fix onboarding state and model sorting @louis-jan (#1009) -- fix: limit analytics events capture @louis-jan (#1012) -- fix: wrong selected model right panel @urmauur (#1001) -- fix: review finder and view as json @louis-jan (#1000) -- fix: enable download app linux @urmauur (#993) - -## 🧰 Maintenance - -- chore: remigrate if there is no models dir @louis-jan (#1038) -- bump nitro version to 0.1.30 @hiento09 (#1036) -- chore: in app copy fixes @0xSage (#1032) -- Separate posthog project for jan app and docs @hiento09 (#1029) -- Update posthog capture url list @hiento09 (#1022) -- docs: second half of "import model docs" PR @0xSage (#1021) -- docs: how to import models @0xSage (#1020) -- fix bug #1013, enable posthog for release app version only @hiento09 (#1019) - -## Contributor - -@0xSage, @hahuyhoang411, @hiento09, @jan-service-account, @louis-jan and @urmauur +## Changes + +- Hotfix model hub @hahuyhoang411 (#1033) +- Update Model.json @hahuyhoang411 (#1005) + +## 🚀 Features + +- feat: app theme depend on local storage instead native theme electron @urmauur (#1014) +- feat: move stop inference button into the send button @urmauur (#1011) +- feat: loader when starting model @urmauur (#945) +- fix: enable download app linux @urmauur (#993) +- fix: remove coming soon windows @urmauur (#986) + +## 🐛 Fixes + +- fix: migrate new models @louis-jan (#1034) +- fix: add input for api key remote model @urmauur (#1031) +- fix bug #1013, enable posthog for release app version only @hiento09 (#1019) +- fix: delete first message then regenerate again will break the thread @louis-jan (#1015) +- fix: #995 - Fix onboarding state and model sorting @louis-jan (#1009) +- fix: limit analytics events capture @louis-jan (#1012) +- fix: wrong selected model right panel @urmauur (#1001) +- fix: review finder and view as json @louis-jan (#1000) +- fix: enable download app linux @urmauur (#993) + +## 🧰 Maintenance + +- chore: remigrate if there is no models dir @louis-jan (#1038) +- bump nitro version to 0.1.30 @hiento09 (#1036) +- chore: in app copy fixes @0xSage (#1032) +- Separate posthog project for jan app and docs @hiento09 (#1029) +- Update posthog capture url list @hiento09 (#1022) +- docs: second half of "import model docs" PR @0xSage (#1021) +- docs: how to import models @0xSage (#1020) +- fix bug #1013, enable posthog for release app version only @hiento09 (#1019) + +## Contributor + +@0xSage, @hahuyhoang411, @hiento09, @jan-service-account, @louis-jan and @urmauur diff --git a/docs/docs/releases/changelog/changelog-v0.4.3.mdx b/docs/docs/releases/changelog/changelog-v0.4.3.mdx index 2426f6280..d34c85f60 100644 --- a/docs/docs/releases/changelog/changelog-v0.4.3.mdx +++ b/docs/docs/releases/changelog/changelog-v0.4.3.mdx @@ -8,62 +8,62 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.4 Highlighted Issue: [Issue #1159: Hotfix Prompt template for models on the Hub](https://github.com/janhq/jan/pull/1159) -## Changes - -- Hotfix Prompt template for models on the Hub @hahuyhoang411 (#1159) -- Update model list for new release @hahuyhoang411 (#1143) -- fix(Thread): #1119 focus on the first thread to prevent blank chat screen @namchuai (#1127) -- fix(Thread): #1064 message being added to wrong thread if switching thread @namchuai (#1108) -- fix(Thread): #1042 allow create new thread by clicking Use in Jan Hub @namchuai (#1103) -- feat(ModelSetting): #1065 update state of model setting between threads @namchuai (#1090) -- Update model version @hahuyhoang411 (#1086) -- fix: cache hallucinations and failed to load model due to race condition @louis-jan (#1071) -- fix(thread): #1043 default model to prefer active model @namchuai (#1070) -- Update issue templates @0xSage (#1058) -- Update ctx\_len and max\_tokens @hahuyhoang411 (#1035) - -## 🚀 Features - -- feat: Add codeQL analysis for push main and pr main @hiro-v (#1128) -- Feature autoupdater for nightly build @hiento09 (#1068) -- feat: copy button for code block @urmauur (#1062) -- Enhancements to Dependency Installation and App Testing @hiento09 (#965) - -## 🐛 Fixes - -- fix: error road map url @hieu-jan (#1153) -- Fix token speed slow in machine has multi gpus @hiento09 (#1157) -- fix: added dialog confirmation clean thread @urmauur (#1142) -- fix: remove remote model from shortcut models dialog @urmauur (#1124) -- fix: ui issue - all models are activated @louis-jan (#1120) -- fix: should not hide empty message away @louis-jan (#1116) -- fix: added tooltip for user cannot change model after starting thread @urmauur (#1115) -- fix: remote model always active badges @urmauur (#1113) -- fix: handle chat completion state with enter button @louis-jan (#1114) -- fix: model active indicator only show when model activated @urmauur (#1110) -- fix: #1096 yield error message upon thread switching @louis-jan (#1109) -- fix: toaster success deleted thread showing id instead of active model @urmauur (#1111) -- fix: update copy setting page @urmauur (#1105) -- fix: search recommended model @urmauur (#1106) -- fix: #1097 streaming response is replaced by error message @louis-jan (#1099) -- Fix auto update windows Bug @hiento09 (#1102) -- fix: added dialog confirmation when delete thread @urmauur (#1093) -- fix: system monitor broken layout when responsive @urmauur (#1085) -- bug: chatbox doesn't resize back down @urmauur (#1084) -- fix: thread is broken after deleted first generated message @louis-jan (#1061) - -## 🧰 Maintenance - -- feat: Add codeQL analysis for push main and pr main @hiro-v (#1128) -- docs: refactor dev docs, guides and specs @0xSage (#1092) -- Correct jq command cause ci nightly build run failed @hiento09 (#1104) -- Fix nightly build autoupdater @hiento09 (#1073) -- Feature autoupdater for nightly build @hiento09 (#1068) -- docs: Update product.md @0xSage (#1066) -- Posthog disable click event and increase timeout for nitro load model… @hiento09 (#1060) -- docs: improve quickstart docs @0xSage (#1047) - -## Contributor - -@0xSage, @hahuyhoang411, @hiento09, @hieu-jan, @hiro-v, @jan-service-account, @louis-jan, @namchuai and @urmauur +## Changes + +- Hotfix Prompt template for models on the Hub @hahuyhoang411 (#1159) +- Update model list for new release @hahuyhoang411 (#1143) +- fix(Thread): #1119 focus on the first thread to prevent blank chat screen @namchuai (#1127) +- fix(Thread): #1064 message being added to wrong thread if switching thread @namchuai (#1108) +- fix(Thread): #1042 allow create new thread by clicking Use in Jan Hub @namchuai (#1103) +- feat(ModelSetting): #1065 update state of model setting between threads @namchuai (#1090) +- Update model version @hahuyhoang411 (#1086) +- fix: cache hallucinations and failed to load model due to race condition @louis-jan (#1071) +- fix(thread): #1043 default model to prefer active model @namchuai (#1070) +- Update issue templates @0xSage (#1058) +- Update ctx\_len and max\_tokens @hahuyhoang411 (#1035) + +## 🚀 Features + +- feat: Add codeQL analysis for push main and pr main @hiro-v (#1128) +- Feature autoupdater for nightly build @hiento09 (#1068) +- feat: copy button for code block @urmauur (#1062) +- Enhancements to Dependency Installation and App Testing @hiento09 (#965) + +## 🐛 Fixes + +- fix: error road map url @hieu-jan (#1153) +- Fix token speed slow in machine has multi gpus @hiento09 (#1157) +- fix: added dialog confirmation clean thread @urmauur (#1142) +- fix: remove remote model from shortcut models dialog @urmauur (#1124) +- fix: ui issue - all models are activated @louis-jan (#1120) +- fix: should not hide empty message away @louis-jan (#1116) +- fix: added tooltip for user cannot change model after starting thread @urmauur (#1115) +- fix: remote model always active badges @urmauur (#1113) +- fix: handle chat completion state with enter button @louis-jan (#1114) +- fix: model active indicator only show when model activated @urmauur (#1110) +- fix: #1096 yield error message upon thread switching @louis-jan (#1109) +- fix: toaster success deleted thread showing id instead of active model @urmauur (#1111) +- fix: update copy setting page @urmauur (#1105) +- fix: search recommended model @urmauur (#1106) +- fix: #1097 streaming response is replaced by error message @louis-jan (#1099) +- Fix auto update windows Bug @hiento09 (#1102) +- fix: added dialog confirmation when delete thread @urmauur (#1093) +- fix: system monitor broken layout when responsive @urmauur (#1085) +- bug: chatbox doesn't resize back down @urmauur (#1084) +- fix: thread is broken after deleted first generated message @louis-jan (#1061) + +## 🧰 Maintenance + +- feat: Add codeQL analysis for push main and pr main @hiro-v (#1128) +- docs: refactor dev docs, guides and specs @0xSage (#1092) +- Correct jq command cause ci nightly build run failed @hiento09 (#1104) +- Fix nightly build autoupdater @hiento09 (#1073) +- Feature autoupdater for nightly build @hiento09 (#1068) +- docs: Update product.md @0xSage (#1066) +- Posthog disable click event and increase timeout for nitro load model… @hiento09 (#1060) +- docs: improve quickstart docs @0xSage (#1047) + +## Contributor + +@0xSage, @hahuyhoang411, @hiento09, @hieu-jan, @hiro-v, @jan-service-account, @louis-jan, @namchuai and @urmauur diff --git a/docs/docs/releases/changelog/changelog-v0.4.4.mdx b/docs/docs/releases/changelog/changelog-v0.4.4.mdx index f61e35c44..ea85cb2de 100644 --- a/docs/docs/releases/changelog/changelog-v0.4.4.mdx +++ b/docs/docs/releases/changelog/changelog-v0.4.4.mdx @@ -8,191 +8,191 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.4 Highlighted Issue: [Issue #1587: Update 2023-11-05-hello-world.md](https://github.com/janhq/jan/pull/1587) -## Changes - -- Update 2023-11-05-hello-world.md @Ssstars (#1587) -- fix(API): #1511 update swagger page @namchuai (#1572) -- fix(Thread): #1212 thread.json not created when user change thread settings @namchuai (#1570) -- fix(Thread): #1336 not allow creating too many unfinished thread @namchuai (#1538) -- Update 01-how-to-get-involved-and-FAQ.mdx @Ssstars (#1555) -- Update 01-how-to-get-involved-and-FAQ.mdx @Ssstars (#1553) -- Update 02-embracing-pod-structure.mdx @Ssstars (#1550) -- Update 01-how-we-hire.mdx @Ssstars (#1551) -- Update 01-how-we-hire.mdx @Ssstars (#1524) -- fix(InferenceExtension): #1067 sync the nitro process state @namchuai (#1493) -- fix(Messages): #1434 create message via api does not display on app correctly @namchuai (#1479) -- Docs for the Integration of Continue and Jan in VSCode @0xgokuz (#1467) -- Chore: Update model.json for UI @hahuyhoang411 (#1448) -- Docs for Installing Models from Hub @0xgokuz (#1450) -- Update about.md @Ssstars (#1436) -- feat(UI): #1404 make left side bar collapsible by hot key @namchuai (#1420) -- docs: Typo in 06-hardware.md @akaMrNagar (#1408) -- fix(API): #1409 fix wrong prefix for threads api @namchuai (#1410) -- Update model hub @hahuyhoang411 (#1383) -- fix(Model): remove unsupported default model setting params @namchuai (#1382) -- fix(trinity): update cover path for trinity v1.2 @hahuyhoang411 (#1380) -- Chore/update model hub @hahuyhoang411 (#1342) -- Update about.md @Ssstars (#1359) -- fix(JanHub): #1158 sort model list @namchuai (#1257) -- fix(Message): open link with external browser @namchuai (#1339) -- feat(Model): #1028 made model.json optional @namchuai (#1314) -- docs: Update onboarding.md @Diane0111 (#1293) -- fix: clean resource on exit @louis-jan (#1290) -- fix: posthog configuration @hieu-jan (#1283) -- docs: update README.md @eltociear (#1277) -- Enable scrolling in the message chat box @Gri-ffin (#1280) -- chore: Update README.md @sr-albert (#1263) -- Adding new model to the Hub @hahuyhoang411 (#1213) -- Feature GPU detection for Jan on Windows and Linux @hiento09 (#1242) -- fix(Thread): #1168 fix newly created thread cannot select model after restart @namchuai (#1176) - -## 🚀 Features - -- feat: add compatibility tag to model selection in right panel @urmauur (#1552) -- Feature integrate antivirus scanner to ci @hiento09 (#1529) -- feat: [hub] update compatibility tags colors @urmauur (#1516) -- feat: hub recommendation labels @urmauur (#1440) -- Feature linux support app image format @hiento09 (#1442) -- fix: render external links @urmauur (#1441) -- fix: add icon collapsible left panel and update keyboard shortcut page @urmauur (#1439) -- feat(UI): update UI footer @urmauur (#1424) -- Fix Bug for Chat Reply Goes off Screen @mishrababhishek (#1393) -- feat: move social media from left panel into footer @urmauur (#1325) -- feat: implementation new UI thread settings @urmauur (#1301) -- Bring social media links @Gri-ffin (#1295) -- feat: added keyboard shortcut list in setting page @urmauur (#1275) -- feat: add swagger /docs to localhost:1337 @louis-jan (#1268) -- feat: update posthog configuration @hieu-jan (#1258) -- feat: Deprecate model.json ready state in favor of .download ext @louis-jan (#1238) -- feat: add engine settings @namchuai (#1199) -- feat: users should be able to switch models mid-thread @louis-jan (#1226) -- feat: temporary link how to import model @urmauur (#1209) - -## 🐛 Fixes - -- fix: #1594 - Model settings - change thread model - go back does not see according settings @louis-jan (#1595) -- fix: #1548 - duplicate command shortcut instruction @louis-jan (#1600) -- fix: switch model caused app crash @louis-jan (#1597) -- fix: #1559 Inference Parameters displayed on new thread with Openai GPT Model @louis-jan (#1588) -- fix: enable user set value manually model setting from input @urmauur (#1585) -- fix: #1569 - Does not apply thread settings when loading model @louis-jan (#1576) -- fix: could not change model params settings @louis-jan (#1547) -- fix: gpu check module export does not work in extension @louis-jan (#1536) -- fix: adjust calculation hub labels using total RAM instead remaining RAM @urmauur (#1522) -- Feature integrate antivirus scanner to ci @hiento09 (#1529) -- fix: allow users to set max tokens variably @urmauur (#1513) -- fix: stop word update @louis-jan (#1457) -- Revert nitro to 0.2.6 @hiento09 (#1491) -- fix: enable text selection codeblock @urmauur (#1466) -- fix: suppress all main node JS error message dialog @louis-jan (#1460) -- Correct AppImage path @hiento09 (#1446) -- fix: render external links @urmauur (#1441) -- fix: add icon collapsible left panel and update keyboard shortcut page @urmauur (#1439) -- fix: GET /models does not work due to new default model dir @louis-jan (#1392) -- fix: model migration stopped working @louis-jan (#1378) -- fix: wrong condition for displaying error message @louis-jan (#1376) -- fix: show hide section engine params @urmauur (#1374) -- fix: copy stream tooltip and hide section when no params setting @urmauur (#1373) -- bugs: fix stop streaming when user delete or clean thread @urmauur (#1347) -- fix: show a proper error message on download failure @louis-jan (#1345) -- Add detect cuda version jan app @hiento09 (#1351) -- fix: Error occurred: Unexpected token "d", "data: ..." is not a valid JSON @louis-jan (#1332) -- fix: app getting stuck at downloading 99% while downloading model @louis-jan (#1320) -- correct type utf-8 @hiento09 (#1311) -- Fix memory on mac included cached and swap @hiento09 (#1298) -- fix: should check app dir before spawning log @louis-jan (#1297) -- fix: disable process logging from server @louis-jan (#1296) -- fix: user should be able to access Swagger docs from localhost:1337 @louis-jan (#1292) -- Switch from Gigabyte to Gibibyte - System monitor @hiento09 (#1286) -- Switch from systeminformation to os-utils to resolve Bitdefender false positive and memory leak issue @hiento09 (#1282) -- fix: swagger CSP issue @louis-jan (#1284) -- fix: support markdown break line @urmauur (#1274) -- fix ci test run failed @hiento09 (#1267) -- Fix wrong linux nitro path @hiento09 (#1266) -- fix: enable command enter on dialog confirmation clean thread @urmauur (#1261) -- fix: input message duplicated due with some input sources @louis-jan (#1259) -- fix: mac users should not see GPU settings @louis-jan (#1255) -- fix: remove redundant gpu detection prompt event @louis-jan (#1254) -- fix: engine settings GUI - feature toggle @louis-jan (#1252) -- Fix bug #1178 high ram on windows @hiento09 (#1241) -- fix: #1183 Reveal in finder does not work on windows @namchuai (#1239) -- fix: remove delay tooltip and click event @urmauur (#1217) -- fix: enable enter command on dialog confirmation delete thread @urmauur (#1218) -- fix: Cleared thread last message is not updated @louis-jan (#1225) -- Fix switch thread crash nitro windows linux @hiento09 (#1214) -- fix: darkmode broken color @urmauur (#1186) - -## 🧰 Maintenance - -- chore: typo model.json @louis-jan (#1599) -- docs: add 04-how-to-get-error-logs.mdx @hieu-jan (#1580) -- chore: teach how to attach logs @0xSage (#1578) -- chore: issues should auto close with PRs through template @0xSage (#1577) -- chore: Update issue templates @0xSage (#1568) -- docs: fix x handles @0xSage (#1532) -- Docs to integrate OpenRouter with Jan without UI/UX @0xgokuz (#1495) -- chore: fix darkmode docs @hieu-jan (#1520) -- docs: fix algolia configuration @hieu-jan (#1518) -- docs: fix algolia configuration @hieu-jan (#1517) -- Revert URL release in readme to version 0.4.3 @hiento09 (#1502) -- refactor: add app and nitro log - resolve dependencies issue @louis-jan (#1447) -- chore: enable agolia @hieu-jan (#1497) -- docs: update troubleshooting and redirects old pages @hieu-jan (#1492) -- docs: minor fix @hieu-jan (#1478) -- docs: initial handbook structure @hieu-jan (#1477) -- Bump nitro to 0.2.8 and change Jan App to support cuda >= 11.7 @hiento09 (#1476) -- Chore update docs jan - add AppImage instruction to docusaurus @hiento09 (#1480) -- Bump nitro to 0.2.7 @hiento09 (#1474) -- chore: error message update @louis-jan (#1473) -- docs: Update 02-import-manually.mdx @0xSage (#1469) -- docs: Update about.md @0xSage (#1465) -- Bump nitro version to 0.2.6 @hiento09 (#1458) -- docs: adding customize engine settings @hieu-jan (#1455) -- docs: add-missing-path @hieu-jan (#1454) -- docs: resize gif @hieu-jan (#1453) -- docs: revenue philosophy @0xSage (#1443) -- docs: jan framework principles @0xSage (#1438) -- docs: fix typo in docs @hieu-jan (#1419) -- chore: clean up use os hook @namchuai (#1418) -- docs: explain each docs page intent @0xSage (#1417) -- docs: Update 01-server.md @0xSage (#1416) -- Update warning url from github md file to jan.ai docs site @hiento09 (#1414) -- docs: improve gpu not used guide @hieu-jan (#1405) -- chore: update README.md @eltociear (#1406) -- Update USAGE docs for linux @hiento09 (#1401) -- docs: gpu not detected @0xSage (#1399) -- docs: Troubleshoot Failed To Fetch @gabrielle-ong (#1398) -- docs: improve docs syntax @hieu-jan (#1394) -- docs: add-install-nightly-guide @hieu-jan (#1390) -- docs: correct href link @hieu-jan (#1338) -- docs: fix chat payload and cURL @hieu-jan (#1360) -- docs: add Chatting Guide @hieu-jan (#1184) -- Chore add docs usage how to switch run mode jan app @hiento09 (#1353) -- docs: configure index page @hieu-jan (#1330) -- docs: Update product.md @0xSage (#1326) -- docs: Update 01-server.md @0xSage (#1327) -- refactor: deprecate the appVersion IPC and use the predefined VERSION @louis-jan (#1309) -- docs: update using models documentation @hieu-jan (#1288) -- docs: update pm handbook @0xSage (#1307) -- docs: contributor docs overview @0xSage (#1305) -- chore: github PR template @0xSage (#1304) -- Fix memory on mac included cached and swap @hiento09 (#1298) -- Enrich discord message for nightly build url @hiento09 (#1294) -- Refactor CI by create shared jobs output @hiento09 (#1287) -- docs: update README.md @hieu-jan (#1281) -- docs: Update README.md @0xSage (#1248) -- feat: Jan Server, API and decoupled clients @louis-jan (#948) -- docs: improve 02-import-manually @hieu-jan (#1222) -- chore: Update issue templates @0xSage (#1229) -- docs: Update 02-import-manually.mdx @0xSage (#1197) -- add sleep 500ms if platform is windows before starting nitro process @hiento09 (#1215) -- docs: improve troubleshoot documentation @hieu-jan (#1173) -- docs: update bug report template @hieu-jan (#1180) -- docs: add troubleshooting @hieu-jan (#1169) -- chore: copy fixes @0xSage (#1167) -- docs: Update 01-start-thread.md @0xSage (#1122) - -## Contributor - -@0xSage, @0xgokuz, @Diane0111, @Gri-ffin, @Ssstars, @akaMrNagar, @eltociear, @gabrielle-ong, @hahuyhoang411, @hiento09, @hieu-jan, @jan-service-account, @louis-jan, @mishrababhishek, @namchuai, @sr-albert, @urmauur and Abhishek Mishra +## Changes + +- Update 2023-11-05-hello-world.md @Ssstars (#1587) +- fix(API): #1511 update swagger page @namchuai (#1572) +- fix(Thread): #1212 thread.json not created when user change thread settings @namchuai (#1570) +- fix(Thread): #1336 not allow creating too many unfinished thread @namchuai (#1538) +- Update 01-how-to-get-involved-and-FAQ.mdx @Ssstars (#1555) +- Update 01-how-to-get-involved-and-FAQ.mdx @Ssstars (#1553) +- Update 02-embracing-pod-structure.mdx @Ssstars (#1550) +- Update 01-how-we-hire.mdx @Ssstars (#1551) +- Update 01-how-we-hire.mdx @Ssstars (#1524) +- fix(InferenceExtension): #1067 sync the nitro process state @namchuai (#1493) +- fix(Messages): #1434 create message via api does not display on app correctly @namchuai (#1479) +- Docs for the Integration of Continue and Jan in VSCode @0xgokuz (#1467) +- Chore: Update model.json for UI @hahuyhoang411 (#1448) +- Docs for Installing Models from Hub @0xgokuz (#1450) +- Update about.md @Ssstars (#1436) +- feat(UI): #1404 make left side bar collapsible by hot key @namchuai (#1420) +- docs: Typo in 06-hardware.md @akaMrNagar (#1408) +- fix(API): #1409 fix wrong prefix for threads api @namchuai (#1410) +- Update model hub @hahuyhoang411 (#1383) +- fix(Model): remove unsupported default model setting params @namchuai (#1382) +- fix(trinity): update cover path for trinity v1.2 @hahuyhoang411 (#1380) +- Chore/update model hub @hahuyhoang411 (#1342) +- Update about.md @Ssstars (#1359) +- fix(JanHub): #1158 sort model list @namchuai (#1257) +- fix(Message): open link with external browser @namchuai (#1339) +- feat(Model): #1028 made model.json optional @namchuai (#1314) +- docs: Update onboarding.md @Diane0111 (#1293) +- fix: clean resource on exit @louis-jan (#1290) +- fix: posthog configuration @hieu-jan (#1283) +- docs: update README.md @eltociear (#1277) +- Enable scrolling in the message chat box @Gri-ffin (#1280) +- chore: Update README.md @sr-albert (#1263) +- Adding new model to the Hub @hahuyhoang411 (#1213) +- Feature GPU detection for Jan on Windows and Linux @hiento09 (#1242) +- fix(Thread): #1168 fix newly created thread cannot select model after restart @namchuai (#1176) + +## 🚀 Features + +- feat: add compatibility tag to model selection in right panel @urmauur (#1552) +- Feature integrate antivirus scanner to ci @hiento09 (#1529) +- feat: [hub] update compatibility tags colors @urmauur (#1516) +- feat: hub recommendation labels @urmauur (#1440) +- Feature linux support app image format @hiento09 (#1442) +- fix: render external links @urmauur (#1441) +- fix: add icon collapsible left panel and update keyboard shortcut page @urmauur (#1439) +- feat(UI): update UI footer @urmauur (#1424) +- Fix Bug for Chat Reply Goes off Screen @mishrababhishek (#1393) +- feat: move social media from left panel into footer @urmauur (#1325) +- feat: implementation new UI thread settings @urmauur (#1301) +- Bring social media links @Gri-ffin (#1295) +- feat: added keyboard shortcut list in setting page @urmauur (#1275) +- feat: add swagger /docs to localhost:1337 @louis-jan (#1268) +- feat: update posthog configuration @hieu-jan (#1258) +- feat: Deprecate model.json ready state in favor of .download ext @louis-jan (#1238) +- feat: add engine settings @namchuai (#1199) +- feat: users should be able to switch models mid-thread @louis-jan (#1226) +- feat: temporary link how to import model @urmauur (#1209) + +## 🐛 Fixes + +- fix: #1594 - Model settings - change thread model - go back does not see according settings @louis-jan (#1595) +- fix: #1548 - duplicate command shortcut instruction @louis-jan (#1600) +- fix: switch model caused app crash @louis-jan (#1597) +- fix: #1559 Inference Parameters displayed on new thread with Openai GPT Model @louis-jan (#1588) +- fix: enable user set value manually model setting from input @urmauur (#1585) +- fix: #1569 - Does not apply thread settings when loading model @louis-jan (#1576) +- fix: could not change model params settings @louis-jan (#1547) +- fix: gpu check module export does not work in extension @louis-jan (#1536) +- fix: adjust calculation hub labels using total RAM instead remaining RAM @urmauur (#1522) +- Feature integrate antivirus scanner to ci @hiento09 (#1529) +- fix: allow users to set max tokens variably @urmauur (#1513) +- fix: stop word update @louis-jan (#1457) +- Revert nitro to 0.2.6 @hiento09 (#1491) +- fix: enable text selection codeblock @urmauur (#1466) +- fix: suppress all main node JS error message dialog @louis-jan (#1460) +- Correct AppImage path @hiento09 (#1446) +- fix: render external links @urmauur (#1441) +- fix: add icon collapsible left panel and update keyboard shortcut page @urmauur (#1439) +- fix: GET /models does not work due to new default model dir @louis-jan (#1392) +- fix: model migration stopped working @louis-jan (#1378) +- fix: wrong condition for displaying error message @louis-jan (#1376) +- fix: show hide section engine params @urmauur (#1374) +- fix: copy stream tooltip and hide section when no params setting @urmauur (#1373) +- bugs: fix stop streaming when user delete or clean thread @urmauur (#1347) +- fix: show a proper error message on download failure @louis-jan (#1345) +- Add detect cuda version jan app @hiento09 (#1351) +- fix: Error occurred: Unexpected token "d", "data: ..." is not a valid JSON @louis-jan (#1332) +- fix: app getting stuck at downloading 99% while downloading model @louis-jan (#1320) +- correct type utf-8 @hiento09 (#1311) +- Fix memory on mac included cached and swap @hiento09 (#1298) +- fix: should check app dir before spawning log @louis-jan (#1297) +- fix: disable process logging from server @louis-jan (#1296) +- fix: user should be able to access Swagger docs from localhost:1337 @louis-jan (#1292) +- Switch from Gigabyte to Gibibyte - System monitor @hiento09 (#1286) +- Switch from systeminformation to os-utils to resolve Bitdefender false positive and memory leak issue @hiento09 (#1282) +- fix: swagger CSP issue @louis-jan (#1284) +- fix: support markdown break line @urmauur (#1274) +- fix ci test run failed @hiento09 (#1267) +- Fix wrong linux nitro path @hiento09 (#1266) +- fix: enable command enter on dialog confirmation clean thread @urmauur (#1261) +- fix: input message duplicated due with some input sources @louis-jan (#1259) +- fix: mac users should not see GPU settings @louis-jan (#1255) +- fix: remove redundant gpu detection prompt event @louis-jan (#1254) +- fix: engine settings GUI - feature toggle @louis-jan (#1252) +- Fix bug #1178 high ram on windows @hiento09 (#1241) +- fix: #1183 Reveal in finder does not work on windows @namchuai (#1239) +- fix: remove delay tooltip and click event @urmauur (#1217) +- fix: enable enter command on dialog confirmation delete thread @urmauur (#1218) +- fix: Cleared thread last message is not updated @louis-jan (#1225) +- Fix switch thread crash nitro windows linux @hiento09 (#1214) +- fix: darkmode broken color @urmauur (#1186) + +## 🧰 Maintenance + +- chore: typo model.json @louis-jan (#1599) +- docs: add 04-how-to-get-error-logs.mdx @hieu-jan (#1580) +- chore: teach how to attach logs @0xSage (#1578) +- chore: issues should auto close with PRs through template @0xSage (#1577) +- chore: Update issue templates @0xSage (#1568) +- docs: fix x handles @0xSage (#1532) +- Docs to integrate OpenRouter with Jan without UI/UX @0xgokuz (#1495) +- chore: fix darkmode docs @hieu-jan (#1520) +- docs: fix algolia configuration @hieu-jan (#1518) +- docs: fix algolia configuration @hieu-jan (#1517) +- Revert URL release in readme to version 0.4.3 @hiento09 (#1502) +- refactor: add app and nitro log - resolve dependencies issue @louis-jan (#1447) +- chore: enable agolia @hieu-jan (#1497) +- docs: update troubleshooting and redirects old pages @hieu-jan (#1492) +- docs: minor fix @hieu-jan (#1478) +- docs: initial handbook structure @hieu-jan (#1477) +- Bump nitro to 0.2.8 and change Jan App to support cuda >= 11.7 @hiento09 (#1476) +- Chore update docs jan - add AppImage instruction to docusaurus @hiento09 (#1480) +- Bump nitro to 0.2.7 @hiento09 (#1474) +- chore: error message update @louis-jan (#1473) +- docs: Update 02-import-manually.mdx @0xSage (#1469) +- docs: Update about.md @0xSage (#1465) +- Bump nitro version to 0.2.6 @hiento09 (#1458) +- docs: adding customize engine settings @hieu-jan (#1455) +- docs: add-missing-path @hieu-jan (#1454) +- docs: resize gif @hieu-jan (#1453) +- docs: revenue philosophy @0xSage (#1443) +- docs: jan framework principles @0xSage (#1438) +- docs: fix typo in docs @hieu-jan (#1419) +- chore: clean up use os hook @namchuai (#1418) +- docs: explain each docs page intent @0xSage (#1417) +- docs: Update 01-server.md @0xSage (#1416) +- Update warning url from github md file to jan.ai docs site @hiento09 (#1414) +- docs: improve gpu not used guide @hieu-jan (#1405) +- chore: update README.md @eltociear (#1406) +- Update USAGE docs for linux @hiento09 (#1401) +- docs: gpu not detected @0xSage (#1399) +- docs: Troubleshoot Failed To Fetch @gabrielle-ong (#1398) +- docs: improve docs syntax @hieu-jan (#1394) +- docs: add-install-nightly-guide @hieu-jan (#1390) +- docs: correct href link @hieu-jan (#1338) +- docs: fix chat payload and cURL @hieu-jan (#1360) +- docs: add Chatting Guide @hieu-jan (#1184) +- Chore add docs usage how to switch run mode jan app @hiento09 (#1353) +- docs: configure index page @hieu-jan (#1330) +- docs: Update product.md @0xSage (#1326) +- docs: Update 01-server.md @0xSage (#1327) +- refactor: deprecate the appVersion IPC and use the predefined VERSION @louis-jan (#1309) +- docs: update using models documentation @hieu-jan (#1288) +- docs: update pm handbook @0xSage (#1307) +- docs: contributor docs overview @0xSage (#1305) +- chore: github PR template @0xSage (#1304) +- Fix memory on mac included cached and swap @hiento09 (#1298) +- Enrich discord message for nightly build url @hiento09 (#1294) +- Refactor CI by create shared jobs output @hiento09 (#1287) +- docs: update README.md @hieu-jan (#1281) +- docs: Update README.md @0xSage (#1248) +- feat: Jan Server, API and decoupled clients @louis-jan (#948) +- docs: improve 02-import-manually @hieu-jan (#1222) +- chore: Update issue templates @0xSage (#1229) +- docs: Update 02-import-manually.mdx @0xSage (#1197) +- add sleep 500ms if platform is windows before starting nitro process @hiento09 (#1215) +- docs: improve troubleshoot documentation @hieu-jan (#1173) +- docs: update bug report template @hieu-jan (#1180) +- docs: add troubleshooting @hieu-jan (#1169) +- chore: copy fixes @0xSage (#1167) +- docs: Update 01-start-thread.md @0xSage (#1122) + +## Contributor + +@0xSage, @0xgokuz, @Diane0111, @Gri-ffin, @Ssstars, @akaMrNagar, @eltociear, @gabrielle-ong, @hahuyhoang411, @hiento09, @hieu-jan, @jan-service-account, @louis-jan, @mishrababhishek, @namchuai, @sr-albert, @urmauur and Abhishek Mishra diff --git a/docs/docs/releases/changelog/changelog-v0.4.5.mdx b/docs/docs/releases/changelog/changelog-v0.4.5.mdx index d2e1d245b..4c9163dc7 100644 --- a/docs/docs/releases/changelog/changelog-v0.4.5.mdx +++ b/docs/docs/releases/changelog/changelog-v0.4.5.mdx @@ -8,96 +8,96 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.4 Highlighted Issue: [Issue #1758: bug: Correct text for Windows](https://github.com/janhq/jan/issues/1758) -## Changes - -- fix(Wording): #1758 correct text for windows @namchuai (#1768) -- fix(Log): server log is not display in windows @namchuai (#1764) -- Release Cut v0.4.5 @louis-jan (#1752) -- chore(nitro): 0.2.11 -> 0.2.12 @hiro-v (#1754) -- fix: Nitro CPU threads with correct physical/ performance CPU count @hiro-v (#1726) -- fix(Model): #1662 imported model does not use gpu @namchuai (#1723) -- fix(API): #1720 host/port provided in the local API server does not fully applied @namchuai (#1721) -- fix: server API reference @hiro-v (#1670) -- fix(Model): refactor model label @namchuai (#1596) -- docs/postmortem v 0.4.4 @hieu-jan (#1617) -- chore(ShortcutModal): clean up shortcut modal @namchuai (#1614) -- chore(Dependencies): upgrade node-fetch to fix vulnerable issue @namchuai (#1598) - -## 🚀 Features - -- feat: update UI allow user change folder @urmauur (#1738) -- feat: error message when not enough RAM @urmauur (#1706) -- feat: improvement ux for local api server @urmauur (#1704) -- feat: allow user to move jan folder @namchuai (#1649) -- feat: HTTP proxy support @markmehere (#1562) -- Feature add schedule clean cloudflare page and r2 @hiento09 (#1653) -- feat: relayout left panel setting page @urmauur (#1648) -- Update CI follow git flow @hiento09 (#1625) -- feat: Implement UI page API server dashboard @urmauur (#1636) -- fix: #1545 long thread title @lucido-simon (#1605) - -## 🐛 Fixes - -- fix: model selection does not show in API settings page @louis-jan (#1828) -- fix: user can't view model setting in local api server @namchuai (#1807) -- fix: cannot change jan data folder @namchuai (#1805) -- fix: model selection does not show in API settings page @louis-jan (#1802) -- fix: user can't use a model in model hub @namchuai (#1801) -- fix: stop openai inference raises something amiss @louis-jan (#1799) -- regression fix: input disabled darkmode @urmauur (#1800) -- fix: clean last message when user clean thread message @namchuai (#1793) -- fix: app log not being printed @namchuai (#1790) -- fix: api settings are not applied on changes @louis-jan (#1789) -- fix: could not delete model @louis-jan (#1779) -- fix: can not start model when server is not enabled from model settings page @louis-jan (#1774) -- regression fix: input port not accept alphabets @urmauur (#1772) -- Correct bash script syntax in ci @hiento09 (#1769) -- Hotfix CI pre-release not trigger @hiento09 (#1757) -- fix: bring back open app directory @louis-jan (#1756) -- fix: input port have range validation @urmauur (#1741) -- Fix error nightly build schedule run failed @hiento09 (#1736) -- fix: active model when start server @urmauur (#1719) -- fix: Change to fixed `localhost` instead of using host variable @hiro-v (#1729) -- Fix autoupdater nightly build error @hiento09 (#1727) -- Correct download url readme @hiento09 (#1724) -- fix: API chat/completion is blocked by CORS @louis-jan (#1705) -- fix: Jan server - v1/chat/completions is throwing ERR\_REQUIRE\_ESM @louis-jan (#1703) -- fix: Jan server is showing blank page @louis-jan (#1702) -- fix: switching loader from remote to local model from thread right panel @urmauur (#1692) -- fix: hot-fix algolia search @hieu-jan (#1700) -- fix: disable api key field while server is running @urmauur (#1694) -- fix: stoping model show starting model @urmauur (#1693) -- fix bug #1650 hogging resources @hiento09 (#1663) -- fix: auto select text when collapse panel @urmauur (#1645) -- fix: wrong selected model ref @louis-jan (#1638) -- fix: enable check for update on all supported platforms @louis-jan (#1626) -- fix: correct footer @hieu-jan (#1628) - -## 🧰 Maintenance - -- Docs publish to github page trigger on push to docs branch @hiento09 (#1783) -- Correct bash script syntax in ci @hiento09 (#1769) -- Combine 2 ci pipeline pre-release and nightly into one @hiento09 (#1767) -- Hotfix CI pre-release not trigger @hiento09 (#1757) -- Fix error nightly build schedule run failed @hiento09 (#1736) -- docs: add troubleshoot unexpected token @hieu-jan (#1711) -- docs: fix about pages @0xSage (#1699) -- refactor: deprecate extension type implementation @louis-jan (#1677) -- refactor: file prefix replace utils \& add unit test @louis-jan (#1676) -- Correct ref branch for update url on README.md file @hiento09 (#1672) -- docs: update 02-somethings-amiss @hieu-jan (#1668) -- Cherrypick cicd to main branch to apply new gitflow @hiento09 (#1665) -- docs: add user and developer guides for extensions @hieu-jan (#1657) -- docs: add QA Script @hieu-jan (#1660) -- chore: Bump nitro to 0.2.11 @hiro-v (#1655) -- chore: Bump version nitro to 0.2.10 @hiro-v (#1644) -- docs: add antivirus compatibility testing @hieu-jan (#1641) -- refactor: introduce node module in nitro extension @louis-jan (#1630) -- Update 02-somethings-amiss.mdx @Ssstars (#1634) -- docs: add integration AzureOpenAI @hieu-jan (#1632) -- docs: add troubleshooting permission denied @hieu-jan (#1631) - -## Contributor - -@0xSage, @Ssstars, @hiento09, @hientominh, @hieu-jan, @hiro-v, @jan-service-account, @louis-jan, @lucido-simon, @markmehere, @namchuai and @urmauur +## Changes + +- fix(Wording): #1758 correct text for windows @namchuai (#1768) +- fix(Log): server log is not display in windows @namchuai (#1764) +- Release Cut v0.4.5 @louis-jan (#1752) +- chore(nitro): 0.2.11 -> 0.2.12 @hiro-v (#1754) +- fix: Nitro CPU threads with correct physical/ performance CPU count @hiro-v (#1726) +- fix(Model): #1662 imported model does not use gpu @namchuai (#1723) +- fix(API): #1720 host/port provided in the local API server does not fully applied @namchuai (#1721) +- fix: server API reference @hiro-v (#1670) +- fix(Model): refactor model label @namchuai (#1596) +- docs/postmortem v 0.4.4 @hieu-jan (#1617) +- chore(ShortcutModal): clean up shortcut modal @namchuai (#1614) +- chore(Dependencies): upgrade node-fetch to fix vulnerable issue @namchuai (#1598) + +## 🚀 Features + +- feat: update UI allow user change folder @urmauur (#1738) +- feat: error message when not enough RAM @urmauur (#1706) +- feat: improvement ux for local api server @urmauur (#1704) +- feat: allow user to move jan folder @namchuai (#1649) +- feat: HTTP proxy support @markmehere (#1562) +- Feature add schedule clean cloudflare page and r2 @hiento09 (#1653) +- feat: relayout left panel setting page @urmauur (#1648) +- Update CI follow git flow @hiento09 (#1625) +- feat: Implement UI page API server dashboard @urmauur (#1636) +- fix: #1545 long thread title @lucido-simon (#1605) + +## 🐛 Fixes + +- fix: model selection does not show in API settings page @louis-jan (#1828) +- fix: user can't view model setting in local api server @namchuai (#1807) +- fix: cannot change jan data folder @namchuai (#1805) +- fix: model selection does not show in API settings page @louis-jan (#1802) +- fix: user can't use a model in model hub @namchuai (#1801) +- fix: stop openai inference raises something amiss @louis-jan (#1799) +- regression fix: input disabled darkmode @urmauur (#1800) +- fix: clean last message when user clean thread message @namchuai (#1793) +- fix: app log not being printed @namchuai (#1790) +- fix: api settings are not applied on changes @louis-jan (#1789) +- fix: could not delete model @louis-jan (#1779) +- fix: can not start model when server is not enabled from model settings page @louis-jan (#1774) +- regression fix: input port not accept alphabets @urmauur (#1772) +- Correct bash script syntax in ci @hiento09 (#1769) +- Hotfix CI pre-release not trigger @hiento09 (#1757) +- fix: bring back open app directory @louis-jan (#1756) +- fix: input port have range validation @urmauur (#1741) +- Fix error nightly build schedule run failed @hiento09 (#1736) +- fix: active model when start server @urmauur (#1719) +- fix: Change to fixed `localhost` instead of using host variable @hiro-v (#1729) +- Fix autoupdater nightly build error @hiento09 (#1727) +- Correct download url readme @hiento09 (#1724) +- fix: API chat/completion is blocked by CORS @louis-jan (#1705) +- fix: Jan server - v1/chat/completions is throwing ERR\_REQUIRE\_ESM @louis-jan (#1703) +- fix: Jan server is showing blank page @louis-jan (#1702) +- fix: switching loader from remote to local model from thread right panel @urmauur (#1692) +- fix: hot-fix algolia search @hieu-jan (#1700) +- fix: disable api key field while server is running @urmauur (#1694) +- fix: stoping model show starting model @urmauur (#1693) +- fix bug #1650 hogging resources @hiento09 (#1663) +- fix: auto select text when collapse panel @urmauur (#1645) +- fix: wrong selected model ref @louis-jan (#1638) +- fix: enable check for update on all supported platforms @louis-jan (#1626) +- fix: correct footer @hieu-jan (#1628) + +## 🧰 Maintenance + +- Docs publish to github page trigger on push to docs branch @hiento09 (#1783) +- Correct bash script syntax in ci @hiento09 (#1769) +- Combine 2 ci pipeline pre-release and nightly into one @hiento09 (#1767) +- Hotfix CI pre-release not trigger @hiento09 (#1757) +- Fix error nightly build schedule run failed @hiento09 (#1736) +- docs: add troubleshoot unexpected token @hieu-jan (#1711) +- docs: fix about pages @0xSage (#1699) +- refactor: deprecate extension type implementation @louis-jan (#1677) +- refactor: file prefix replace utils \& add unit test @louis-jan (#1676) +- Correct ref branch for update url on README.md file @hiento09 (#1672) +- docs: update 02-somethings-amiss @hieu-jan (#1668) +- Cherrypick cicd to main branch to apply new gitflow @hiento09 (#1665) +- docs: add user and developer guides for extensions @hieu-jan (#1657) +- docs: add QA Script @hieu-jan (#1660) +- chore: Bump nitro to 0.2.11 @hiro-v (#1655) +- chore: Bump version nitro to 0.2.10 @hiro-v (#1644) +- docs: add antivirus compatibility testing @hieu-jan (#1641) +- refactor: introduce node module in nitro extension @louis-jan (#1630) +- Update 02-somethings-amiss.mdx @Ssstars (#1634) +- docs: add integration AzureOpenAI @hieu-jan (#1632) +- docs: add troubleshooting permission denied @hieu-jan (#1631) + +## Contributor + +@0xSage, @Ssstars, @hiento09, @hientominh, @hieu-jan, @hiro-v, @jan-service-account, @louis-jan, @lucido-simon, @markmehere, @namchuai and @urmauur diff --git a/docs/docs/releases/changelog/changelog-v0.4.6.mdx b/docs/docs/releases/changelog/changelog-v0.4.6.mdx index 37bc2d493..58bbe258b 100644 --- a/docs/docs/releases/changelog/changelog-v0.4.6.mdx +++ b/docs/docs/releases/changelog/changelog-v0.4.6.mdx @@ -8,57 +8,57 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.4 Highlighted Issue: [Issue #1918: Regression fix assistant extension codesign](https://github.com/janhq/jan/pull/1918) -## Changes - -- Regression fix assitant extension codesign @hiento09 (#1918) -- Release cut 0.4.6 @louis-jan (#1888) -- feat: add factory reset feature @namchuai (#1750) -- chore: add react developer tools to electron @Helloyunho (#1858) -- Sync Release 0.4.5 to dev @louis-jan (#1830) - -## 🚀 Features - -- feat: integrate umami @hieu-jan (#1809) -- feat: Add default value for ngl @hiro-v (#1886) -- feat: add start/stop model via http api @namchuai (#1862) -- feat: add snackbar component and update style side banner @urmauur (#1874) -- feat: move open app directory into icon folder @urmauur (#1879) -- chore: Bump nitro to 0.3.3 @hiro-v (#1877) -- feat: put timestamp under thread name in left panel @urmauur (#1820) -- perf: remove unnecessary rerender when user typing input @namchuai (#1818) - -## 🐛 Fixes - -- fix: umami analytics send app loaded event @louis-jan (#1928) -- fix: migration loading indicator @louis-jan (#1913) -- fix: broken manual import model with NA fields @louis-jan (#1912) -- fix: openAIEmbedding now requires top level API Key configuration @louis-jan (#1902) -- fix: load model fail overlays thread message error @louis-jan (#1901) -- fix: show generate response on message send @louis-jan (#1895) -- fix: display error message on model load fail @louis-jan (#1894) -- fix: the selected model auto revert back to previous used model with setting mismatch @louis-jan (#1883) -- fix: add dialog confirm when move folder and next dest isn't empty @urmauur (#1880) -- Increase timeout for explore.e2e.spec test @hiento09 (#1844) -- chore: Bump nitro to 0.3.3 @hiro-v (#1877) -- fix: auto collapse retrieval setting while update config @urmauur (#1866) -- fix: loader show while error global when change folder @urmauur (#1870) -- fix: retrieval always ask for api key @louis-jan (#1856) -- fix: all input text box are disabled @namchuai (#1855) -- fix: add loader when user change folder @urmauur (#1850) -- Add code sign step for darwin assistant extension @hiento09 (#1841) -- fix: preserve focused thread when navigating in jan app @namchuai (#1814) -- fix: highlight menu dropdown server options @urmauur (#1831) - -## 🧰 Maintenance - -- chore: mark RAG as experimental feature @louis-jan (#1882) -- Increase timeout for explore.e2e.spec test @hiento09 (#1844) -- chore: Bump nitro to 0.3.3 @hiro-v (#1877) -- chore: Jan Data Folder setting is no longer an experimental feature @louis-jan (#1847) -- chore: resolve main conflict @louis-jan (#1833) -- Update release url on README to default branch instead of main branch @hiento09 (#1832) - -## Contributor - -@Helloyunho, @hiento09, @hieu-jan, @hiro-v, @jan-service-account, @louis-jan, @namchuai, @urmauur and James +## Changes + +- Regression fix assitant extension codesign @hiento09 (#1918) +- Release cut 0.4.6 @louis-jan (#1888) +- feat: add factory reset feature @namchuai (#1750) +- chore: add react developer tools to electron @Helloyunho (#1858) +- Sync Release 0.4.5 to dev @louis-jan (#1830) + +## 🚀 Features + +- feat: integrate umami @hieu-jan (#1809) +- feat: Add default value for ngl @hiro-v (#1886) +- feat: add start/stop model via http api @namchuai (#1862) +- feat: add snackbar component and update style side banner @urmauur (#1874) +- feat: move open app directory into icon folder @urmauur (#1879) +- chore: Bump nitro to 0.3.3 @hiro-v (#1877) +- feat: put timestamp under thread name in left panel @urmauur (#1820) +- perf: remove unnecessary rerender when user typing input @namchuai (#1818) + +## 🐛 Fixes + +- fix: umami analytics send app loaded event @louis-jan (#1928) +- fix: migration loading indicator @louis-jan (#1913) +- fix: broken manual import model with NA fields @louis-jan (#1912) +- fix: openAIEmbedding now requires top level API Key configuration @louis-jan (#1902) +- fix: load model fail overlays thread message error @louis-jan (#1901) +- fix: show generate response on message send @louis-jan (#1895) +- fix: display error message on model load fail @louis-jan (#1894) +- fix: the selected model auto revert back to previous used model with setting mismatch @louis-jan (#1883) +- fix: add dialog confirm when move folder and next dest isn't empty @urmauur (#1880) +- Increase timeout for explore.e2e.spec test @hiento09 (#1844) +- chore: Bump nitro to 0.3.3 @hiro-v (#1877) +- fix: auto collapse retrieval setting while update config @urmauur (#1866) +- fix: loader show while error global when change folder @urmauur (#1870) +- fix: retrieval always ask for api key @louis-jan (#1856) +- fix: all input text box are disabled @namchuai (#1855) +- fix: add loader when user change folder @urmauur (#1850) +- Add code sign step for darwin assistant extension @hiento09 (#1841) +- fix: preserve focused thread when navigating in jan app @namchuai (#1814) +- fix: highlight menu dropdown server options @urmauur (#1831) + +## 🧰 Maintenance + +- chore: mark RAG as experimental feature @louis-jan (#1882) +- Increase timeout for explore.e2e.spec test @hiento09 (#1844) +- chore: Bump nitro to 0.3.3 @hiro-v (#1877) +- chore: Jan Data Folder setting is no longer an experimental feature @louis-jan (#1847) +- chore: resolve main conflict @louis-jan (#1833) +- Update release url on README to default branch instead of main branch @hiento09 (#1832) + +## Contributor + +@Helloyunho, @hiento09, @hieu-jan, @hiro-v, @jan-service-account, @louis-jan, @namchuai, @urmauur and James diff --git a/docs/docs/releases/changelog/changelog-v0.4.7.mdx b/docs/docs/releases/changelog/changelog-v0.4.7.mdx index 7e7942096..57e5cfa16 100644 --- a/docs/docs/releases/changelog/changelog-v0.4.7.mdx +++ b/docs/docs/releases/changelog/changelog-v0.4.7.mdx @@ -8,111 +8,111 @@ For more details, [GitHub Issues](https://github.com/janhq/jan/releases/tag/v0.4 Highlighted Issue: [Issue #2121: Release cut v0.4.7](https://github.com/janhq/jan/pull/2121) -## Changes - -- Release cut v0.4.7 @louis-jan (#2121) -- chore: update models @hahuyhoang411 (#1829) -- add docs for entire advanced settings @hieu-jan (#2063) -- docs: Fix #2040 : added /v1 path to apiBase @ldebs (#2041) -- fix: ui for disabled state of gpu acceleration @namchuai (#2034) -- feat: Initialize POM structure with fixtures on Playwright @Van-QA (#2015) -- Alternative solution for `Thread titles should auto-summarize Topic` @0xgokuz (#1976) -- Update authors.yml Rex @hahuyhoang411 (#1956) -- Update authors.yml Louis @louis-jan (#1955) -- Change env Dockerfile.gpu and update README @hiento09 (#1963) -- chore: Update authors.yml for Van Pham @Van-QA (#1954) -- Sync dev branch to docs branch @hieu-jan (#1948) -- sync current docs branch to dev branch @hieu-jan (#1947) -- feat: Playwright capture screenshot of Electron desktop app (Jan) on failures @Van-QA (#1934) -- Sync main to dev after release 0.4.6 @hiento09 (#1929) - -## 🚀 Features - -- feat: Add nitro vulkan to support AMD GPU/ APU and Intel Arc GPU @hiro-v (#2056) -- fix: flow edit message @urmauur (#2113) -- Feature helmchart and ci jan server @hiento09 (#2106) -- feat: improvementUI GPU acceleration @urmauur (#1990) -- feat: add edit messages users @urmauur (#1974) -- feat: revamp ui dropdown list model option @urmauur (#1977) -- feat: add modal troubleshooting guideline @urmauur (#1968) -- feat: integrate umami script locally @hieu-jan (#1958) -- feat: User Selectable GPUs and GPU-based Model Recommendations @hiento09 (#1730) - -## 🐛 Fixes - -- fix: correct vulkan settings @louis-jan (#2128) -- fix: chore UI @louis-jan (#2125) -- Regression: bump nitro to 0.3.13 @hiento09 (#2124) -- Regression: Linux vulkan binary path @hiento09 (#2123) -- fix: revert back menu actions @louis-jan (#2120) -- fix: mismatching between nightly build and version - jan about @louis-jan (#2114) -- fix: flow edit message @urmauur (#2113) -- fix: tools section should be expanded by default @louis-jan (#2110) -- fix: failed to bind port - nitro error message copy @louis-jan (#2101) -- fix: remove caret down icon when tab selected into remote model @urmauur (#2102) -- fix: openai client sdk compatible @louis-jan (#2096) -- Fix bug #2005 docker blank website @hiento09 (#2093) -- fix: check if port is occupied before start local server @namchuai (#2098) -- fix: broken model.json update @louis-jan (#2099) -- fix: make text input scrollable @urmauur (#2083) -- fix: failed to send message blocks thread creation @louis-jan (#2091) -- fix: server crashes on missing module @louis-jan (#2089) -- fix: expand assistant and model settings by default @louis-jan (#2081) -- fix: move jan data folder - error handling for no write permission granted @louis-jan (#2077) -- fix: check for updates should show no update are available on the latest build @louis-jan (#2075) -- fix: infinity showed when haven't get total size @namchuai (#2066) -- fix: should stop running the model when GPU settings are changed @louis-jan (#2067) -- fix: settings page state loop and dark theme @louis-jan (#2065) -- fix: Fix Nitro windows with error 3221225781 @hiro-v (#2057) -- fix: message should only be interrupted when i start another thread @louis-jan (#2053) -- fix: local server start error should not change to started state @louis-jan (#2052) -- fix: update copy of message queue @louis-jan (#2051) -- fix: download mutilple binaries @namchuai (#2043) -- fix: disable gpu drop down box if there's no GPU ready @namchuai (#2046) -- fix: app should generate thread title with length restriction @louis-jan (#2037) -- fix: factory reset not remove jan data folder @namchuai (#2027) -- fix: content setting right panel default to collapse @urmauur (#2026) -- fix: local server blank parameters if there is no thread selected @louis-jan (#2028) -- fix: model path backward compatible @louis-jan (#2018) -- fix: resolve state update loop infinitive rerendering @louis-jan (#2017) -- fix: lack of auto-cleaning mechanism for logs @louis-jan (#2003) -- fix: app stuck regenerating assistant response @louis-jan (#2001) -- fix: decouple thread summary update @louis-jan (#1994) -- fix: app fails gracefully with clear error messages @louis-jan (#1993) -- fix: retrieval stuck at generating response @louis-jan (#1988) -- Fix macos auto update failed on nightly build @hiento09 (#1991) -- fix: model downloads broken on nightly @louis-jan (#1984) -- fix: RAG enhancements @urmauur (#1965) -- Update docs run Jan Server in Docker mode @hiento09 (#1960) -- fix: update conditional check last status message @urmauur (#1951) -- fix: markdown render for chat completion role user @urmauur (#1944) -- fix: avoid users to create so many threads at the same time @urmauur (#1930) -- fix: download model will close panel item hub @urmauur (#1923) - -## 🧰 Maintenance - -- docs: improve integrations guide \& import model using absolute path @hieu-jan (#2076) -- chore: add app version into log @namchuai (#2116) -- docs: add integration docs Mistral AI API @hieu-jan (#2070) -- docs:add-advanced-settings-https-proxy @hieu-jan (#2054) -- chore: refactor watch system resource hook @louis-jan (#2048) -- docs: Updates Guide Using the Local Server @SamPatt (#1924) -- server install core using link instead of file @hiento09 (#2025) -- chore: prettier fix @louis-jan (#2019) -- chore: bump nitro 0.3.9 @louis-jan (#2016) -- refactor: reduce IPC \& API handlers - shared node logics @louis-jan (#2011) -- docs: update 03-gpu-not-used with RTX issues @hieu-jan (#1992) -- docs: add Jan installation using Docker @hieu-jan (#1981) -- chore: reduce bundle size @louis-jan (#1970) -- docs: add author.yml @hieu-jan (#1973) -- Update authors.yml hien @hiento09 (#1953) -- chore: server download progress + S3 @louis-jan (#1925) -- chore: add author james @namchuai (#1952) -- chore: Add author - Ashley @imtuyethan (#1950) -- chore: Add Author - Hiro @hiro-v (#1949) -- docs: adding new feature for v0.4.6 to release checklist @Van-QA (#1927) - -## Contributor - -@0xSage, @0xgokuz, @SamPatt, @Van-QA, @hahuyhoang411, @hiento09, @hieu-jan, @hiro-v, @imtuyethan, @jan-service-account, @ldebs, @louis-jan, @namchuai, @urmauur and James +## Changes + +- Release cut v0.4.7 @louis-jan (#2121) +- chore: update models @hahuyhoang411 (#1829) +- add docs for entire advanced settings @hieu-jan (#2063) +- docs: Fix #2040 : added /v1 path to apiBase @ldebs (#2041) +- fix: ui for disabled state of gpu acceleration @namchuai (#2034) +- feat: Initialize POM structure with fixtures on Playwright @Van-QA (#2015) +- Alternative solution for `Thread titles should auto-summarize Topic` @0xgokuz (#1976) +- Update authors.yml Rex @hahuyhoang411 (#1956) +- Update authors.yml Louis @louis-jan (#1955) +- Change env Dockerfile.gpu and update README @hiento09 (#1963) +- chore: Update authors.yml for Van Pham @Van-QA (#1954) +- Sync dev branch to docs branch @hieu-jan (#1948) +- sync current docs branch to dev branch @hieu-jan (#1947) +- feat: Playwright capture screenshot of Electron desktop app (Jan) on failures @Van-QA (#1934) +- Sync main to dev after release 0.4.6 @hiento09 (#1929) + +## 🚀 Features + +- feat: Add nitro vulkan to support AMD GPU/ APU and Intel Arc GPU @hiro-v (#2056) +- fix: flow edit message @urmauur (#2113) +- Feature helmchart and ci jan server @hiento09 (#2106) +- feat: improvementUI GPU acceleration @urmauur (#1990) +- feat: add edit messages users @urmauur (#1974) +- feat: revamp ui dropdown list model option @urmauur (#1977) +- feat: add modal troubleshooting guideline @urmauur (#1968) +- feat: integrate umami script locally @hieu-jan (#1958) +- feat: User Selectable GPUs and GPU-based Model Recommendations @hiento09 (#1730) + +## 🐛 Fixes + +- fix: correct vulkan settings @louis-jan (#2128) +- fix: chore UI @louis-jan (#2125) +- Regression: bump nitro to 0.3.13 @hiento09 (#2124) +- Regression: Linux vulkan binary path @hiento09 (#2123) +- fix: revert back menu actions @louis-jan (#2120) +- fix: mismatching between nightly build and version - jan about @louis-jan (#2114) +- fix: flow edit message @urmauur (#2113) +- fix: tools section should be expanded by default @louis-jan (#2110) +- fix: failed to bind port - nitro error message copy @louis-jan (#2101) +- fix: remove caret down icon when tab selected into remote model @urmauur (#2102) +- fix: openai client sdk compatible @louis-jan (#2096) +- Fix bug #2005 docker blank website @hiento09 (#2093) +- fix: check if port is occupied before start local server @namchuai (#2098) +- fix: broken model.json update @louis-jan (#2099) +- fix: make text input scrollable @urmauur (#2083) +- fix: failed to send message blocks thread creation @louis-jan (#2091) +- fix: server crashes on missing module @louis-jan (#2089) +- fix: expand assistant and model settings by default @louis-jan (#2081) +- fix: move jan data folder - error handling for no write permission granted @louis-jan (#2077) +- fix: check for updates should show no update are available on the latest build @louis-jan (#2075) +- fix: infinity showed when haven't get total size @namchuai (#2066) +- fix: should stop running the model when GPU settings are changed @louis-jan (#2067) +- fix: settings page state loop and dark theme @louis-jan (#2065) +- fix: Fix Nitro windows with error 3221225781 @hiro-v (#2057) +- fix: message should only be interrupted when i start another thread @louis-jan (#2053) +- fix: local server start error should not change to started state @louis-jan (#2052) +- fix: update copy of message queue @louis-jan (#2051) +- fix: download mutilple binaries @namchuai (#2043) +- fix: disable gpu drop down box if there's no GPU ready @namchuai (#2046) +- fix: app should generate thread title with length restriction @louis-jan (#2037) +- fix: factory reset not remove jan data folder @namchuai (#2027) +- fix: content setting right panel default to collapse @urmauur (#2026) +- fix: local server blank parameters if there is no thread selected @louis-jan (#2028) +- fix: model path backward compatible @louis-jan (#2018) +- fix: resolve state update loop infinitive rerendering @louis-jan (#2017) +- fix: lack of auto-cleaning mechanism for logs @louis-jan (#2003) +- fix: app stuck regenerating assistant response @louis-jan (#2001) +- fix: decouple thread summary update @louis-jan (#1994) +- fix: app fails gracefully with clear error messages @louis-jan (#1993) +- fix: retrieval stuck at generating response @louis-jan (#1988) +- Fix macos auto update failed on nightly build @hiento09 (#1991) +- fix: model downloads broken on nightly @louis-jan (#1984) +- fix: RAG enhancements @urmauur (#1965) +- Update docs run Jan Server in Docker mode @hiento09 (#1960) +- fix: update conditional check last status message @urmauur (#1951) +- fix: markdown render for chat completion role user @urmauur (#1944) +- fix: avoid users to create so many threads at the same time @urmauur (#1930) +- fix: download model will close panel item hub @urmauur (#1923) + +## 🧰 Maintenance + +- docs: improve integrations guide \& import model using absolute path @hieu-jan (#2076) +- chore: add app version into log @namchuai (#2116) +- docs: add integration docs Mistral AI API @hieu-jan (#2070) +- docs:add-advanced-settings-https-proxy @hieu-jan (#2054) +- chore: refactor watch system resource hook @louis-jan (#2048) +- docs: Updates Guide Using the Local Server @SamPatt (#1924) +- server install core using link instead of file @hiento09 (#2025) +- chore: prettier fix @louis-jan (#2019) +- chore: bump nitro 0.3.9 @louis-jan (#2016) +- refactor: reduce IPC \& API handlers - shared node logics @louis-jan (#2011) +- docs: update 03-gpu-not-used with RTX issues @hieu-jan (#1992) +- docs: add Jan installation using Docker @hieu-jan (#1981) +- chore: reduce bundle size @louis-jan (#1970) +- docs: add author.yml @hieu-jan (#1973) +- Update authors.yml hien @hiento09 (#1953) +- chore: server download progress + S3 @louis-jan (#1925) +- chore: add author james @namchuai (#1952) +- chore: Add author - Ashley @imtuyethan (#1950) +- chore: Add Author - Hiro @hiro-v (#1949) +- docs: adding new feature for v0.4.6 to release checklist @Van-QA (#1927) + +## Contributor + +@0xSage, @0xgokuz, @SamPatt, @Van-QA, @hahuyhoang411, @hiento09, @hieu-jan, @hiro-v, @imtuyethan, @jan-service-account, @ldebs, @louis-jan, @namchuai, @urmauur and James From 86af902d19e0b91068a401002fe9103b08809626 Mon Sep 17 00:00:00 2001 From: NamH Date: Mon, 11 Mar 2024 21:24:31 +0700 Subject: [PATCH 07/37] fix: message from quick ask not get the selected model (#2307) Signed-off-by: James Co-authored-by: James --- web/hooks/useSendChatMessage.ts | 19 ++++++++++++++----- 1 file changed, 14 insertions(+), 5 deletions(-) diff --git a/web/hooks/useSendChatMessage.ts b/web/hooks/useSendChatMessage.ts index 9e88e763a..11a57a598 100644 --- a/web/hooks/useSendChatMessage.ts +++ b/web/hooks/useSendChatMessage.ts @@ -79,6 +79,8 @@ export default function useSendChatMessage() { const setIsGeneratingResponse = useSetAtom(isGeneratingResponseAtom) const activeThreadRef = useRef() + const selectedModelRef = useRef() + useEffect(() => { modelRef.current = activeModel }, [activeModel]) @@ -91,6 +93,10 @@ export default function useSendChatMessage() { activeThreadRef.current = activeThread }, [activeThread]) + useEffect(() => { + selectedModelRef.current = selectedModel + }, [selectedModel]) + const resendChatMessage = async (currentMessage: ThreadMessage) => { if (!activeThreadRef.current) { console.error('No active thread') @@ -128,11 +134,13 @@ export default function useSendChatMessage() { type: MessageRequestType.Thread, messages: messages, threadId: activeThreadRef.current.id, - model: activeThreadRef.current.assistants[0].model ?? selectedModel, + model: + activeThreadRef.current.assistants[0].model ?? selectedModelRef.current, } const modelId = - selectedModel?.id ?? activeThreadRef.current.assistants[0].model.id + selectedModelRef.current?.id ?? + activeThreadRef.current.assistants[0].model.id if (modelRef.current?.id !== modelId) { setQueuedMessage(true) @@ -213,7 +221,7 @@ export default function useSendChatMessage() { { role: ChatCompletionRole.User, content: - selectedModel && base64Blob + selectedModelRef.current && base64Blob ? [ { type: ChatCompletionMessageContentType.Text, @@ -242,7 +250,7 @@ export default function useSendChatMessage() { ) let modelRequest = - selectedModel ?? activeThreadRef.current.assistants[0].model + selectedModelRef?.current ?? activeThreadRef.current.assistants[0].model if (runtimeParams.stream == null) { runtimeParams.stream = true } @@ -344,7 +352,8 @@ export default function useSendChatMessage() { ?.addNewMessage(threadMessage) const modelId = - selectedModel?.id ?? activeThreadRef.current.assistants[0].model.id + selectedModelRef.current?.id ?? + activeThreadRef.current.assistants[0].model.id if (modelRef.current?.id !== modelId) { setQueuedMessage(true) From c8ea16e661bc4ae74e50e4b1bbd01a9e7961b06f Mon Sep 17 00:00:00 2001 From: Louis Date: Mon, 11 Mar 2024 22:58:18 +0700 Subject: [PATCH 08/37] fix: quick ask blocks app update (#2310) --- electron/handlers/update.ts | 2 ++ electron/main.ts | 31 +++++++++++++++++++++--------- electron/managers/tray.ts | 38 +++++++++++++++++++++++++++++++++++++ electron/utils/tray.ts | 24 ----------------------- 4 files changed, 62 insertions(+), 33 deletions(-) create mode 100644 electron/managers/tray.ts delete mode 100644 electron/utils/tray.ts diff --git a/electron/handlers/update.ts b/electron/handlers/update.ts index 3f52c401e..5ea261e54 100644 --- a/electron/handlers/update.ts +++ b/electron/handlers/update.ts @@ -7,6 +7,7 @@ import { autoUpdater, } from 'electron-updater' import { AppEvent } from '@janhq/core' +import { trayManager } from '../managers/tray' export let waitingToInstallVersion: string | undefined = undefined @@ -22,6 +23,7 @@ export function handleAppUpdates() { message: 'Would you like to download and install it now?', buttons: ['Download', 'Later'], }) + trayManager.destroyCurrentTray() if (action.response === 0) await autoUpdater.downloadUpdate() }) diff --git a/electron/main.ts b/electron/main.ts index 8fe4247d6..e72c7f95a 100644 --- a/electron/main.ts +++ b/electron/main.ts @@ -1,4 +1,4 @@ -import { app, BrowserWindow, Menu, Tray } from 'electron' +import { app, BrowserWindow, Tray } from 'electron' import { join } from 'path' /** @@ -27,7 +27,7 @@ import { setupReactDevTool } from './utils/dev' import { cleanLogs } from './utils/log' import { registerShortcut } from './utils/selectedText' -import { createSystemTray } from './utils/tray' +import { trayManager } from './managers/tray' const preloadPath = join(__dirname, 'preload.js') const rendererPath = join(__dirname, '..', 'renderer') @@ -39,6 +39,8 @@ const quickAskUrl = `${mainUrl}/search` const quickAskHotKey = 'CommandOrControl+J' +const gotTheLock = app.requestSingleInstanceLock() + app .whenReady() .then(setupReactDevTool) @@ -56,11 +58,23 @@ app windowManager.mainWindow?.webContents.openDevTools() } }) - .then(() => process.env.CI !== 'e2e' && createSystemTray()) + .then(() => process.env.CI !== 'e2e' && trayManager.createSystemTray()) .then(() => { log(`Version: ${app.getVersion()}`) }) .then(() => { + if (!gotTheLock) { + app.quit() + } else { + app.on('second-instance', (_event, _commandLine, _workingDirectory) => { + // Someone tried to run a second instance, we should focus our window. + if (windowManager.mainWindow) { + if (windowManager.mainWindow.isMinimized()) + windowManager.mainWindow.restore() + windowManager.mainWindow.focus() + } + }) + } app.on('activate', () => { if (!BrowserWindow.getAllWindows().length) { createMainWindow() @@ -73,6 +87,10 @@ app.on('ready', () => { registerGlobalShortcuts() }) +app.on('before-quit', function (evt) { + trayManager.destroyCurrentTray() +}) + app.once('quit', () => { cleanUpAndQuit() }) @@ -89,12 +107,7 @@ function createMainWindow() { function registerGlobalShortcuts() { const ret = registerShortcut(quickAskHotKey, (selectedText: string) => { - if (!windowManager.isQuickAskWindowVisible()) { - windowManager.showQuickAskWindow() - windowManager.sendQuickAskSelectedText(selectedText) - } else { - windowManager.hideQuickAskWindow() - } + windowManager.showMainWindow() }) if (!ret) { diff --git a/electron/managers/tray.ts b/electron/managers/tray.ts new file mode 100644 index 000000000..18661e58e --- /dev/null +++ b/electron/managers/tray.ts @@ -0,0 +1,38 @@ +import { join } from 'path' +import { Tray, app, Menu } from 'electron' +import { windowManager } from '../managers/window' + +class TrayManager { + currentTray: Tray | undefined + + createSystemTray = () => { + if (this.currentTray) { + return + } + const iconPath = join(app.getAppPath(), 'icons', 'icon-tray.png') + const tray = new Tray(iconPath) + tray.setToolTip(app.getName()) + + const contextMenu = Menu.buildFromTemplate([ + { + label: 'Open Jan', + type: 'normal', + click: () => windowManager.showMainWindow(), + }, + { + label: 'Open Quick Ask', + type: 'normal', + click: () => windowManager.showQuickAskWindow(), + }, + { label: 'Quit', type: 'normal', click: () => app.quit() }, + ]) + tray.setContextMenu(contextMenu) + } + + destroyCurrentTray() { + this.currentTray?.destroy() + this.currentTray = undefined + } +} + +export const trayManager = new TrayManager() diff --git a/electron/utils/tray.ts b/electron/utils/tray.ts deleted file mode 100644 index 2ab3e6dcc..000000000 --- a/electron/utils/tray.ts +++ /dev/null @@ -1,24 +0,0 @@ -import { join } from 'path' -import { Tray, app, Menu } from 'electron' -import { windowManager } from '../managers/window' - -export const createSystemTray = () => { - const iconPath = join(app.getAppPath(), 'icons', 'icon-tray.png') - const tray = new Tray(iconPath) - tray.setToolTip(app.getName()) - - const contextMenu = Menu.buildFromTemplate([ - { - label: 'Open Jan', - type: 'normal', - click: () => windowManager.showMainWindow(), - }, - { - label: 'Open Quick Ask', - type: 'normal', - click: () => windowManager.showQuickAskWindow(), - }, - { label: 'Quit', type: 'normal', click: () => app.quit() }, - ]) - tray.setContextMenu(contextMenu) -} From bbfc686c33334e0b619ed07f80084293aed3aa4a Mon Sep 17 00:00:00 2001 From: Louis Date: Mon, 11 Mar 2024 23:29:39 +0700 Subject: [PATCH 09/37] fix: quick ask not show (#2315) * fix: quick ask not show * fix: window resize does not work on windows 11 --- electron/main.ts | 13 +++++++------ electron/managers/window.ts | 1 + 2 files changed, 8 insertions(+), 6 deletions(-) diff --git a/electron/main.ts b/electron/main.ts index e72c7f95a..78577ac68 100644 --- a/electron/main.ts +++ b/electron/main.ts @@ -68,11 +68,7 @@ app } else { app.on('second-instance', (_event, _commandLine, _workingDirectory) => { // Someone tried to run a second instance, we should focus our window. - if (windowManager.mainWindow) { - if (windowManager.mainWindow.isMinimized()) - windowManager.mainWindow.restore() - windowManager.mainWindow.focus() - } + windowManager.showMainWindow() }) } app.on('activate', () => { @@ -107,7 +103,12 @@ function createMainWindow() { function registerGlobalShortcuts() { const ret = registerShortcut(quickAskHotKey, (selectedText: string) => { - windowManager.showMainWindow() + if (!windowManager.isQuickAskWindowVisible()) { + windowManager.showQuickAskWindow() + windowManager.sendQuickAskSelectedText(selectedText) + } else { + windowManager.hideQuickAskWindow() + } }) if (!ret) { diff --git a/electron/managers/window.ts b/electron/managers/window.ts index 796a5d54a..eed80c37c 100644 --- a/electron/managers/window.ts +++ b/electron/managers/window.ts @@ -101,6 +101,7 @@ class WindowManager { expandQuickAskWindow(heightOffset: number): void { const width = quickAskWindowConfig.width! const height = quickAskWindowConfig.height! + heightOffset + this._quickAskWindow?.setMinimumSize(width, height) this._quickAskWindow?.setSize(width, height, true) } From 6f899eb1186860a07bf8880a112473cbc40eb760 Mon Sep 17 00:00:00 2001 From: Service Account Date: Mon, 11 Mar 2024 17:43:46 +0000 Subject: [PATCH 10/37] janhq/jan: Update README.md with nightly build artifact URL --- README.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index f847d63d2..ef9516b81 100644 --- a/README.md +++ b/README.md @@ -76,31 +76,31 @@ Jan is an open-source ChatGPT alternative that runs 100% offline on your compute Experimental (Nightly Build) - + jan.exe - + Intel - + M1/M2 - + jan.deb - + jan.AppImage From de433bbe6e76c97c5aa98ecc533b253ce2f6bb5e Mon Sep 17 00:00:00 2001 From: Service Account Date: Mon, 11 Mar 2024 20:20:46 +0000 Subject: [PATCH 11/37] janhq/jan: Update README.md with nightly build artifact URL --- README.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index ef9516b81..496bbd434 100644 --- a/README.md +++ b/README.md @@ -76,31 +76,31 @@ Jan is an open-source ChatGPT alternative that runs 100% offline on your compute Experimental (Nightly Build) - + jan.exe - + Intel - + M1/M2 - + jan.deb - + jan.AppImage From ad7e483df606ac4230279b90db9dddd6e16e5e9c Mon Sep 17 00:00:00 2001 From: Louis Date: Tue, 12 Mar 2024 09:21:28 +0700 Subject: [PATCH 12/37] fix: jan app tray blocks app update (#2319) --- electron/managers/tray.ts | 1 + 1 file changed, 1 insertion(+) diff --git a/electron/managers/tray.ts b/electron/managers/tray.ts index 18661e58e..4e1c1a4ff 100644 --- a/electron/managers/tray.ts +++ b/electron/managers/tray.ts @@ -27,6 +27,7 @@ class TrayManager { { label: 'Quit', type: 'normal', click: () => app.quit() }, ]) tray.setContextMenu(contextMenu) + this.currentTray = tray } destroyCurrentTray() { From 259698eaf720f99125cfed3a102018b55822dc51 Mon Sep 17 00:00:00 2001 From: Service Account Date: Tue, 12 Mar 2024 02:36:22 +0000 Subject: [PATCH 13/37] janhq/jan: Update README.md with nightly build artifact URL --- README.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index 496bbd434..107a35d29 100644 --- a/README.md +++ b/README.md @@ -76,31 +76,31 @@ Jan is an open-source ChatGPT alternative that runs 100% offline on your compute Experimental (Nightly Build) - + jan.exe - + Intel - + M1/M2 - + jan.deb - + jan.AppImage From facfaa190a285cccabc4200c097ee5116dabd31f Mon Sep 17 00:00:00 2001 From: Faisal Amir Date: Tue, 12 Mar 2024 10:03:00 +0700 Subject: [PATCH 14/37] feat: quick ask support darkmode (#2316) --- web/app/search/SelectedText.tsx | 8 ++++---- web/app/search/UserInput.tsx | 5 ++--- web/app/search/page.tsx | 2 +- 3 files changed, 7 insertions(+), 8 deletions(-) diff --git a/web/app/search/SelectedText.tsx b/web/app/search/SelectedText.tsx index 742eba956..2bb90775e 100644 --- a/web/app/search/SelectedText.tsx +++ b/web/app/search/SelectedText.tsx @@ -29,15 +29,15 @@ const SelectedText = ({ onCleared }: { onCleared?: () => void }) => { return shouldShowSelectedText ? (
- +
-

{text}

+

{text}

) : (
diff --git a/web/app/search/UserInput.tsx b/web/app/search/UserInput.tsx index 3bf148f40..85cb9843e 100644 --- a/web/app/search/UserInput.tsx +++ b/web/app/search/UserInput.tsx @@ -56,7 +56,7 @@ const UserInput: React.FC = () => { } return ( -
+
{ {
- inputRef?.current?.focus()} />
) diff --git a/web/app/search/page.tsx b/web/app/search/page.tsx index 0822c2676..912a843bb 100644 --- a/web/app/search/page.tsx +++ b/web/app/search/page.tsx @@ -4,7 +4,7 @@ import UserInput from './UserInput' const Search: React.FC = () => { return ( -
+
) From b9b421a495876a58465710da872425a8e045dd57 Mon Sep 17 00:00:00 2001 From: Louis Date: Tue, 12 Mar 2024 17:25:50 +0700 Subject: [PATCH 15/37] fix: quick app bugs (#2327) --- electron/main.ts | 20 +++++++------ electron/managers/mainWindowConfig.ts | 2 +- electron/managers/tray.ts | 36 +++++++++++++++--------- electron/managers/window.ts | 4 --- web/containers/Providers/KeyListener.tsx | 4 --- 5 files changed, 35 insertions(+), 31 deletions(-) diff --git a/electron/main.ts b/electron/main.ts index 78577ac68..b7ac0a74c 100644 --- a/electron/main.ts +++ b/electron/main.ts @@ -43,6 +43,12 @@ const gotTheLock = app.requestSingleInstanceLock() app .whenReady() + .then(() => { + if (!gotTheLock) { + app.quit() + throw new Error('Another instance of the app is already running') + } + }) .then(setupReactDevTool) .then(setupCore) .then(createUserSpace) @@ -63,22 +69,20 @@ app log(`Version: ${app.getVersion()}`) }) .then(() => { - if (!gotTheLock) { - app.quit() - } else { - app.on('second-instance', (_event, _commandLine, _workingDirectory) => { - // Someone tried to run a second instance, we should focus our window. - windowManager.showMainWindow() - }) - } app.on('activate', () => { if (!BrowserWindow.getAllWindows().length) { createMainWindow() + } else { + windowManager.showMainWindow() } }) }) .then(() => cleanLogs()) +app.on('second-instance', (_event, _commandLine, _workingDirectory) => { + windowManager.showMainWindow() +}) + app.on('ready', () => { registerGlobalShortcuts() }) diff --git a/electron/managers/mainWindowConfig.ts b/electron/managers/mainWindowConfig.ts index 184fb1c86..4f1715a94 100644 --- a/electron/managers/mainWindowConfig.ts +++ b/electron/managers/mainWindowConfig.ts @@ -5,7 +5,7 @@ export const mainWindowConfig: Electron.BrowserWindowConstructorOptions = { width: DEFAULT_WIDTH, minWidth: DEFAULT_WIDTH, height: DEFAULT_HEIGHT, - skipTaskbar: true, + skipTaskbar: false, show: true, trafficLightPosition: { x: 10, diff --git a/electron/managers/tray.ts b/electron/managers/tray.ts index 4e1c1a4ff..e7e07a0a1 100644 --- a/electron/managers/tray.ts +++ b/electron/managers/tray.ts @@ -13,20 +13,28 @@ class TrayManager { const tray = new Tray(iconPath) tray.setToolTip(app.getName()) - const contextMenu = Menu.buildFromTemplate([ - { - label: 'Open Jan', - type: 'normal', - click: () => windowManager.showMainWindow(), - }, - { - label: 'Open Quick Ask', - type: 'normal', - click: () => windowManager.showQuickAskWindow(), - }, - { label: 'Quit', type: 'normal', click: () => app.quit() }, - ]) - tray.setContextMenu(contextMenu) + tray.on('click', () => { + windowManager.showQuickAskWindow() + }) + + // Add context menu for windows only + if (process.platform === 'win32') { + const contextMenu = Menu.buildFromTemplate([ + { + label: 'Open Jan', + type: 'normal', + click: () => windowManager.showMainWindow(), + }, + { + label: 'Open Quick Ask', + type: 'normal', + click: () => windowManager.showQuickAskWindow(), + }, + { label: 'Quit', type: 'normal', click: () => app.quit() }, + ]) + + tray.setContextMenu(contextMenu) + } this.currentTray = tray } diff --git a/electron/managers/window.ts b/electron/managers/window.ts index eed80c37c..7b97f8349 100644 --- a/electron/managers/window.ts +++ b/electron/managers/window.ts @@ -73,15 +73,11 @@ class WindowManager { hideMainWindow(): void { this.mainWindow?.hide() this._mainWindowVisible = false - // Only macos - if (process.platform === 'darwin') app.dock.hide() } showMainWindow(): void { this.mainWindow?.show() this._mainWindowVisible = true - // Only macos - if (process.platform === 'darwin') app.dock.show() } hideQuickAskWindow(): void { diff --git a/web/containers/Providers/KeyListener.tsx b/web/containers/Providers/KeyListener.tsx index d832059c2..a4702783c 100644 --- a/web/containers/Providers/KeyListener.tsx +++ b/web/containers/Providers/KeyListener.tsx @@ -24,10 +24,6 @@ export default function KeyListener({ children }: Props) { useEffect(() => { const onKeyDown = (e: KeyboardEvent) => { - if (e.key === 'Escape') { - window.core?.api?.hideMainWindow() - } - const prefixKey = isMac ? e.metaKey : e.ctrlKey if (e.key === 'b' && prefixKey) { From 32b76499d973e67586b1aad460b4577a0a9c50fb Mon Sep 17 00:00:00 2001 From: hieu-jan <150573299+hieu-jan@users.noreply.github.com> Date: Tue, 12 Mar 2024 20:54:59 +0900 Subject: [PATCH 16/37] fix: existing-changelog --- docs/plugins/changelog-plugin/fetchData.js | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/docs/plugins/changelog-plugin/fetchData.js b/docs/plugins/changelog-plugin/fetchData.js index 351ab3932..a9b970b3a 100644 --- a/docs/plugins/changelog-plugin/fetchData.js +++ b/docs/plugins/changelog-plugin/fetchData.js @@ -1,5 +1,6 @@ const fs = require('fs'); const path = require('path'); +const fetch = require('node-fetch'); async function fetchData(siteConfig) { const owner = siteConfig.organizationName; @@ -70,6 +71,14 @@ async function fetchData(siteConfig) { // Process the GitHub releases data here for (const release of releases) { const version = release.tag_name; + + // Check if the changelog file already exists for the current version + const existingChangelogPath = path.join(outputDirectory, `changelog-${version}.mdx`); + if (fs.existsSync(existingChangelogPath)) { + console.log(`Changelog for version ${version} already exists. Skipping...`); + continue; + } + const releaseUrl = release.html_url; const issueNumberMatch = release.body.match(/#(\d+)/); const issueNumber = issueNumberMatch ? parseInt(issueNumberMatch[1], 10) : null; @@ -94,4 +103,4 @@ async function fetchData(siteConfig) { } } -module.exports = fetchData; \ No newline at end of file +module.exports = fetchData; From af5bcea773e0f0e574663fcf1f399f3e961f6a63 Mon Sep 17 00:00:00 2001 From: Louis Date: Tue, 12 Mar 2024 20:10:59 +0700 Subject: [PATCH 17/37] fix: gate quick ask with feature toggle (#2331) --- core/src/node/helper/config.ts | 2 +- core/src/types/config/appConfigEntity.ts | 1 + electron/main.ts | 13 ++++++- electron/managers/tray.ts | 4 +++ electron/managers/window.ts | 4 +++ web/containers/Providers/DataLoader.tsx | 9 +++-- web/helpers/atoms/AppConfig.atom.ts | 2 ++ web/hooks/useFactoryReset.ts | 1 + web/screens/Settings/Advanced/index.tsx | 45 ++++++++++++++++++++++-- 9 files changed, 74 insertions(+), 7 deletions(-) diff --git a/core/src/node/helper/config.ts b/core/src/node/helper/config.ts index 71e721578..06f2b03cd 100644 --- a/core/src/node/helper/config.ts +++ b/core/src/node/helper/config.ts @@ -4,13 +4,13 @@ import fs from 'fs' import os from 'os' import childProcess from 'child_process' -// TODO: move this to core const configurationFileName = 'settings.json' // TODO: do no specify app name in framework module const defaultJanDataFolder = join(os.homedir(), 'jan') const defaultAppConfig: AppConfiguration = { data_folder: defaultJanDataFolder, + quick_ask: false, } /** diff --git a/core/src/types/config/appConfigEntity.ts b/core/src/types/config/appConfigEntity.ts index 81ea0b30f..1402aeca1 100644 --- a/core/src/types/config/appConfigEntity.ts +++ b/core/src/types/config/appConfigEntity.ts @@ -1,3 +1,4 @@ export type AppConfiguration = { data_folder: string + quick_ask: boolean } diff --git a/electron/main.ts b/electron/main.ts index b7ac0a74c..70314b169 100644 --- a/electron/main.ts +++ b/electron/main.ts @@ -5,7 +5,7 @@ import { join } from 'path' * Managers **/ import { windowManager } from './managers/window' -import { log } from '@janhq/core/node' +import { getAppConfigurations, log } from '@janhq/core/node' /** * IPC Handlers @@ -95,7 +95,15 @@ app.once('quit', () => { cleanUpAndQuit() }) +app.once('window-all-closed', () => { + // Feature Toggle for Quick Ask + if (getAppConfigurations().quick_ask) return + cleanUpAndQuit() +}) + function createQuickAskWindow() { + // Feature Toggle for Quick Ask + if (!getAppConfigurations().quick_ask) return const startUrl = app.isPackaged ? `file://${quickAskPath}` : quickAskUrl windowManager.createQuickAskWindow(preloadPath, startUrl) } @@ -107,6 +115,9 @@ function createMainWindow() { function registerGlobalShortcuts() { const ret = registerShortcut(quickAskHotKey, (selectedText: string) => { + // Feature Toggle for Quick Ask + if (!getAppConfigurations().quick_ask) return + if (!windowManager.isQuickAskWindowVisible()) { windowManager.showQuickAskWindow() windowManager.sendQuickAskSelectedText(selectedText) diff --git a/electron/managers/tray.ts b/electron/managers/tray.ts index e7e07a0a1..b81b1e556 100644 --- a/electron/managers/tray.ts +++ b/electron/managers/tray.ts @@ -1,11 +1,15 @@ import { join } from 'path' import { Tray, app, Menu } from 'electron' import { windowManager } from '../managers/window' +import { getAppConfigurations } from '@janhq/core/node' class TrayManager { currentTray: Tray | undefined createSystemTray = () => { + // Feature Toggle for Quick Ask + if (!getAppConfigurations().quick_ask) return + if (this.currentTray) { return } diff --git a/electron/managers/window.ts b/electron/managers/window.ts index 7b97f8349..da8dd4b17 100644 --- a/electron/managers/window.ts +++ b/electron/managers/window.ts @@ -2,6 +2,7 @@ import { BrowserWindow, app, shell } from 'electron' import { quickAskWindowConfig } from './quickAskWindowConfig' import { AppEvent } from '@janhq/core' import { mainWindowConfig } from './mainWindowConfig' +import { getAppConfigurations } from '@janhq/core/node' /** * Manages the current window instance. @@ -43,6 +44,9 @@ class WindowManager { }) windowManager.mainWindow?.on('close', function (evt) { + // Feature Toggle for Quick Ask + if (!getAppConfigurations().quick_ask) return + if (!isAppQuitting) { evt.preventDefault() windowManager.hideMainWindow() diff --git a/web/containers/Providers/DataLoader.tsx b/web/containers/Providers/DataLoader.tsx index bc1461d5b..7d38a29d6 100644 --- a/web/containers/Providers/DataLoader.tsx +++ b/web/containers/Providers/DataLoader.tsx @@ -10,7 +10,10 @@ import useGetSystemResources from '@/hooks/useGetSystemResources' import useModels from '@/hooks/useModels' import useThreads from '@/hooks/useThreads' -import { janDataFolderPathAtom } from '@/helpers/atoms/AppConfig.atom' +import { + janDataFolderPathAtom, + quickAskEnabledAtom, +} from '@/helpers/atoms/AppConfig.atom' type Props = { children: ReactNode @@ -18,6 +21,7 @@ type Props = { const DataLoader: React.FC = ({ children }) => { const setJanDataFolderPath = useSetAtom(janDataFolderPathAtom) + const setQuickAskEnabled = useSetAtom(quickAskEnabledAtom) useModels() useThreads() @@ -29,8 +33,9 @@ const DataLoader: React.FC = ({ children }) => { ?.getAppConfigurations() ?.then((appConfig: AppConfiguration) => { setJanDataFolderPath(appConfig.data_folder) + setQuickAskEnabled(appConfig.quick_ask) }) - }, [setJanDataFolderPath]) + }, [setJanDataFolderPath, setQuickAskEnabled]) console.debug('Load Data...') diff --git a/web/helpers/atoms/AppConfig.atom.ts b/web/helpers/atoms/AppConfig.atom.ts index 75343d722..2c678b77d 100644 --- a/web/helpers/atoms/AppConfig.atom.ts +++ b/web/helpers/atoms/AppConfig.atom.ts @@ -6,6 +6,7 @@ const PROXY_FEATURE_ENABLED = 'proxyFeatureEnabled' const VULKAN_ENABLED = 'vulkanEnabled' const IGNORE_SSL = 'ignoreSSLFeature' const HTTPS_PROXY_FEATURE = 'httpsProxyFeature' +const QUICK_ASK_ENABLED = 'quickAskEnabled' export const janDataFolderPathAtom = atom('') @@ -19,3 +20,4 @@ export const proxyAtom = atomWithStorage(HTTPS_PROXY_FEATURE, '') export const ignoreSslAtom = atomWithStorage(IGNORE_SSL, false) export const vulkanEnabledAtom = atomWithStorage(VULKAN_ENABLED, false) +export const quickAskEnabledAtom = atomWithStorage(QUICK_ASK_ENABLED, false) diff --git a/web/hooks/useFactoryReset.ts b/web/hooks/useFactoryReset.ts index 06a637572..da0813060 100644 --- a/web/hooks/useFactoryReset.ts +++ b/web/hooks/useFactoryReset.ts @@ -30,6 +30,7 @@ export default function useFactoryReset() { // set the default jan data folder to user's home directory const configuration: AppConfiguration = { data_folder: defaultJanDataFolder, + quick_ask: appConfiguration?.quick_ask ?? false, } await window.core?.api?.updateAppConfiguration(configuration) } diff --git a/web/screens/Settings/Advanced/index.tsx b/web/screens/Settings/Advanced/index.tsx index 70cf335c2..0433ba75c 100644 --- a/web/screens/Settings/Advanced/index.tsx +++ b/web/screens/Settings/Advanced/index.tsx @@ -2,7 +2,7 @@ import { useEffect, useState, useCallback, ChangeEvent } from 'react' -import { openExternalUrl, fs } from '@janhq/core' +import { openExternalUrl, fs, AppConfiguration } from '@janhq/core' import { Switch, @@ -23,7 +23,7 @@ import { ScrollArea, } from '@janhq/uikit' -import { useAtom } from 'jotai' +import { useAtom, useAtomValue } from 'jotai' import { AlertTriangleIcon, AlertCircleIcon } from 'lucide-react' import ShortcutModal from '@/containers/ShortcutModal' @@ -42,6 +42,7 @@ import { proxyAtom, proxyEnabledAtom, vulkanEnabledAtom, + quickAskEnabledAtom, } from '@/helpers/atoms/AppConfig.atom' type GPU = { @@ -56,6 +57,8 @@ const Advanced = () => { ) const [vulkanEnabled, setVulkanEnabled] = useAtom(vulkanEnabledAtom) const [proxyEnabled, setProxyEnabled] = useAtom(proxyEnabledAtom) + const quickAskEnabled = useAtomValue(quickAskEnabledAtom) + const [proxy, setProxy] = useAtom(proxyAtom) const [ignoreSSL, setIgnoreSSL] = useAtom(ignoreSslAtom) @@ -87,6 +90,14 @@ const Advanced = () => { [setPartialProxy, setProxy] ) + const updateQuickAskEnabled = async (e: boolean) => { + const appConfiguration: AppConfiguration = + await window.core?.api?.getAppConfigurations() + appConfiguration.quick_ask = e + await window.core?.api?.updateAppConfiguration(appConfiguration) + window.core?.api?.relaunch() + } + useEffect(() => { const setUseGpuIfPossible = async () => { const settings = await readSettings() @@ -361,7 +372,7 @@ const Advanced = () => { Vulkan Support
-

+

Enable Vulkan with AMD GPU/APU and Intel Arc GPU for better model performance (reload needed).

@@ -426,6 +437,34 @@ const Advanced = () => { />
+
+
+
+
+ Jan Quick Ask +
+
+

+ Enable Quick Ask to be triggered via the default hotkey{' '} +

+ {isMac ? '⌘' : 'Ctrl'} + J +
{' '} + (reload needed). +

+
+ { + toaster({ + title: 'Reload', + description: + 'Quick Ask settings updated. Reload now to apply the changes.', + }) + updateQuickAskEnabled(!quickAskEnabled) + }} + /> +
+ {/* Clear log */}
From 2082b81d64835edc714e0552f0c928950bc9cfbf Mon Sep 17 00:00:00 2001 From: Service Account Date: Tue, 12 Mar 2024 13:40:59 +0000 Subject: [PATCH 18/37] janhq/jan: Update README.md with nightly build artifact URL --- README.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index 107a35d29..9fc133a8b 100644 --- a/README.md +++ b/README.md @@ -76,31 +76,31 @@ Jan is an open-source ChatGPT alternative that runs 100% offline on your compute Experimental (Nightly Build) - + jan.exe - + Intel - + M1/M2 - + jan.deb - + jan.AppImage From e5369ed151238acec7c5105360e1282f45c20a72 Mon Sep 17 00:00:00 2001 From: Louis Date: Wed, 13 Mar 2024 01:11:21 +0700 Subject: [PATCH 19/37] fix: do not migrate extensions from quick ask window (#2336) --- web/containers/Providers/index.tsx | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/web/containers/Providers/index.tsx b/web/containers/Providers/index.tsx index 73445f10a..e70a56ca8 100644 --- a/web/containers/Providers/index.tsx +++ b/web/containers/Providers/index.tsx @@ -4,6 +4,8 @@ import { PropsWithChildren, useEffect, useState } from 'react' import { Toaster } from 'react-hot-toast' +import { usePathname } from 'next/navigation' + import { TooltipProvider } from '@janhq/uikit' import GPUDriverPrompt from '@/containers/GPUDriverPromptModal' @@ -29,6 +31,7 @@ import { extensionManager } from '@/extension' const Providers = (props: PropsWithChildren) => { const { children } = props + const pathname = usePathname() const [setupCore, setSetupCore] = useState(false) const [activated, setActivated] = useState(false) @@ -40,6 +43,11 @@ const Providers = (props: PropsWithChildren) => { setTimeout(async () => { if (!isCoreExtensionInstalled()) { + // TODO: Proper window handle + // Do not migrate extension from quick ask window + if (pathname === '/search') { + return + } setSettingUp(true) await setupBaseExtensions() return From fb893b89f3d2f1b04b54ccbafd6482d1ca2282f9 Mon Sep 17 00:00:00 2001 From: Service Account Date: Tue, 12 Mar 2024 20:19:19 +0000 Subject: [PATCH 20/37] janhq/jan: Update README.md with nightly build artifact URL --- README.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index 9fc133a8b..a50691950 100644 --- a/README.md +++ b/README.md @@ -76,31 +76,31 @@ Jan is an open-source ChatGPT alternative that runs 100% offline on your compute Experimental (Nightly Build) - + jan.exe - + Intel - + M1/M2 - + jan.deb - + jan.AppImage From bff20ab87bdc3a4c5270226aa9bae470228a807f Mon Sep 17 00:00:00 2001 From: Louis Date: Wed, 13 Mar 2024 10:18:20 +0700 Subject: [PATCH 21/37] fix: put quick ask feature toggle under experimental feature (#2338) --- web/screens/Settings/Advanced/index.tsx | 52 +++++++++++++------------ 1 file changed, 27 insertions(+), 25 deletions(-) diff --git a/web/screens/Settings/Advanced/index.tsx b/web/screens/Settings/Advanced/index.tsx index 0433ba75c..6853a5104 100644 --- a/web/screens/Settings/Advanced/index.tsx +++ b/web/screens/Settings/Advanced/index.tsx @@ -437,33 +437,35 @@ const Advanced = () => { />
-
-
-
-
- Jan Quick Ask -
+ {experimentalEnabled && ( +
+
+
+
+ Jan Quick Ask +
+
+

+ Enable Quick Ask to be triggered via the default hotkey{' '} +

+ {isMac ? '⌘' : 'Ctrl'} + J +
{' '} + (reload needed). +

-

- Enable Quick Ask to be triggered via the default hotkey{' '} -

- {isMac ? '⌘' : 'Ctrl'} + J -
{' '} - (reload needed). -

+ { + toaster({ + title: 'Reload', + description: + 'Quick Ask settings updated. Reload now to apply the changes.', + }) + updateQuickAskEnabled(!quickAskEnabled) + }} + />
- { - toaster({ - title: 'Reload', - description: - 'Quick Ask settings updated. Reload now to apply the changes.', - }) - updateQuickAskEnabled(!quickAskEnabled) - }} - /> -
+ )} {/* Clear log */}
From f1db7a4f86090d1d1662a381824e5b7e9a488b17 Mon Sep 17 00:00:00 2001 From: Service Account Date: Wed, 13 Mar 2024 09:31:13 +0000 Subject: [PATCH 22/37] janhq/jan: Update README.md with nightly build artifact URL --- README.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index a50691950..23604f01b 100644 --- a/README.md +++ b/README.md @@ -76,31 +76,31 @@ Jan is an open-source ChatGPT alternative that runs 100% offline on your compute Experimental (Nightly Build) - + jan.exe - + Intel - + M1/M2 - + jan.deb - + jan.AppImage From 7048c85465eae1ee568f75b9965d16e22ee9e414 Mon Sep 17 00:00:00 2001 From: Service Account Date: Wed, 13 Mar 2024 20:17:51 +0000 Subject: [PATCH 23/37] janhq/jan: Update README.md with nightly build artifact URL --- README.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index 23604f01b..bc206f4eb 100644 --- a/README.md +++ b/README.md @@ -76,31 +76,31 @@ Jan is an open-source ChatGPT alternative that runs 100% offline on your compute Experimental (Nightly Build) - + jan.exe - + Intel - + M1/M2 - + jan.deb - + jan.AppImage From 24c6dd05bec712b768e3860ec380a02417279ca2 Mon Sep 17 00:00:00 2001 From: hiento09 <136591877+hiento09@users.noreply.github.com> Date: Thu, 14 Mar 2024 13:12:30 +0700 Subject: [PATCH 24/37] Add icon file contain image size in file name to fix linux icon (#2344) Co-authored-by: Hien To --- electron/icons/512x512.png | Bin 0 -> 38651 bytes 1 file changed, 0 insertions(+), 0 deletions(-) create mode 100644 electron/icons/512x512.png diff --git a/electron/icons/512x512.png b/electron/icons/512x512.png new file mode 100644 index 0000000000000000000000000000000000000000..289f99ded85e1e3cc91b17af6b607d2cafd97cd7 GIT binary patch literal 38651 zcmd3NWmjCmwr%4Uf(Q5D5G=U6yL%EW1b4UK1a~L627 z(G9z*man;Hg()jap&}C?gFqlu8EJ7<5C|Ih6&i#H4?M2jW*>nElAW}UBM5ZD{q_T8 z&oB)90>XEa&~{R@HFI+P>R<|Tb#?t@VPolN{MF9%ldXe!`WZg~2t)>w5f@Q+%Q#te z|6-ugba#1PbBFJsSV(Qb%YvJuz(BFo1ShX2j&b1EkHYO==}_rWY1nlbu6yJML!SpN z0p?aUY0PXeiDi%p9zkhy>@K2P$52}Ci^a0nQ4;T{Jd^LkVYpPaz^!}BP2JZ zV_sR8MO7MZs*_}@1M>^Z88lMOI8qIsPy_)FMoUmN)_*=B@H>I+eESBTj1THB@GOTK z{Qu}rXg|DdyxjjJ3qdPVE0K;wH%N-zgfvs&C4oa_zxq)~Wv-ZAJINH4yW@wh!M=w_ z4PyNIHCkdvfZ@#Q3FG(gDJ%FQ{Fb1KHif4Tu{Nb%)b!j%ZzD;>FP6s@j~&!L{D zuEZtV&HH_&0VSF4=6mO;JB#4>ZBOysTNs*Fc&O~tQ_E?U>(a9@gh=sX?^&X)XFEP7 z{KNkU-*-W;S;sP1TbACova zh$++D;L%B=T@gKY#|5fLeGVeR_~gxN;6ME5MK^n4Gq?4)LVSrlGsh11yD9FXl0dmD(^GRWw?u{ja8t#4YqB6QIT$>iAH5h5rx zMC8 z*}?FGu-osNbO`rz-qMNr%In4%PMIc5e0Rww-_dZ#O`aPpBgCTiIir7?ok<*%&xiNA zq+TL1!htE7J59&t}go z!DF=a*y&JjBfZ*NL*6SE+(abh_gtFw4de{gNI1dtGew0xUROnOu*D;N1#LAZkh;Js z@(=8SxOENAY>j9&zV6UFu^-c{Zfq&QcR$&Ez~Ao5lS>w|N`@_8r~<5>0I~Ygh%5rh z5pf_=#pfSJ==R?2Z=&Jv)yHOd{w4;ao|}`kzVYO&i5HRGw4;3Eg?d zS;;rip{jBkmxH2x{_-^Dpbh-%_^5?iGEc6WpKtKzUb4B9^JIbh@7d=8e;*NAkS41i z%bfF=#d^*@yMABHvRTSrhUR&#P%#x5dwAT8`9?fDU07&L=8j21hu~esKxIitt#lSa zF=BYf@rQUmFV$0IAehasdO7~&X-}SSw@dj*4{ksaRPJNE#&0j^%v~2}<|a>f!_W5d z-)8NV>YT(9iF13ao>~Ja_9`z5YI3nUb3Ex3P!A%FR(4rg_h7A%tI)*(o^$mAvb=AP z&=QhII_`qyUWOMq{$cn^@@^f{Ff9;f5LxOWDJ|xNIPPh#nErXA(wYPj@L_u0WrBjy zA+QaU4Q!{ArT22Lq*NshNXp zJ1?qA>u#Lbsr6Cy3=J7qBfE^;We5gJr~2C*ufOZ<4ZZ}arcwvHdH+xc3j4Buphs#9 zxRlUQiRXAWEL{v(2=;9uNl)Vwm_@5wOIlklP4z{(_Y1rvLgoBMdXaNXu$EFn@Yu5$ zmBh^aa%M{7a))}ZPOS8!VN26EW#))S$AoS4WyEuG=Hn^N-A)x^I!m=jm?{+@QEpB; zy)DdiaQowBWlE(r4zwq7C%sqnd}fOK!NF5}qJ}HgF#Y-&wOr>ct9_1sVx>oSJQK=J z{2nrHU_~)ZtBQx+*#NXZIgZRHcA{O>Yo`vSa26e69iKaJVuy|cq!dfOkT+rRM$KoZ z#7CDS!Vro#SIYCO3G&*JnD(|Fa5ChD?Bn;wV+TeX7yH`uFK2P3U8c2{$Y8Q76Bzxo_3*qT6kVx6Mgy%7T#TVk!tpb%^hYnM%Zb5PUC{@}D~eR&VrBA+ z>$zge6Q$-t&V5`aWP75D_jH({DQS#-;@2JVGOKQP0cD>)t6#;uo~*$ zoD?u<8hmZ7z~Om2`N)xqbHe-|==NXJ`&!X96f(_bhyhVPD2Gcq+%}BT;Udr-HafKH z^k+%6V5eE6+kvn?8C38M+(YeOYO^DEUrt@6lD_S|7Oisq*h-%#_97)4>_oJjG!6s3 zjVrn2%H~cd0|s3nT;N32WY^{|Vf@)o(*y@Edbt$kjC zYim4<@jgGl6#kM`25VOe(yH;C`OF=fWn+tvE}BPLqP0_1fX45Ilw&pP*_Jv(!$t4# z710w~Jn;D3@N!VR$N3N9B9fm5-u%2baxKQ7x4)ZEul;KsKRScVLnUR#y6!lp==~&8 zwoN#1mw)JX42k~nwn!9{l3mpyKdN~~HEL{jO2t!JE%!Luxnb22TBNT&{-Z)jO_5TQ z!PEq(FxoAVc8xiCrDJ$OQL*geACezV8%N$J_=lm_`NsfQ?Jjj!3gx)T`X`)R0{)b~8kqvM3a%@VOY|5Ce*+>0{k!qai46H3Y z;q!iBj0gT;8YpNfZhHe3+pL$91GJM4aRyyWsjL8L>HQSXFvP<&@`U#{9aDN6tADIl zB;sM0^kcLJtk0kgaGfJyFfTnB|0a<5eBm9}F^^m&KEGCkIb%2QB-IGN8v&Hn++1}T zCLX%YQsV=jqZnm3(oy6)oJ}-hH(D6x0_-BU=CN_RvyCaj_Vp>Au`=K49;5vZj&{__ zJ(Uo306AH8QebtvDJj>nTxZFZ*T=eEwsZ)9dgaAq~9dG4z zauw~509!sQC`CueltdQM0DG5EE~%0}lyfG@7cdP8<1Fr8;|Ae%*T?i2TBbZJ>ED#= zq_&11lhcN=cwl7({#OBL6{q9{r$=neDgXpMX!_gPW0$oeL}pkRufAt$wHs0 zT3LU$JoG7*I8<&{S8UrI1BF*HtNy|;<=ctuC0Cn@A8hTsV`(eNG`5XlmMmf07hQ@* z+3@K2SC_=0&S5}ySV*$)#3!$rFKTWB76SU%q8~wK^GXH_RYl(XcIMi~AhFng-~xqB z;wsDU-v`e|oib3$S#`B}rk$s+BIj$t!)tI~c8&5HiPY|3S=g%Im7aJ#zyU=aSnsP8 ze}5o2xO>$C!fYu8dGDq)%lf{5#LU9?)?#+#^g2ygt~?}mmvn9~CKu|j(?h8el70~S zd~;hVIFpCH&__({v7q(DUn;~y(Iv67iu$;;C_6>K_-H}_ShlAf>M!_wFt!M5H$=z} zu|O;Y;gCo>c4+cmH1A%>IE6XORSEE`KKs?Rc2pkK;k}58DMaY=h86_>Sba>D|5S) zZ;RFL@Jh7Qbr)cg=W5r-)RxGj_c!1~h(&cFjv!zX;N!t5TZEDb-u`x3%MkLYa|rfF zhL!tarY&EsuY#c~7|wz1(g}%Jau1JSz9x5aclVlWA@htW4xnsR@-vKBw4H`fTzop+ zLz7ESb66){M5rlJc=;5}#l=e@g!rVLXPT4=(tg01u~(6Dqonz8wHyNv^3Mswtl+{M zYK=O{L|&}aPUi81}O}@QW=Pj zSBT$^A({`9#__U4wOBq5st@_x8yL9kY=k!gg&I%eS~`dztfw%Cy;%B@NlsV^k4SxCP zc*GGPy{V|YMy8>$U8WS!d41al_D826p2aW9sQxps{9=P4YSF>SQl_vigp%i?boI*! zla(@D%!XuqcpwegPKz{z*92!Hy#;A-MRA$s>fP6K6o$~tbAPm@g=yt1% zlJHqIiIB&Q%h1zuVepKir7LgTIeGreho4?2XE+Rz>lX`?lar_J8AX#e#U}FK95PQ) zmHz>ib z%>VHtpcV=K{>zmp&wx5fo_NJCUQwg4PzYsOH|`Q{_*DCyw{t(fxswL3z`2w*k*K@9 zx{ZFl*IY><%0y|l2h9Pr+Ku=Q*b@j8PObgO(Qigx(25Osi=BS*^%scsPbl%m3IJWe z4>uXVM@54^&W)PVx^$wutgUQ#pvtlHG2yOJXWE=9)a{3S^(|sj-TaxUsb{{>8^V3ig(g;t_KQA z{o^|H^lsp=CK$fDQj=OJF!T~774}DsK=tf3*U6g_?h;T$99q&YL9j31jMdcsMsjG< zzRA4ON@+?S)n~ zlYd#7&4R_J9^o}fYs_u*J~uFr(r}5t6@B*#yR*g!0k=LIlmG@jkbTn`79yeJdX7Sn(=>d+q$0LlDsq8WXF65j zKh!^T%z&yX(YLhUC?J?gMOU%4mQgxPTeC4t63MB4vUQXLTgI5|ATy%CixWYWyTQ6@ zHeDg1g-5N>%9&_i0C`z{UW1!+(`offHMJjuK(b$2n{A<-WtP=b~2kJi=u1WI83d98_YmXhgcn))g zbT7j2umV1<%`*5SZTPn;5~K3_($rZQQ2Ekynu-Xg(%|_j5Dw}(^>z87-?wCzSxpswWekGPKa4b!d4o?y`|C5)h)hS z@_dtbm>9B_Fn6Eb&u{s5__5q0R31cTF&uX9a>;&IFYi84*TAxR_77=MrU;Bw%Inp8 z?V8IcW$+kpt&2Z|`Kwl0QD3>kIG4XcsNwxP`+Op25xl{qmriNoI_|#-qDB(=Y~%fo zk>C~nU3jWPJ^=e|4Es1Fo1K{y6kWeU7e7i3uJl~~l?|nd&qP0dC!gJRmvJ`2^4C`? z5C!*jZ?oGnUZC?$3I#^u#GRGb8#zAd(Cehd;wkLmZXIzh1=B(F}RyAEC2`_Donw!tlGE)AXzBvULzk%KZ1_0IDhGm z@O-$d@Jf!lL`L){TZPS1293q{qEs*W*0AbRtn+52sv)j z@XEAcuP|gUcGZ|=bx0&qKW$OVJHGfs9En-HP$RS7YB{Vc^15uIfbp{Ctza2x&_6RN z{9#iLUY2+&OJS5#c>p{Fgpcp;i+Uoi;>0~@VuaAqwoI&G*|~`7dj-TlDg%Q-q4#U( z#Vn0QA?J8e*n^`M@wV`7W7;AGr6DAFcY>p7oHv5AokJKQ5H|Gxd)-ZaUKMyuU|)Slm95GlhEAg#vffCzU)V2d zfyH=fDRL)HOsRE9;4AfnvtpH$O(t^Y%~_Vy-rj>n&}mX6-1OMc@nZd(qs%+=rGn1V z-w1(`vZ8cK09kV#a5$W{AteF0X*}n59YhPL4%EaTf28wqY{?MbexX-niakS-eHq3? zurI$_%51A&bgBaOXvQO5G{d)()>g!UyM%Xa9!w7KnJBPnttejKU!NadE0=4%0PL$w z*rXlxNFuYl(nNx)D(;Kt1OL6tmGzXN&J#0<5Pg3GTS=Pk`W;8;#j5WbeZV^@WRu(4 zp(~Ndn-S&4>$4CN-f8u=JEinTw8=iGk$1!{&w7!cl(fA8j9T+6UdIDzF|2nc9x=vr z0E{hSIa%=3m~K1ZAs>eY06`}bDk*O@@SwkAP-2-dA;EGvomZl3{Zg2g6c_2eApgg( zf-pSx4bH5K01)puv8R*D$@J~?k5%+69#nk{Bq}6je_0PF$ zhwt!Y6TiPML)ozgROj%TO4HChJOC229o%U(RbUuMErB!7T4RC2=5L!9@`BA%$$iUVe?C&W7UiQIV9dLXpGtnL8zdRC z0;Sm=>n6}aB>gJQX%IUqrV-*)D@#7+PR9De?1vbZ72z4xC9O#cxV45h(Ssywg}}M# zH!(81TD?9~_$^qiz+_kv+jwdk0NqXPY|+KV`Dd!&SJQH%im_;wpJF#Wji5OF~#jNA$8;7UdZd&l2vpZN;5KeKTWQM=}W4 ztA}0RZ93JzNTuLR=XsqLl_ur6NmakR^)#-D6?4C$vpb|&+|Km+x3(&6*qAu&O$-A= zhVV{O#64UX(_MV9{rV}SFQy!&v=Xi|=)$?>E3S@Y{ynRolL?}*{{c>?8D1R5;Eut) z*&DA#;5ozUxzu~p2?mAfPh6U%{`ZyUCwoOxTc__4(dFRR(L-6rno+y1YuqcVfPbT~l^KM+b3VRTcw#~fZP~?>Tem%mt=PtBs+w!=Bfi*f)N;;H zG(l+}pszEj`DsdYt zAo?B+CKmMF+2aiTYx+5V)>4I^`&&leO#7RIe-jSITk1$zL=R=qe*t^R5K8YD=~qe~ zgntfDDk(i88(|%&y}t0JzDluEeCfr-ZZ|DVyVK|#G?m7kT{u^nOMljyybI4^obLic zipD>-mP&YzS$GZtYEIZvw%q)TreKPPt6K%;+&*`4HKJton+*`${kJDko_T!=_Ik~a zmEx>+@@i`yci@l-WW%E5^j%4^ZT681Dqrmz7d4%7n74rPSHY(pi+Kx85N6tp7uLt; zn1W|Y`J->^D2}wf;Zj05USPcRy?1LbR>kMkfxut=QKu^RMp$*~z5f*a%`-V~cs2w( zCVq{0%9~cbAAR%qyOlyC5G0gW5`C-LpJ1kynncYe1J~5nFDyM~})|DU>)~|GJLnvZCefTSJ%A0vtiqt5(S^jO3)l|Q? zk-uW@&W8wQK|aS1x`s0360diNTi{lE74QM=T#=%qK$Hty9(_X{=+2&c>SJ;8N zKhA!=_uSC&RT>!rpn|)Fk;Nl+z%s*o?osbRXp+Q~c;p6zturNv$>%>3Huz9p2q3*; zw9(?6KEfqkuj|mEcKVJC$#@aQy-&x-T*w6X3BmNJa%V<3_Zm2zD@~398a6i%Zi|z* zXP49t+6Zsu|K*@rhyLgST+bS+zrU>j@$6iW1~vGedNJ1<1zj#FhOzScr4`Z0T8B}J zAi^YYfpQ(z1)ee9X&VfY-+{W=_|V~sTkLkume{i-^tRQ3d>T>-R1Tpqw`wLUmlfTw*N5yvgb3H!6hdFlpUxl)@M3EJ)(tE$FpItZ7dD-1rl$>z zVWicwpDg?wtUDHEr^aaV&!ysdVEUAE;0lAQ+B7Ll1Zt+bvYa#!B>B@+WZ~Rp&_A2~ z7I*`j+a)gQ#zXO!(2Mn_tEJa|kSU)XHJpebP2LIUZgGp8|G=A39XD6lWi1#xY=UxQ zRmRvnS+6uCKEN3=M7_Bo?o<_;xHV}Dqh`~70!B@mm-rpKIJB*Wj7@CS>)Z1&hL5B4Vny_3DL_*7lKf-=hq@i8~4lEEo5F*rJmg+mu(n-(m& zc5Jys!QI}~j<`eZp`CrltT5b<`$Op$N3Xctf=9F>A4eDtf7nvi6tPRE9=&@;yidu| z_P&dv$ueC%kW9D&3)YjDbcr3f+x%J|fZNq2Fg>v}U_cR7Lb-|$*kuTRv?idUcqD~2 zQaoUiy!~Y7=%Gdx^JD&=HoO(bEeDh7teqZ4bD$dRSiNenUwWitNIg?@x&xs5{F%2~ zCD)H`&0cgBo(RUIzMoFYb9B`f{00qF1`z3I>OzPoAnf1qofGf_#usqlDR#|XQe3wV zdBgAj%~$-XoUF?mySvJLR+@Y8$n28#XELY)Ltrlghr7__&D{b9(koNw2KtY8Q8) zLEvfDmA9d-T0*OB*KmGT0gc_UnA*w;RugF4inXf)cYb#gL-%3%$mZR*gU1`Hl=y=s zSuR;?Y!n*S&T>AUTzHa?`bun1-D1CpIv-Z*0a==U#t;`97Zf!JXg|HRT0} zgxJ5NqO*IOr_hqs(ORr0St;L_@gHf+hLA{B=T0wBvoXi#)YL6nc-xGmw$!&h3cK9s zk(iAg43U3C%Lc>!J}@#ZD0^px7}#7-(r!hebhd*q6=4~S^-Bd42pt*}G8gRk@M9H1 zB?@H9>+kF{77M`}GUHDmqNxUa!gUW~J}6qJNScqp$j2e$U$%1L?W+(RS1g&1p~zv( z<~!a!?%`n4O%Wau?x8!vqVJNCWx17zVfwtJziA~8Q%88X5-yBM&imLW|GsIhWF9k{gkhWya_ zG~q@f$w{I6=6XR^bM|7~yie$DTAL&n4RN7wIQ;Wp$`Woc-d2HWQ*^;_R+8D`|Hc1PnO3#+oPev*Sgh z%UUbQ5-E6$_hoh;#{lzpZ|_suJVBFpHP8dV#AevsU?Y%rOqRFPrgY@b6P6P?LRa1A zA@X>qhf>MWlKQ~zhLG--b%iA#e%+U2ez(~6l4#@+9nqRaqc{lhs3zU!6bJy%6dR`~ z>soHO{$VkGg60u&%OPqa#LXorM=|J)GsS6?fsFZ!<3WT5zMe=5V9jEzH%NtVm<$oL z68f#&LKS1XFHCd;J#)}ORWnwXUX)$LI_9N%S~z7ixyI?a^=no==il3DDdsws?Wl+BYH64&?`>`xq z9z%{1Cch<5z7*#@ROKXk47|t+a?ikd=DvwXd{_UksSW4NBYBV>AXzwNjtX*JDZVb_cKx z57zb=SvV_hcwL~CrN5bKk!@zUJL@M`_e7ED9G*wo?ce^HYpJ(}P)5%xa1f1^1GOzB z<`gS;O`K^WK#cez0gJHvc;ntt&q-*t?+#^$C9@7`J7L=Fu8TF8y@p3cjprk7>^r}T z%A1ilO$8cC*U*cQfY+F#Ph&^!z_&1UgSBtf6>{HFY!P(!dq+f!oLZNZb8jdEYo^_n zG@}Pp%Om8c;0k*}AGl_=BRqek1bF>lK4W2&!}l3sqTZaZCHojmk7w3^ibAnA7V_EZ z-2I#6oT0JWkRSr=5egtsWYG#5?MhpRFYN`Hp$j-;rcltzPW~9xiecj?*?Y_5c_?>J zgT|siL95FRj~<*Vnrr$l|1RS3TP*T6wfFdW%`l@Hkde2Swa{A{Y)zH){C@{uJiG!JrgH-K#^D>!TbHcR1*=|Rz_AyEspOkCGsMSmby1I8j81<2WMwF0i^`} z5r>?q=0h-L)O_6~!cF1)`x^q>j!Wt-m&^bM={I`$pC_`x%mi!#VQ_rtsH$QS5+1cnvjGjqT&b ziYQ`Am$oGbW(7*wcLQui*`E>E>VWR1X6eHS5lrmkxMO%TE4rWb#-1?Qv+D?jds(V@ z*U8{XJIaM~{n{MAEp45s={iWtIbA6eD4 zD961|mi~r&pw=gu$9Th_*VW~bB2f5p9GJ47xV>h#&+|z1{z^_9r{%LWkAEncCiWXK)hakFK&bq|{i@b*%gtE{N@~}%75&ufLBhiY?nwU0%Q<+;OWZYe zcxRy1otq}J`iLj6$I)A}dG|i`JoQ-}uL(IC(Na=DY^SGlQTXq#Y4rw*6-jDG%J}T? z=ZVQPsL<&iDJ}mG;voV6bj|sUXq3LEF*kIw^io!DR221|T-WaKUg=THk<9H>JB{$9 zFO#0tzNxB8CBn&tFaLFtd_K zV+X%<_>SnrNkE@I9suLgLG&1A^k24LtXEIVg#@S5-=D&g$zEQ&(PrrNd>g`)4S9ce zQ@-y$x3qd=30|`+MC<2Yi@Zn#(#;nuF|vsCETo3>-eVIsA%@%APT_Nom98`xa=Xi@ zAcxi*>Gq~if=VPjL8}NKLld2M%kepq^Z9>WP{tUG5zxVdzgC}ZzkNm>X2R3{MbHcW z1^DNI!+&RsWJc+ynP76AJ>tt^|o;C!(!lWz;<&2nh^A5gGASuS(D{) z_vHMV#&*7C)HH zGoz$~(i(H@D3F_@ubaeWc_hg~(LyjMM_0Ukq5A81Ww}L!-^&FWor>;6$d@ZN->LXK zJ*@0ocl9ZkT+@)ezrw#pxDd!5!r@+-|0d)S?hilF3o9SA_Z`sqjk{ z4V*pCVAn@zD`eQUqr#@VjZZs6W*smD2QJB?v9wH-hPodbqY!Z$*63*PLY=;6eem|& zX#1Hph%Rj^P{D>hkZd)tmvqctD((|b?n+TkDA3@FuS}n-=x7HPnUC}^`j__)NjM*% z9${TelolqXoGbW`S{d{<0Qjic)*f-^3SgvhB>c0zL<|)=d;6xitGwN|y{bl+X9AhLjlm#I04^`~ z4cViy9Wd(*M!yyh-nYYvjAMzX>9%JGZQ2}bm-1@gy;=mou)l#cYNeH8b0;jIoOH_RUTt8Hxx}ltna8a=D066vPqHG z&$TJK_)v*u<&K$z|7P{Wbt`snH=1+`NWD%Xa4?~Iyw7J-0_?gJ*W^PAm-l- z8GCNj)~}wR&7_9hn!Fc#^%`JASO#=A)8v9&aL5)?+SO$$Ugz|mZKey?b5}^6-(iUd zPV2(>)^`7{Q0(sD+Pv0sfw~eSrBRYCh2}yKCHiFDBmrJ~2T!(~cNZrX4UaK)7%B3R zM6fx0I!=g8K7u#g*8~2z=eXhr;OV97Oa2#b0GJ!gSou-s@F{RO)&Sq=J|a#VT+Wc`5=JUq3VL`4WcG4aMiog&w=(%ge#5{KrOtXtm5y zd(O@u+J8=Zgpw|oyeDNoIQGltCKu6~!wei!kzy6in2cgrLK43A!FBf2fALiNeg2A1 zYN_)PV#%gPW;ns+m)lVc-x<~??}a51xIKdDH7A@e_HO4T#b-K8M_gJ9;p`)W&Tbm* zr#W8p#{_eU%o;d%(~ho~0sdAsvjwFGx%Qe}6Ro{{;o|gt@bJu}&AVDGzv9IHq<{}z zLVjEh-&lu(AJ(u#5w&U=6YlV$b|-qx4dL1eo!_a&C2BMdZIm%q^dzU@%4GUWBXGVB zb@){vI!Ex!CTO8EAg!#r%FFS@t`xC4F6Afc*xn%(`>1@y5T&$hlB)Nazoy#;$gU5S zDmzCt9C^$z4*(n=5qH74D@)!ngwHY`)eM(`k_(`<3sIad)~n~Z!R_D2=&>v(7RIRA zAErO&209H%2Il2OeG@fPqTw@i2n@gAW6LkRVM~W` zlIq5jW#d!5O4h3pZrU+{a|6`KIGiqsHqP2yYfLmjsM8lL&h?OkVC8|qJZ*Z3T*fa5 zRyq1aKHsUQK!0!V8z{hzkINxK7rN^a%0QeUPx)&Br~Fv}YK_)_@AIJL1TT|0>lN~E zzJ+|6b6Aovz<5g1$V0K7+vyN0{<){awnA>ds1M2JpV}$N&CSBwY*BZ+Tchl%i^V=x z(kRl38F#=VAFpAkeQZs-oPtc(fixzw#WTB^F>e*}iL-~+3B;Q_4Kit;_ z5e(1XZq7$Q=vdRwsplI9AQ9q}1$$55rfjIrNE+IbcPmU<{F_UK?yABKb2z_O7qiJC z0LY0}y=&yr&{zPtH$jU%RG$qqrZR9b64vL*1^1YL5mes?#t4XgPW}6pV-gr_sm1Lk zQgarN+jZUYy*iz`bai>5_vq~gojI^3(o*XtM3pS4ypNR!NvC97Zcw&z#irZeVyU9~A-{?!i`Hsa^e1`mWy>OH$+GL&rwY``>rm zds~x7Qk~L{WW^^g5A9+6FX4?`A6EYa^88zECH50ICC(vk1KdaSjZu&=0`wQ^7N;yL zf#;3~J$=T9C|a@?ck?3OZ~1=fQI364ZZ7;_AHYoY>tZ+a>n(iD=-hRDzwm{w<8J~O zqqU&ux)pM@pSVXf^HJ@f*gF6$Vq84?_vegJS;-bw(UMqV-Fla%ld=z2R4p>ht z;=~n9p)+hMmINGlqH75RRpaw(xLsR=W%~I7Un7knY&dB^w;JFjc!Vd?bO~$P^V~6n zyZJmX0j3Iwb+|hzO!xB@PwV<^q9YN*M|A7__}=z| z5=>aQo^CrD?7F_n?^gNXvrIlHWhDs7o;J`khxXexA~@*)bGq=Nk)Jj{o4uYWc$U?( zGC00}xNOH;ye~v*Ti_wQ2WM&cT>Q``oXPTbz&}V*{vq{I^V@G8bh7S-p3@=a4H80j zq|NSPP2DF#oiE?CZB9)PfhP0##iM*C=E+D63)Nx@D5i7c-D&in+>8=@8w)S%hnBPHx21So8%I%m&XB9 z9f;to#O5>5@fgCyuF11;!SIQd5XY|V33n+OBYL%ZS7AR0+zFNwI}3_l4+M(M)|A1F zy*_D076mR*ECZgg*obchR|3z@!5efNvR4oCu)* zrP9GIwDB87STVX~Rw*^WG~iVdR>hU3b)Kf;tdxV}-C-=mzHB2QZRi0$qPGyWaQA#T zZKv0ZmeC>HCwuD!?c@n)!~si2HU$20()})slCqVkZ>s5cj`PlNgB$N@i&0NI_F5e6 zHX%GkTk3+hK?-h;cuoUPOe|Fj5`>+tH!>w?uQ*HTX9O_rNoma@0cfL-3y?pPJihD` zAv}eE*qSf(3`K4z9=jF~#C71P0-)Q5Ao?Q|4}m58JOFgSNB?p}DD829j`imU#^%;M z9GDX(*3foqCSB#>&t)?P`z#S+(1_8!dxD&?d-dZWc9bJ9?d_hgySj0G<-XWiY1&lu zJB9s?C;zV&wvpoqj~bwg!4OMeZ8pCVe7mH*_md1!tfEX#(7A2OU5$0^>~Z!&oFv=U zt>Hh<5lR29|GWw@`n8?p4UI?Q9Klj@lx_$q3-*Kt0Vso8LUZ;hnpB>L?R$PAnmt<{ zqr`lK4y(73O_a-b26>e%l%5TG}6 zfk>Xe5rjPZXn*;7yT0;zeX^>uC42*RQ+4nBr6TQ%W=L7(Do(j3ezm#6tdl-21 z!>cQV-Fr1+_zjmfle_MuT#S34P(%|)8rH4uNf0z7T3=l^e*wX;v{ygmSoC-luTFkI zitOi*PnPrf%0!qg%?cK%O-#Gb_@ppI`?SA@m7yycclKflr6dtpDDAAMZoXxQXOU`L zXZfTgNU+j#?^tIRB=3DJ^VH#))rCojy?O0K-!!^fg>{YU*@{zXtk)<7>QV%SPiUX| zXGPy6KON1~iNX8{3TLadjveL1E3s{@sDydh(<|YAUmzC21gGEdN9DQBmp`WQD*n~w zBYJ1#fC2{sK_UJpl&oUEP@hjnh!2yiRoOEqFkVM@>m1@szC2;6=Y&+|j_@x-Fr#&t)VDvMBDsXjb$Q_yXmFe~A1E3#?+Aswx&o z5sxxs1CBBmqMjwWD_<%MH(BP05IX}H>em_J=N%}j3(wxMcxBXrOj}xpTuyx%l8+au zKxNDT@d0kCKxKfXpw#OTWFfmhbF#mee98nT zZR8lnpElzBvo1G;l1WIlc>mF^G5UkD#hmg>*cD;MHNINP1C{3Y3`Z;ebw0H#Z&!@{;8s>~Y}*Z8fSeLZCZt8zns z88l8u%pQM;wc$r+$?)Bo;bw;~X0xo|ufFCGqnbM4f@4PXI~V9l(uj31@>Z&C4*>>O z)^98A-&}t7N{waKRYb-hkFqVR&;8M<>W@P0!8Ao$m^EVO2(N=R z=ZzbP7cmrY_v*k%(Wf5g8q1FaZ2s#fkTcyT7bYEbKd?`z=LYbajgZv(e*V6s)odmm zBdGwP&S>%W&%kL#XEB+`6gHff$%`qiV7)SbFEwD${w-JeU^QYZ4a zWKX06cEC*g6h6`s^YTk0sq7}5x(@OwnDUZoNUJx)~#dqfxa$wjW{yCyVxF}sm> z9;dFazh>OcEfdk+Q_K739|zjW<;f^uitTseD;TRN@A!ypFK@;lzFCK4Dp zS`}TAlk;amqFq^=L`=#)`FMZ(+Ya*uawX(_>lacWUHUH9?OS$Yy?ONAUT2T}H>jl% z2DLH-$F?SQ1QuWbA>xbu$nBm~j%RWNV_|Y=w+CNQJ*Ee3g+@|jm-++a9no^qj$C1Q zsz#?R&ATJPUf51&aKkGaAfS#p+r@wZDgUhEyp*jsh$9-&qx!r%DG2k`V0(Q0-2XP+ zIMp-pJ-*qFNdbMbRfoVvq`$eTrYhz^RD`$bCWE_m%p7j?TqfdM#qw|V z>eln-yuFpb4zvca86H~z=gGS+^DWb&`!5&3^VoPOuCryjZ!?4mWmh1W;&lBVu4|ry zMX|9s&`f!x&%xq4etTxTM~&jm_Hn#+hD^_yrNJQOSCa(QeNG`JtAy~mk~+8>eFLs1 zr0&@~zyD9-1gI==4>~1D>7V z-b^2zfZO-U<(X3buR4}_coxnLzNN*6@YCy@H@T{PuXv4^&e6 zLv74oc~h34Z%_iwznqkvXsQY*95Y2H&cp!3e!o!|ZH1f`)&-hHIl^l6FM2Ny0 z(@A{}`LW;H#kg;8AeL`eLPZNmhJ&tat`0BNdG7^|x~axccA+W#Z_T^=nWP~{BOtS6 zXul@^BxwFhqilFT0Gf{=Mq%uuY|bkM298kd6of+2@9UmOJF$S5b#nsrm)}N_yulVg zhWtsAh>1~Qt3wtqZ|V_`+(xuJ2egyDv#Nf=n43w|9db}uTmlh0u2>q_m<6W&0OaoV z+cZ|(5hh!Y6NcqlDBrGRKlbp1e`g}h0QxWZ+jF6%A3$4A(bVO8$Aw=!WE95BXknAk zqhRWhsY2-b^R5nPkjhPzIc&oG)u4v&8=`kmUb7c`ARj<14Mo7MJSpu9D!{XdW`-A`^ISrFuds;#l*sc3@IwqFUA-g-2NOA+Y$7LE@_!l|a zI?9g}38<_!*c)A@ZifRF!_bYSf}|h~0!m1XfPi#&r$|ee z3?bd!E!`j>Nau%;2I-LQZkYG-{XOsB^V~CcpL6!wYp=b}xmbReq4dKbx9zp&pu6wY zyn9A4%6%qaF7DxSW!dAbu)!@z?vlt`LkRc%0!}I)q}7s`k(aU%jE`whB{H4VRdX?r z<5zg_oj05K8oDsSC=QP=%ALjoef$K8_{0hR#~*y{_FKjL9)jY+ob#3;G%c1aoZeap zv`k~$$JZ6$9*mBd=cKyU14(FD;h>Zn+Ma5mz3rJOhDBxf-v|~(Jre1I-~{Xv+_^Le z@ZGg8@$NSZ>~>i${LZ9?A79kHunTXPXwSc_mDkSXWB=g!%7qiv=@!wuEt)@Re?dRm zG84v_BXP!_P@bCVmKj8v!oOD6(bOpVGC~^jIF*dyvAmy%D>O9pv*S2n8inia7TV&_ zvlaZHin}0HwTS9X7Sr{w2XR?q%q4d$nq&3WjVj#nB;X&l;bzWZPyG^yjNOp(0Ppt z_@g^RUPr7=NQA!ZmoI>hHPHO=Z*7J2&(Q-c%5xV?qc$Mi4?ji57j7ymI$$vQ+Om_U ze4#T3w<9apf{p(ab1;@@m3?e`DgmZvlEDT6TG5i;ztu3ojkK3C-MSsTIGiAkSu{6~ zSyLe}T|Q2Sb(Udk6565Tm{`1dp2he_y{|1WoNP(nP&YlRoQ%2{x2bE`|z*BHkIn5 zzISUT*|%KYS{Wedc=v!S*FERHGvo}rlh@np8k!Hs!xS8?k)pH#3Bo)=dt;(Kr?C8E zL4(DE5nr{T-HF<-0-Kywy97esg{WO`2`iFlRp;`UJYRIWOUA+Xwnt}sJ8T_jY?Mz> zxM%ggz_gC%91QaxLrhCAb*VOZJ4HI3QmuX151N>(wive%Gm7-&5IaGl*h zqBGWP4tZ@3r8{QM_*{VnZI$>TN8!P?ls46P5^#)Ck0 zg6&_HzOPfZ;Yw`yQz&`GPTgm+A75%~4LRA0@sIezqbK=cgR^j01&L&J)f;PD{H@kO z;h7IqiaBD3@WCW4NMLvL`U7kDCo;Dj6Uq#i1*&>YxukT~XKa;<(-grN78zOH!_s~& zRH}nvY|pt<_#Us9AA*kq2X`nEhyY2$3gMH~_G}tE2^MA%DstUo$pKUvP0ccE?5%IL zJiWcsH#|CKiuAp2*&xwzLQ@g;WYr%MkH4 z)E~P>v*Kos)TjfokcA(w{x-ca0Z8pgucb=tR!{)T8l;sLQV9dx!Fy8seHLa9YPggZ zf2d?;xZgm|sDEU^2;^a+v2`x7JNKn)oR5UbwDls|s9b+XIRFH!V>|rj7!H5Vc6K86 zkqJV$Y=iR8#q7k3QCU0EP|)$1`%bYd? z$?tBHfC>!)=?NSn+?SD`d+-bAJLBqzNKBZD5eJ|9k|Xr0wtICms%^z{9}s`ism0%R zcD9uC7Dw87Feg=_Al)HavB@EHWw!Lx_#w6|l&qew3SxPHxGZdLvO=2W6+8Do15idq=25a_`OR4 z_q800TD-n|FrzLCwV@*Gm$RWT;-N4*1k6XQz}GktiTX^R|3Mu8VZ8iUg3oV_;58rl zXh-8d5}}W3Qi53Vp&imvc-c_E-)!k^nxG@w&RH|!(8mlpYg?r225I9PGRqgn{-uhF z{6p$-1&T^qxKANES$n*XvQ>m;d}wqyw;RX&~X6{ZN+bxih4XGpvV+ zk0L>~J2;)*3h|vv+wD=F*Oyqm$$3-3@hcT=iPidHhi#iGIdNkA;h`fOV?*OifK)}G z;9ds;5>?af4ZMAk6WzC+I~iwLhiQ_IsvGP5l#cB12>rsNzC@Y(xG0XRwq9@Dn? zOv{b!`f=9P*yliB!!lqsR#{wnkbs>@j8mPF!zZniem4~gfq%UIr1KKT+HSQ+=`|gm z{v*ue4gugZ{60`-hQ?Ze2$lV>4$pa`_yuNAr2W?`ty=V}`r$lOq!b-84Pr%n(Lcs# z5f%~#`(O55-emk6s6;YcEUE<<_dcdZJ|@6d8EE_El z>#)LFD)hWEYRWH-8R<643V-jh+oSkpB-Wb^b2>L+0MANd{oPJDJ_c89%&J$Cg>(Jo7&LZ#XHfyrX-IKszc5s330tfi5#{P=(`odi2GI zTK9fzV&(oGb;729RgH92m{zn-RFCBwDpI2FGWLojn&K-{5%a(V*?*Oc|6X3zJWsx7 z2H;Biud~s_7>i!mB2?LPC<1kuLN8^nD`!4ZG2*L6gCCOH z{@-WSPAU9M+bqeYcS6q_Ea`&Vwk&bku}Fw$b$7j9FhkqsxTDe|L=SfFXxcp~Ueldw z=Zo4TnWBDknH>8|&InL*^m3(z@Q`jxsRLYLt5KItfFH`oKTS>qy~*u_nF{y|bN7yb zmEdrPSxgNkF`Jq{>gv%Q$!JJD-6R-(B;-kTJG~_s_Z#olXaaW#ysm-~ubgi~fYY^I zi6$0+6$KVZkfL8`UnOPvtM=D4Pm1HU@&?$jaUZ73NpaS z*>5c9GHE9RJ#jh9g1Jp8(b67$96*KO2Hm|}Yt!DG*zS64v(79zrTLa<3xP3NyD}v} z1}l+(^dAI<03{GhK^6CD~^Xwe~%#y(3?IIV{?{Dy2Bul4j zFv1X2aKXJ5dew1kXpFM6#@Tc9@tDo;WOGadlxvPVwh((@URT+i=Lzk^HDvQusTCmB6c#*cx z+0m}PV}=nhyL`KheQsctCOPvLH(L4|PJA;jXTaTU8kEU4O!WGu;%>*EUamLf3)PB( zIx|MT>yNh5{TGM!ELpw-wMek}b!SY;yyvKxB@%osu8$MmSc2Bibr=Q$q6d$fB|z~1PiwX+1kuLcpY zR${XMWP%+|E{RGvwak1u0`2^O*wNd$(eZxj%;r@-@9|lM9?{UEG3gMY!F-E zj+QKBL(g23whQ{EA*8`s5Qdicc$N!r#_g~n{C^ufqMeol$qSPh(RLaT_@l(ju@AVH zDV9EihRw^YYT4^sfbG@OLW0&zQD9P(&{ua)JO9DZ5rEylEW`-DCkgIV)phiyg5<;p z8szw5!>80qh0O{kAo#!=m&SN$ZEzU@K%Fyp%4~J*S%3^h#%P4^aiayBe_z`pJIvr@ zefGI&0G)KJ(3{mVMaZM|9z);t@DK*XxJ1z)GtNJ91Sa(s+TVaMnJFGM8r`v5 zYHP1F3oP-kY>Irw$YNmlCD(Ol40QfJ4i`56@e2i}47lYtJK#j4+!cri;}ko2{I%vI zQ#eXi$drYN`ymL!)fF-`K8$@3OAq&W1Xk~Y^zvQ=J4E0BBmxSs^;#*d@8g>=b!O+p z8be`=0#B6aAHcB$P9ZRKCJ^CEJ+7$umO{H-39V^Dto? zAh0f;d?`f5@mahBejvi)$`B_ec`PwPs=R15vH#^zDl`wHvm{7gVN$bdL5^cDcq_Td z80BV^G~XiS^LzY~{mgyrx*(Tw^q)|@10`IN`*8t#A{;(JPJ!0#^fJn2QDMhsAkyj^ zVYWe$6l|R%Vc0jBKl*Uo?On}T+t}nvnA-L|E6{8A4JCl|x4cRAxgkD^*voKy)b{q* zbC@R<#C_8f<~?>0H@n`fG$UUr)!`i2NExNQm^h&XJ6gqTQc4&LZjT5aWnah~uFq^cZtL*!9ga91sH% zY420=Ief6dOabK(avV8hna?87e}ZmA+sSbyETvI8U!uj4+6 zyyB>U6f~e1eozPQ^GbI{3~kQ^BAj9hO37#_Cf`o}6q1S~E^zVky)Fke=WTAGXH0Aw z&Lyd^>ksotO?ERd4*|{ivj+A6PaeJD2qSZuPZVv&=Ju(j6sbgjurkO#?w9ooje=qc zI>Q-x1nci0x3119FZ)t-t|0eX>mBL<(2c6WME+|aoIlVvNqbZ2;*C#q_91&x|9jlN zSSpH7N)$fElpAN-9NVan44u3Mx}-UoMen|odlgn(G{*D>^v}6+RXRe zDQU~w^72Z&)sTcm+vlcH&XZd3#v~)}Rqo%=;}N?5_eZycu2<<X?OyfiTA+Bs7a_$@OrE)wQ@-X5O3L*8F@WDRWnmme1fYdI7lAZ_ zRBAso9!IolboH}#r=(mb;rqjaqSUfT6Wz0~oXZ(87B?`&wc$?_4LnZ?%U?D0x-nUQ zl^xv_rs|Unes}`-07y){z6|vu|NTucH0$o30Dq$I1bs(`PgJ6&IC9b%nLm;8-#d7N z`MQ>Hc6L;k((~2eJ(R$HZx?-Q$=v5DSt5y_T#^-cV7R7VNhF*l}!$(XydLk@SzGa8VCc%afnoL6YLP3Z^#d9q8PWZ0CJrt zBsua}TZUFWnMW%pX*6a3`A?sR{!-$=D00V1#E9Q^Lxo7+Yup8FMIJu%{5{4O?`XGt zRxz)t&1N6WJ;@SmzCZL&ONXRGj(??=duLDn1je(@wh6VV;eI(?FPFE>R+&0e&=M1m$FFMZM$xG(WepBp@`84YzG&G>E-<21EH!ia$i;Fp3Rbjpd3@x97^;@HZVX zU`*Iosui^faBRjy;r1^-$AU*ksRL3|-XrNhZ{uUoQ4An-+^+xh#^-C$T5KZGy5GEM z)%AFbkoYJsIc2X*hLV%i?$lH|a`L90HGj>!Di%YNiI5sLfwP^1A8yD<*T*IPXf!A~ zKg}i`c-10X0#H_q1eoo*99JXop#Qe>}U%sfVmG6tWa! zW|R)(7I6*pN=NWT0U)+TXaE;L%|(%3yG~g&|KtI#(mwZPgiDpP&9n=~LKIJVnP>Y)-2%kUYf;a1iaRk5&Jjs>4zu*>H1;(85 z0>GUGK>G=iZWD_{t*1Sq3!edv0428I&H*E4+<0Q0fFCw^W6zmS1!&86Z~=~ABnH8P z<_v4wi_MZ+zaI}1fB7g*9b&?26hK4XCdN?@LGCC1jIq-_``CfM72!ul)=iRWnWB(K zobTLNogUD`OC9!}8*^3CI(2)Y`em6twj{g3!~5h61{e@6+rc~EuMGTsQ&IeUsKGrW z5VONiRE>L=r9<=A7DPbpV^x>zXAk0T5qtW2eD~m4(XBrRRk6FfU+A6EmYP=7Sp|BQtvG4W{Or1Z>BLCs%NY7M*dqm*F2kjv8nR^rX%&%uU!GfJb+e^Qo z#Y>bw9J+e67CG41Tt>jy_zDGPJ9@)_zjP*mL!4{okI$;em0}z4QFHc*C8A#%S$A^T z0kwA8mY6XS9+&&OcDH%a=E(TL8;=)lQ0odnGkGtodP^DLdsFYu0h`(CrtqW%H+Dxb zI1XIk+RO5kzB=phijg)Tq zw3|cPkEf1bG6AZ=D6J?fW|b3)zC15MN9VI8s#_+@s8;qSXeUXwMb37CssvLcZtseZ z9A;E><0i##+;1;;{Y^;$x$S!98*;+YveFnjj#p+8egM8vIHYX5DGQZF>HsmEVEc^p2T0CyO+^SiMcSi^{+W1SG+@;vI~T5NWqEGJ9X zc#Xg_w?ct|k2i*p>p)m|D^L$fnOQ+8VV1u&ir#7R=!QkD1T*+#*axW6)e@wsq0k{| zagdFMr1v;Yux=L`NjmPm3^c@zU1?lIW8V1>1_KanbIH&-Ytd9DJ zCIT3AKLPcFG)W7A$hSs=?>G@jeq{muFN$0BosQ;KZyjkj{7=mnj^4})K@D&AGWJSM zzb2Y3AswrOQ~PNGuMYd2>URob$b}8wxM^QZN!k(ggcgM#W7X4~D@2}@{DHzyJU4jV zsIQH#=1!801#th|nTVS_tp>#E7v{#ABP;S*=f@^?0Dvflm7B*1^cVEYsI ztg{VQ_sV2q$@jAacUxH4cmFq>uooIUNdR7&(*`)#R&a+pa| zWKTKP_E$-zUb{sM3HgjR{58X&@QRFiBZjpKky1mcgn_BneG8wTA2H#-)fT--$u6D- zAi&J2`qDUHhxUP!q;5EmgFjSxk&mj|IrQtZP7M#8OX}H)ga6G2esHg@kG#u&5lHaA z8{T=c%|5d=n!c22`2_Mn9(N8}R(bjJqd(ZP@Adg?_;~&i1EEd3*OE1b^ARom2KOG^ zkDiCRX{2@r!Z!HIw^jg-a!NeDzO<8bwpHQ#i3eN~R>rxu^P}6m8tIw_w2b-X>WkT$ z(!abusr)K{;A~YZmnFRqdgm07i}r}Xe)V<*p|AgTC(Uu|))c?1Ns0TzTwZ@in|jC} z+zS)Q=U4+sjg&vFD=T+IB=jJ9I@#}-Xnq^@{`*3f#90<1>iN!lx5fDUj8%2t=HGaEz%^f!dylQU4G|U54Od3_y+}{d((5)Znd-IgJ z+}0>kS`EIdOPV}13%%hLdPg17O`a*iPZ%3Cm?M(JV;zeGOHuV)jxZFqG!!ZK9$B;= z=R8@|f3?HR+00kn^?r3tS$tqh?g0frlCNehv&m4Xop?N+O(Xws-8>!;(4P&u;1%9O zOkihD*N@q#urDR~fTF}i9Z&1jS-|(f>A!cVJAt3l-XUF=fbU1ItR48ICL|*h0Yz46 z3a%?I(noCuqVl%Q!3#;d2OsSA^k=eE6n|6=yRtjl|J%7}^jT_d{nX8Fo}T$`eJU zYFqfvPIKB;7U(Zd9HPB)K3c+H**H8&kh-A`z)Xm^CF!#_mc;}T=Em=DjC@axNE>5A z^xo=R`8vG)TBdf~vA~Rk&IqiMg4&^o&X&@_Ndt=JH1EghVAp-In&ycqivXEHe5T=F z^{-lgmX3%V=`Fu9CeL|u5cRq7zX(-|=gW<#D^zY7fob4Cm*mjSuJS@7D==Z7UP3F{ zt;%v+U8e!$ome0=s$6jx=lhEXyWD;#--8oBMU^@6-(XgLf2c)@zKDR6@RqgU42j5W zaSuto`nP#JXQ`q~nVG@;p9JLV$;A|SN8bMpbu~V7;q*quEs(pWyr2>Y zjkmDfd84TqES0kFh>56;{d9ilZ08UwCH5hK=hmUS<1LJBC+n)h?pbwdJqZJR%f`!T zl8zSo@#9kFvzGfm;Vi?d_ z(L>rN0YTi;4%~u?I!~^}zL)&W4$0In#@d;CRCvgx z4HDOQ9lNMk(Sw&1Pot^0t$pwH0r#7MbK(8j(uEvtWR)2xAx``V(1!QPe*oh_B(pOOS8+pm&(A;+ zkcMX~KMEI_AV#Vs)d5|YrumV{{V%gY{fJ_D$&}sRHpo4E#?sWUfSi#+fCAKSa!~d& zX7IW(6$MXU1<35n;gWP>sdI(2dLQaVuoY3D$|Z^U4Dc1@feH4c$qI&NR{r>s;gv-^ zL!LHL1H8ItkB5t1Q!!vI<~7~B#xI8cpTPPr>r?#~)cPnqsJFkxiZV`4)R5a3zn01O zipIWH8v0Kf)A@F*B@NuixSBN{&NQ#9(5T-#vKlDtFA6u+*H5cvCNo*OF%f2%F5esljO=LfZkgalt+pAvQe4PF|sk+IB6hnp?gzI_`ZPQw#%(U;h z=d6>p@hULMpwFegEMYKZId$T-!XfguRvMMaTg5j#56{n1cpL@?4ez67EZshEBcT(M z&}j>!qlg9OW`89EOON6sf8D3GEI?xqnX`r?yz@iP*!lm68}Chez;2= zP}NmhytV?OKstgi8Ucn6u%uVf9f?yBYc8^u{Ivw&^zV>NF{<@PjnZ24KUU-vs~ z9r1H?b*jR@=`OcJIYNQRHcxD5!YdntuEHu5r|m$ox%;R4}GBiWmiJ~HLet;y_1+_z!SP?1#=>Y z=ZYcteg7D7&mMQ5jj%M=m1|wU2j&ZJM+aPnfE*GOo4{j4dF zs$`%?jS4)1NMZo3^0^~LrxEU!rjOa)SEc0D7A8w&eYn$q6_3UO$oBS-T?>Rjx*FxD zYE{hF2>@LAnbGx+EQmL9YQhz%qpEl-@|6?#qS&$^urUhWVQ2e{DijrITFS0QRtPuq zPE9(N1JpQq<9EPj5qR9}yUG0X-=m+fY4`rj+qDV3ulUqD<@KJz+7q3I%RhzS=x?E* zt@SAW#+#lo{+;zW<2#EyE7Orgu|! z;vwcn@zoRJN*SzYQNt?Hs_hb!KwY8Wc3#BtMW-3Dv`Hw$z6#mUd8>m2Ak?otDK6cc zu2AEu54|g-;MYUp%L#`zl=Q!*19N_iQI6m8V(1QhED6d>m%PgAG-wFH71Ka{tvAZM zTn@cIk=kVqoo};pV__Ty70f~TMvKtwkx$rxmj@tyFjdfD3-yLRDfeRNCq^L^xIa$n zU_g$s8JnuOK)Nd6L1_&cYIQRvWPEtPE8h4?4fm%?!9L{L$U>=6llxlo3)cX=nu8&x zGl9C!JtGu<4SFK~69txE);Xr@u^RAzWZ2d0UWEbxhGjJ;#Y|xug1i!1nAQ)|4*0p` ziL3?}4k$aOy|WQ^9qZ)Tw-qt?og~~T@aL{|WH&V0)WL=wM#E|!1Nk(8m1^Fn*#Phz z{WGFsY`s3H;c|~;)kHqGdKoyT#r;j}kFAgCbU;B5b-z16<3&C8qy9vpPPg^JbI=7P z`fI&Uc49LNkZNvKL#gAXrzhrLqm@Wfmc@@NntcBG;f$A2>a}Ch&YmGZ9(BMB7*hs} zcRHy3szuFC0&w0fuVTL}sCe2(fB(f6|KnqC;f!?wp5&ig(hui9ni z_v=X<_OFYY@|~Rtz$7mvsVH`^SwVc{15Q#r^z>j^=UrQuiZC$%z*fK*7l~mIm`i{4 z%W+E^`(P^)N`NYvx9;_pChm`|7Z%Af`qR6B@ol%_(GqL|$k~7Jj5t6DM3fG7FCrOYek+1$ zL>&nIxn=>>t6kuaX{c_zv6-vmBY4pKz&PISquJzS>2k4@Ur-6X4gwJ(MuSZ1@|+Rl zV``wHQb??3CzB^JY{>>PIT^l#RMhduBE?~qZg7UV`I|<-wGiyK1tZeXiB|R88^wXk zg7vl08uQaTaP`Xeb`+%Q7<6>fMkl^J`|oS!q8MB~6P$(ip{V(LMii$qmQNHp z!NWPFLyE3cs=nbijFu~H^AD^L-sy7P_p)m2{|U_>p`9_%caSx(=~sJ%f=|0=85it_#I^@xSF9mJl}76P;v{&{}Iz^z|*=T5tDZrXJ#t#+gP z&Uu^=5c$FIBBLAQwk1{bIZuTphUOnA`YNhgmyhW9b)V0^EoA*4O*=IB815e>8Jm8e zSYrx&ZnuHg{3koZTH*63@UfapGRYkx+Sn2P-tJbS3777fT6BbB2%9GgJfV^W+_{O5 z_~|Y7b`?6U9{D9+>=^<#ZvofRP_)l&?&~RPg~ee>5EK3oKo63528b5LOY?C6pdanr zZ;82J0memXm_J|=SnwTE*}osDbqJBJuuz=56tqNvCeN}X;BPsjZjKQd9xo410CCS? zc$A$7x%F^$O9g$vHNH8P`ndZ@lHmuJmuv&1^;kfsa8EH`>`5FwNm(1;ji7IZfpVKi ze2G`Z<0QWgz-09fd1VL}_g=#2_fYca{j2PvPqFy!f25iD9x5$s?vGTg5zWfY1A^%T zZjfjSE5S$gADw;pOIT$|+uM}EJS_ILvNT?+i{ z5E8GL2NA+%t|&nI77FYj`3(lFd07d+H~c$J>4;MNv*eVHO&`Al95#VxVLMl0yIEAF z%=!|OH5gnKfvN5oqU&tLEMUW!ub*rH`ClFzZ~2icHUOw{x+P3s6nPzS!i^N4(8tZ5 zToc06B&H^aL<_%^T67%2Q6}c5>$O|JNY)TtJV2*8jZCp`MFLd-B2(Ihl@)x$ZszG`1p=FVtw>)9=xGM%*B+Kdo^oRyZW5!j1z7K=DtExMwq_6!A|q-prxyw@+J z5wHI`2T|lk8imeYSkYx^Cbkg61L#5UZbwE;TVhY|RNN)8si-q7dz}q`;yTePy1$f3 z7FeI;{#MAUU!cm1Ac8M4MT)M!9b82>qZOM@hnQJbg=w?|WA{+CsrgH|HQZvF$w$W2zO16x*Vy-tf@ zxzP6)3B0B&Ms&UDk-O)zm4?zVzzn6m^6%!U2;4h)uvI{3V<#Y_an_R;CVwSP+f?wN z{Ymt0@ySH7nJrG$NLy1mx&yngy{xQkc5_r!-WGICg9`y1nwswuNg~17D1KjJ2?n%W z5D#3JMZ?|v3Rv(tEDmcl9Uo=<-8ihI%9I@HKX<}YK8h#g%qB>|_~#%02Uly2hcA}z z-{WtBay$NP`z4mK2w0oALkOFDdIseKeBs6PKL26&xtjC5Y%8()tmxt3(mVO74aqHJ zAQ`)&ZfKG_O87xHVW6=BtKJ)Tg#qM3oktX`1W@4I-o^RZ2rsU>8#v6Q+1t1CJe>x_ z0WY4ftYl!Wc}ek=EdsqRJw^&+Vz(2{3YBas6f;6p^Rqa~Nzm;}0b-i_A(y1ahjW;e z5<@3Wy`5V>NqdYE*M$ixMZ+K3mBf!MGwP-NJ8(oYU$q)t6RKelxo41}7Y|uorlNbq znDf7e->8GPeqaDuwTQ%mXYV8Kp3k!aQuRRO{&?8PVu^41Pqf2A!jl%6Bz84zQ3C@%bX}(_OA&MRGqjK54eYkpeczdlUs`WY)>5Hn z{{{tLTHR+0Kaa+K$QX=7#=-2~dXolNJW;_mLo{OGzsI)96qhT+^ZA!-IKR4GQQF^5 zFgSe{Ad1K}!?AvF-g&|z!Xs#9na5_cp?;CT;kT^NAXg!^1~2z&nUkqvGeVLIq-(uO z6@pH@^z-8hCowHbQ6Uz~v-BpsBm{)&iOwyzJd6M>C5#JI`Ih%uN@Z8W8SiJtJv(QR z&i!`(0;+J2mkYK5HKJ-|8nh;<~F+KEI>tkg(M^6J28R4e7O)rni&-4P)=c0h8+-> z%8z;HS%6n8fMxS(?te#+vN9A70{PLpNpE`>`K@4cc_9m^r63k8YxnO%BUv~-Xxn2F zI8Ms7El2F;2B6&{{}El`Gw!#Relo-58s$&(j+C)XT97LGYS%lvlxlsWeg*at1MlWe zx1+RgOY0s;3-7Td^SRGBNTw=p@+x)E6y{D*$!28YsAHniS^vcm0j*{X)~+Q-WL!bs zGf<%44=Eee3V3FJV<>XH2Ay-XwWY}DKvwMp_nr4{{%fAxL9Jn3e>&QgO_~F51rFV6 z(HoaQuBnJ9=yl8!(X9qOsPQ!55o@KY>NH5tSijP1(DZ$z zffY9n69b)XM1^^>(?kGZKYOF1Z>%Yz`aBN4Ysi7XVPC)o;98_|?uoO%rKY_6^Jo_o zJHh*B$~;ziG_p+k0(x<%)t7=lp^J=DoS(UJNozLtNi$eyS-#_80^&c_r#q4oJm&B) z=D2|V@IJm*uz%>Ruxx@SRd)SQAK!O`YiP0wg!CbYGyZz;qjH47M^DjhVe>Nz3==Sr z?y=!8bRp&K$dIqbG}VJB>R=IO^BZmq?k?-rd>!5*>hoHD%89Jd<&cO-If8PY(JnN& zZKU3Wy?7zL;H{q8v}SBQhDp#su8=*ZsQ3TX{Qlapb&R#sa_}lb9^8bJwc6 z1C}{T3?3|d9F|=_4gYO9I{!HXcxCB~xKXW2*Z5Vkl^A&X?v1=C0vj+4TPaxO{KRc* z-%lq!+*R~j)oTD=zY%z*mJm?#vtkfh=?qhSJ?br!splQD)_BzHJXd!OE>NY#${rxw zkO>8-u~E)yK3-UTJ4}Q6P08te@WN?7e6%C&6a$Bny|G&59ST>?8XnY5BmycvDV~M_ z?^tljjl(Ftoo;q(iTfU+Q&e6*s7^S%Ss4{CxXXk8|fMyoor7ebia-CU{WJg-8R?{CfwJC2ribb5+}N&(&q@uGip&1o?-N1ug&1Em2ji7q z&G>h}+^3BUlZ}1|L+#@l+|hF~7m0zvcyQe6=6i+q(k`pnpAz=WiWLj*^;Fw7U6G48 zkkQ{O50MUM3d~;E%FC!%@F#U#`G3xzADKI%X<>BwM{m+M$L{cQp7RS2>j4VzF>`}D z!1%vRTc_45@SCz>V1H5mP#c}_=~!!neo%t+j|YC*`_^;m)}HV2@&DZt!=a+y2M9dF z3ypm;HOS{ds#1doy8#m_ZriubW?NEnuzjGOZ(lnBt&PQ^#6uao3l!&49Y3PSL99An3`! z(}KCe%|-mt^t}+7G(v31a2ECWyJ^G2C0Dnbx)F25f3x7MN;95(G!{HmcJuGJV%7b1 zOU~XSk~=r2?PxyS?I}t0U7oIYm|B_3asa9A2W{lDr~b|Dz9*+6tOw`MoU53R&`7eT zZ&UPDUad8mzG_;;`>#zvQE~C79=dbO#s&(p7h`&18WoQZi>}KyM@vz1CVi1|@*+Ot zSuVt=k_1M$a1Dw8neU8svTC-{g0tr@IYoAyUx@xSYXGY9f#Yrvf>R>ZLZy~b%J zQQ>-Mx;}(&CDoEpzzI_i*~N~F5JDPF>}K@oVOPmk=%1VUWO}XF?c^tJIqCH7@c9n6 zEw?PAUKu@yhkDudr21;Rx-KpY_}Gp!v*=%=fN`o$%YtX@rnKX5EfF^WAkq9J{y{Zn zyWR(PmaAG9T#u(lTIg=3r>!EV84n=X+tXK~IX$b(pUz1O3FqE%L~NB;5S^a0^Qgbs zNCYjL;+NgeY(H=7K?TXQ2Nc&u58KXyE^iUly^B@!&~cU;4Pa!@x{5H z5>j7d^SYK_{0DY%KVD~{jyomQJnI|(WkMeNt^o)>DIPDa-8%lN&t5z&pHlPM=zGDl z&p*>fXpYUQt8T_SZ?_POPvelf>W>hvu4syt zTvQDkW)R>UZ!tvWo)E~iKjTME&;4M~@_8eb-&XiFhwR^RU)8S+a1xMM8)qZMXwGF- z$9}~h#R@rOe=XFX`9gotb|}X|ZVBYqiaIJ2YO*0I$gMMruN}hCuh=RK9vYf92E!J; z)g#lcZ}+J8q=8w-Zy(lw!gx|IEx(s2Xe+UR>^7~@^Wr!=AA)m?(uP-uDmM9&KNS1Ihr?)D&4-GoNb(zqJbI z3}?9|QUGlzUEm}Lxdf&E8+0488keo~5PSPpLNf3RxjpB!L~ODZ&@~k}1~|@MY&4B>|`bmKwHBvs;g&?d7X}n>YVN zAY~^YcZqzBFINe0{s8(J%_TZ#PZ3DpcTB^HfS_a$*N6u(z0x5wp8vS*N#Y()*AlD_ zLEW>srFZql8f;n}cQz2Nu4#}|>c0PP*kU99M8_2=8N+mDM<>RB zr%UQiAUh?vIX;3fK5oOuSWw@A&cdwu^tUx~d}QD5Os0h|0+kxt!v7yqnTYd+v)4hY z*Q9Rew8tj6j_cdUPCLN!aKDmcfZGP0x#e#Yms_>vHzOUU zC@XxRE_b3uXXs54E^w9A`vR85flk`iRP=aLeJH{aq{{mTdQ1pP25X;@P(DQxTO? zg9N*2F`Kb($P7!U_a}gu0JCYnSVN;RxKHA6{5jNK>v!Hu>sjH=o`OJ6)y+6dWgx%U zWguRNyFk8myr#?&ndjGiqa&&C*@8Q}2wlUZ^}gRHZpi*As9fL@OXq>8kw4tavv!^Q z-20$hUUX7Jw<8TKSfMZh)~U`N{0com@%QiBmr-l2hW&ZO+J?&i^6%~UHW;2-17E|n zuS3m|4aHHHUi`JS5ODAcsgT!{*JdX?R8NZF%?IqWU51joK^}P|drq%_awcTxvbx4D zlSlpjkF7RNjV&tmjBY?>04AC+t~glp?F?^yC&}Hj#Io0(S2BJQN(5j|eycIx0#V=R zkMvjyvWM!4tnax+Na(kJsbHrP;;DLx1|9~(b$bV-F@ehx{X;vqc@XS#!}94KPQM^b z1N%QdPxr*Vlg|=xs7Eg{+pNB^WIbw&GY&TllT zs~x0rL$c{UYvdTEC*sdyFCA++1PP-8G+*#h^(rv76g|%az-werO=VBcplcT1v3uMY zL1*%;PLUVI>2SmC)mt-JHdfX0;)}&NrNWXD0}fNRW9~2vle+@0(FmJbCdR0S_5Gn$ zRVpwa$E$7n{S6bmqXsrz18)1j`m(#kn!2?2R+Pe{zPG54lIj8@R(rVadn_hhF}bgo z{hGc}%tt@x95drx*VD&L>1s(e_HvMG|1DAlE#V(>YAn(>7b+oP+!UY0+? zSL2JE*Q}6-wa{@RI1r|BPWINa+KyHGl>f=}7SLHT``IVd>dn_0g)chL@ZtA2gh2^8 zLvi)Oxz1qiKBod2{^`a>x$6X&=%*emp5{Jt3>b*uw@{hij*^k?n8l={%;t(K5)lwO zIEp`qu?&<2YXysmg3BD{o$^iDG&siGjrgaw9G1y!ev~L2Pmh`_0!&Sc{tx=H!R0$z z!K^$fC|1GaD;snmh$X_Dx6Gs_79P(Cv)l|{|43{{=0kA6eN4VuW}Jc}&HG4+VmUbh z-8+^V)}_-fg-YkU8km_?Aa18iPWpDkNTXhq5!dd<=bZn`$pGecaD$( zl2G_%ix3a4$YmP(rH@6bY!juT|L{0Jn?4!t&<*v@0_+<& zowl}wA>z$E=agezETA76TD0*F+RtIPY?I*HO1|5ycH8|>X0ZfJWUK`Rg3YdIT7J*|N2yjkjH;rV!40oz;r(*JsKaz4QU7(`hcWW z-w%Qo9-+kL(RLae8P7tg=4>z3^Z=+Tzz1(l((TL9meBiv5%bj-<99US(5%AhTw`fu zWvK~+M3_F-9ys7~w!l=Y)vz|MZYQ1@vx(^ah?n-Fc2g5{=nw?dpgT( zgr4@vyC6v1-+Xnt5Lv=l9@aFg5jnf)L@t`@6}A}uqFP{1Q2E#)RZDBogtpB{@Mh(1 zz5s%L=gXJNv2XcX<)WNTu;%f`2Jucfjl!+w!u9L@7*Kj4;;DQ#6rj1t zGk5cq({Ly;2qH$mRMjURB@bL|B{=2bYo=KjOg& zZvaL3thAfC+jV|T$-geajkFQm-zx0I*G&>F*WzU$kvq-^29l8vBAvFPw-=L6(rDYQ zx_=?Do1QZ*J z3eB{tZ&v)8Z|0ojZ_(}q@N{?20zxJ)AhWc6-ob7DJ3~xFn#P-TgX0I0v8~Tpwq4*$gCIoA;vOIf9h+oF$KwEfxCK*KbA#x`Kq(Gi1s1Ku-K` zw+Qq?*k&d0V4By(k}~@FUI)=uCHnm<8~q=zQ>cv{FdRIs0<)uwcmm1y(a!8&cRs?x z_=tMa#yJ|N)djgyZGoeaH$P%VVfc6S;Drz%6ZHtxC7i?p<1u`ONuEoITU&-XAQJ6I ztLIgkyy+VMTOHmkHLY)5AEEh z#I=;qnzs`t`9V}}^jT({yPs;(gcm}1n&aQsHdHk%x@2Utd}98Nl&b%Dt@JxtN=JW10t8_2R*%v5_}*j&zE#>*u?192)N#G+_FZW5 zcGbQtP=7u^cTc}N>}42jXiR=m^^vD<^~7CU0>b%3#arI1VcRNor5KpkSQ@OZeuZBT zIMs*-fD$&(e^n#ZIWwoqIEF-C=3{?WYUi*n^hA+MTa2c*axoea(A5y)Q2-*t!#;Xq z=&WhPv04Sn-gooSJtEI}@mN|C3yNb|);^}sl?wz6Cgz9zE6A1M@x{x@7`}5F&lyu9 zkfhz@?JfceWHYTw3zb7>#1S8Yw#KBTs?jeDW{T4n**Ps(!`CEb<0eHJo+;`Bt@w^Z zk_xn=4bg#t;2*_bx9SQCWteg0d(2K&DV(q#$xL#c&Mi+-A&`+0Tx~8NKX`luM_mAZ zZM{mm`NnyV{et-wAq*S$D@=+blB{Jn#e=DLuE}5OX49z!O=4tX)cTzt)BE^auiVa$@{> zyU&n^AoV0i>9;_6hLg$Y?dHC}J4!@u=fR|S1`2sahK6I%uB-)j^on_kwRrgUm=QIhX__LcYX*+n<|s5{BykKGq;~|7K9eJc9jP@%pYCb+g?yn z@d{qZbj)_-zSJ`rLJqStL2D;W4=-0Z5hUC1;Km($ni*e@`D8;~t-hhvC`_ zFCUDab3cMtzoT4K%-m8gs0)bAH(C!&PWwgL-ye3w@0(mYJUNlRn4!Ulvj6AO57V?v zwHH+_Z{D_VY)tfzTfD8UkLvk7*2(Lh)R~C!JbI`UFox{C18LnLeR%$Ja~Dxo3^X@h zIbY1QI_PzseM4bEjWj}VbexzxvBG)dJ7}VVrMo=}$ynBvRbX;cqOW>|c{iiwILDM! zQiR02*0prCu_02O=6@z!wTso<-m}|9f6HcKvDejb{5L56_&1&PoQ->;iEvKMhl9uQ zA^1l?ZRgaSPoT%cNl4won_Kcm_Acd9F?`th`g~e>@~byJJtxu19hZ^H`c5yoc3?cp$?mK9gvp1a26Cd1Uq zux!-Y_-50@OL>Mz@Iol^pYadR_oh~fjr9Q90%E`)E0fjC37cMaUp-nh|tqXA;3^ks*_Vm=g5G>%N5MG zsK(q4e*v{PbW?1(Ukub-Zdz*so~i_<(><~r?xZm$%qoSQ_s3M zDsXPC*>7vFt=O`uTD*7#!$lqmcUjmx6!kXX?TJ6)~s3bp?C^vT{}CCZ>2N_HE9397Ulg z*;K7>L;#Dzw+D^}32T~s!qCJi3sM5p%O1toW{LThpsYB;x|ex_t#$8Sx1S$nyx#g0 zXo4&#vA5pxOC?MwkHE^jtz0nt&S;Zu9WrJi@5;h{cif?)CPxbC9Y80;&3{@T z+Hclh&OwF#AzFL%(nKJ>TEFi}pL13F45p0ikUz0q-Sj;vI>%)gZF`st{&UwEbQfVj zy@@>(%RKT0t-X}8mu*w)GffIEdfrnWKWj_a=8Mg_OL#kowq;ed?$l>QX+`-wIkEbt z&hm}FCx&lWLCcwW=0;ZclK+KkEzv!T&>G9^ik8^5 zjzRyHpKn}e`WaI-&Q}{E-LyVJ7>CvE`Clz@&D6JwF-s=IFBD&KY>7syw&h&CV!pU} zwEIUlo|1U06D%IV@V5_7d+47RwXMGCt7>`eRgKe|Ab~OE5oP~tCsD#a1UsR2^;92( zTcTBrLW7>%y|-SzNmYm5uNvZY Date: Thu, 14 Mar 2024 14:07:22 +0700 Subject: [PATCH 25/37] feat: Nitro-Tensorrt-LLM Extension (#2280) * feat: tensorrt-llm-extension * fix: loading * feat: add download tensorrt llm runner Signed-off-by: James * feat: update to rollupjs instead of webpack for monitoring extension Signed-off-by: James * feat: move update nvidia info to monitor extension Signed-off-by: James * allow download tensorrt Signed-off-by: James * update Signed-off-by: James * allow download tensor rt based on gpu setting Signed-off-by: James * update downloaded models Signed-off-by: James * feat: add extension compatibility * dynamic tensor rt engines Signed-off-by: James * update models Signed-off-by: James * chore: remove ts-ignore * feat: getting installation state from extension Signed-off-by: James * chore: adding type for decompress Signed-off-by: James * feat: update according Louis's comment Signed-off-by: James * feat: add progress for installing extension Signed-off-by: James * chore: remove args from extension installation * fix: model download does not work properly * fix: do not allow user to stop tensorrtllm inference * fix: extension installed style * fix: download tensorrt does not update state Signed-off-by: James * chore: replace int4 by fl16 * feat: modal for installing extension Signed-off-by: James * fix: start download immediately after press install Signed-off-by: James * fix: error switching between engines * feat: rename inference provider to ai engine and refactor to core * fix: missing ulid * fix: core bundler * feat: add cancel extension installing Signed-off-by: James * remove mocking for mac Signed-off-by: James * fix: show models only when extension is ready * add tensorrt badge for model Signed-off-by: James * fix: copy * fix: add compatible check (#2342) * fix: add compatible check Signed-off-by: James * fix: copy * fix: font * fix: copy * fix: broken monitoring extension * chore: bump engine * fix: copy * fix: model copy * fix: copy * fix: model json --------- Signed-off-by: James Co-authored-by: James Co-authored-by: Louis * fix: vulkan support * fix: installation button padding * fix: empty script * fix: remove hard code string --------- Signed-off-by: James Co-authored-by: James Co-authored-by: NamH --- .gitignore | 20 +- core/package.json | 12 +- core/rollup.config.ts | 2 +- core/src/api/index.ts | 3 + core/src/core.ts | 36 +- core/src/extension.ts | 44 +++ core/src/extensions/ai-engines/AIEngine.ts | 60 ++++ .../extensions/ai-engines/LocalOAIEngine.ts | 63 ++++ core/src/extensions/ai-engines/OAIEngine.ts | 116 +++++++ core/src/extensions/ai-engines/helpers/sse.ts | 67 ++++ core/src/extensions/ai-engines/index.ts | 3 + core/src/extensions/index.ts | 5 + core/src/extensions/model.ts | 3 +- core/src/extensions/monitoring.ts | 3 +- core/src/node/api/processors/download.ts | 36 +- core/src/node/api/restful/helper/builder.ts | 50 +-- core/src/types/file/index.ts | 35 +- .../miscellaneous/fileDownloadRequest.ts | 8 + core/src/types/miscellaneous/index.ts | 2 + core/src/types/miscellaneous/networkConfig.ts | 4 + .../types/miscellaneous/systemResourceInfo.ts | 28 ++ core/src/types/model/modelEntity.ts | 1 + core/src/types/model/modelInterface.ts | 7 +- core/src/types/monitoring/index.ts | 1 + core/src/types/monitoring/resourceInfo.ts | 6 + core/tsconfig.json | 4 +- extensions/huggingface-extension/src/index.ts | 7 +- .../inference-nitro-extension/download.bat | 2 +- .../inference-nitro-extension/package.json | 2 +- .../inference-nitro-extension/src/index.ts | 3 - .../src/node/accelerator.ts | 237 ------------- .../src/node/execute.ts | 9 +- .../src/node/index.ts | 20 +- extensions/model-extension/src/index.ts | 87 ++++- extensions/monitoring-extension/bin/.gitkeep | 0 extensions/monitoring-extension/download.bat | 2 + extensions/monitoring-extension/package.json | 29 +- .../monitoring-extension/rollup.config.ts | 68 ++++ .../src/@types/global.d.ts | 19 +- extensions/monitoring-extension/src/index.ts | 19 +- extensions/monitoring-extension/src/module.ts | 92 ----- .../monitoring-extension/src/node/index.ts | 317 ++++++++++++++++++ .../monitoring-extension/webpack.config.js | 35 -- extensions/tensorrt-llm-extension/README.md | 79 +++++ extensions/tensorrt-llm-extension/models.json | 49 +++ .../tensorrt-llm-extension/package.json | 75 +++++ .../tensorrt-llm-extension/rollup.config.ts | 73 ++++ .../src/@types/global.d.ts | 10 + .../tensorrt-llm-extension/src/index.ts | 147 ++++++++ .../tensorrt-llm-extension/src/node/index.ts | 191 +++++++++++ .../tensorrt-llm-extension/tsconfig.json | 20 ++ web/containers/DropdownListSidebar/index.tsx | 3 +- .../InstallingExtensionModal.tsx | 87 +++++ .../BottomBar/InstallingExtension/index.tsx | 52 +++ web/containers/Layout/BottomBar/index.tsx | 2 + web/containers/Layout/index.tsx | 3 + web/containers/Providers/EventListener.tsx | 59 +++- web/extension/ExtensionManager.ts | 4 +- web/helpers/atoms/Extension.atom.ts | 40 +++ web/hooks/useActiveModel.ts | 12 +- web/hooks/useDownloadModel.ts | 28 +- web/hooks/useDownloadState.ts | 202 +++++------ web/hooks/useGpuSetting.ts | 21 ++ web/next.config.js | 1 + .../ExploreModels/ExploreModelItem/index.tsx | 13 +- .../ExploreModelItemHeader/index.tsx | 18 + .../CoreExtensions/TensorRtExtensionItem.tsx | 225 +++++++++++++ web/screens/Settings/CoreExtensions/index.tsx | 110 +++--- web/services/appService.ts | 24 ++ web/services/coreService.ts | 7 +- web/types/index.d.ts | 1 + 71 files changed, 2497 insertions(+), 626 deletions(-) create mode 100644 core/src/extensions/ai-engines/AIEngine.ts create mode 100644 core/src/extensions/ai-engines/LocalOAIEngine.ts create mode 100644 core/src/extensions/ai-engines/OAIEngine.ts create mode 100644 core/src/extensions/ai-engines/helpers/sse.ts create mode 100644 core/src/extensions/ai-engines/index.ts create mode 100644 core/src/types/miscellaneous/fileDownloadRequest.ts create mode 100644 core/src/types/miscellaneous/networkConfig.ts create mode 100644 core/src/types/monitoring/resourceInfo.ts delete mode 100644 extensions/inference-nitro-extension/src/node/accelerator.ts create mode 100644 extensions/monitoring-extension/bin/.gitkeep create mode 100644 extensions/monitoring-extension/download.bat create mode 100644 extensions/monitoring-extension/rollup.config.ts delete mode 100644 extensions/monitoring-extension/src/module.ts create mode 100644 extensions/monitoring-extension/src/node/index.ts delete mode 100644 extensions/monitoring-extension/webpack.config.js create mode 100644 extensions/tensorrt-llm-extension/README.md create mode 100644 extensions/tensorrt-llm-extension/models.json create mode 100644 extensions/tensorrt-llm-extension/package.json create mode 100644 extensions/tensorrt-llm-extension/rollup.config.ts create mode 100644 extensions/tensorrt-llm-extension/src/@types/global.d.ts create mode 100644 extensions/tensorrt-llm-extension/src/index.ts create mode 100644 extensions/tensorrt-llm-extension/src/node/index.ts create mode 100644 extensions/tensorrt-llm-extension/tsconfig.json create mode 100644 web/containers/Layout/BottomBar/InstallingExtension/InstallingExtensionModal.tsx create mode 100644 web/containers/Layout/BottomBar/InstallingExtension/index.tsx create mode 100644 web/helpers/atoms/Extension.atom.ts create mode 100644 web/hooks/useGpuSetting.ts create mode 100644 web/screens/Settings/CoreExtensions/TensorRtExtensionItem.tsx create mode 100644 web/services/appService.ts diff --git a/.gitignore b/.gitignore index ae0691605..d9787d87b 100644 --- a/.gitignore +++ b/.gitignore @@ -22,16 +22,16 @@ package-lock.json core/lib/** # Nitro binary files -extensions/inference-nitro-extension/bin/*/nitro -extensions/inference-nitro-extension/bin/*/*.metal -extensions/inference-nitro-extension/bin/*/*.exe -extensions/inference-nitro-extension/bin/*/*.dll -extensions/inference-nitro-extension/bin/*/*.exp -extensions/inference-nitro-extension/bin/*/*.lib -extensions/inference-nitro-extension/bin/saved-* -extensions/inference-nitro-extension/bin/*.tar.gz -extensions/inference-nitro-extension/bin/vulkaninfoSDK.exe -extensions/inference-nitro-extension/bin/vulkaninfo +extensions/*-extension/bin/*/nitro +extensions/*-extension/bin/*/*.metal +extensions/*-extension/bin/*/*.exe +extensions/*-extension/bin/*/*.dll +extensions/*-extension/bin/*/*.exp +extensions/*-extension/bin/*/*.lib +extensions/*-extension/bin/saved-* +extensions/*-extension/bin/*.tar.gz +extensions/*-extension/bin/vulkaninfoSDK.exe +extensions/*-extension/bin/vulkaninfo # Turborepo diff --git a/core/package.json b/core/package.json index 2bf3e1735..2f4f6b576 100644 --- a/core/package.json +++ b/core/package.json @@ -45,11 +45,12 @@ "start": "rollup -c rollup.config.ts -w" }, "devDependencies": { - "jest": "^29.7.0", "@types/jest": "^29.5.12", "@types/node": "^12.0.2", - "eslint-plugin-jest": "^27.9.0", "eslint": "8.57.0", + "eslint-plugin-jest": "^27.9.0", + "jest": "^29.7.0", + "rimraf": "^3.0.2", "rollup": "^2.38.5", "rollup-plugin-commonjs": "^9.1.8", "rollup-plugin-json": "^3.1.0", @@ -58,7 +59,10 @@ "rollup-plugin-typescript2": "^0.36.0", "ts-jest": "^29.1.2", "tslib": "^2.6.2", - "typescript": "^5.3.3", - "rimraf": "^3.0.2" + "typescript": "^5.3.3" + }, + "dependencies": { + "rxjs": "^7.8.1", + "ulid": "^2.3.0" } } diff --git a/core/rollup.config.ts b/core/rollup.config.ts index ebea8e237..95305bf25 100644 --- a/core/rollup.config.ts +++ b/core/rollup.config.ts @@ -64,7 +64,7 @@ export default [ // Allow json resolution json(), // Compile TypeScript files - typescript({ useTsconfigDeclarationDir: true }), + typescript({ useTsconfigDeclarationDir: true, exclude: ['src/*.ts', 'src/extensions/**'] }), // Allow bundling cjs modules (unlike webpack, rollup doesn't understand cjs) commonjs(), // Allow node_modules resolution, so you can use 'external' to control diff --git a/core/src/api/index.ts b/core/src/api/index.ts index e62b49087..f97593934 100644 --- a/core/src/api/index.ts +++ b/core/src/api/index.ts @@ -33,6 +33,8 @@ export enum AppRoute { stopServer = 'stopServer', log = 'log', logServer = 'logServer', + systemInformations = 'systemInformations', + showToast = 'showToast', } export enum AppEvent { @@ -56,6 +58,7 @@ export enum DownloadEvent { onFileDownloadUpdate = 'onFileDownloadUpdate', onFileDownloadError = 'onFileDownloadError', onFileDownloadSuccess = 'onFileDownloadSuccess', + onFileUnzipSuccess = 'onFileUnzipSuccess', } export enum LocalImportModelEvent { diff --git a/core/src/core.ts b/core/src/core.ts index 6e2442c2b..b8cbd3162 100644 --- a/core/src/core.ts +++ b/core/src/core.ts @@ -1,4 +1,4 @@ -import { FileStat } from './types' +import { DownloadRequest, FileStat, NetworkConfig } from './types' /** * Execute a extension module function in main process @@ -17,18 +17,16 @@ const executeOnMain: (extension: string, method: string, ...args: any[]) => Prom /** * Downloads a file from a URL and saves it to the local file system. - * @param {string} url - The URL of the file to download. - * @param {string} fileName - The name to use for the downloaded file. - * @param {object} network - Optional object to specify proxy/whether to ignore SSL certificates. + * + * @param {DownloadRequest} downloadRequest - The request to download the file. + * @param {NetworkConfig} network - Optional object to specify proxy/whether to ignore SSL certificates. + * * @returns {Promise} A promise that resolves when the file is downloaded. */ -const downloadFile: ( - url: string, - fileName: string, - network?: { proxy?: string; ignoreSSL?: boolean } -) => Promise = (url, fileName, network) => { - return global.core?.api?.downloadFile(url, fileName, network) -} +const downloadFile: (downloadRequest: DownloadRequest, network?: NetworkConfig) => Promise = ( + downloadRequest, + network +) => global.core?.api?.downloadFile(downloadRequest, network) /** * Aborts the download of a specific file. @@ -108,6 +106,20 @@ const log: (message: string, fileName?: string) => void = (message, fileName) => const isSubdirectory: (from: string, to: string) => Promise = (from: string, to: string) => global.core.api?.isSubdirectory(from, to) +/** + * Get system information + * @returns {Promise} - A promise that resolves with the system information. + */ +const systemInformations: () => Promise = () => global.core.api?.systemInformations() + +/** + * Show toast message from browser processes. + * @param title + * @param message + * @returns + */ +const showToast: (title: string, message: string) => void = (title, message) => + global.core.api?.showToast(title, message) /** * Register extension point function type definition */ @@ -134,5 +146,7 @@ export { log, isSubdirectory, getUserHomePath, + systemInformations, + showToast, FileStat, } diff --git a/core/src/extension.ts b/core/src/extension.ts index 3b3edc7b3..22accb4b4 100644 --- a/core/src/extension.ts +++ b/core/src/extension.ts @@ -10,6 +10,22 @@ export enum ExtensionTypeEnum { export interface ExtensionType { type(): ExtensionTypeEnum | undefined } + +export interface Compatibility { + platform: string[] + version: string +} + +const ALL_INSTALLATION_STATE = [ + 'NotRequired', // not required. + 'Installed', // require and installed. Good to go. + 'NotInstalled', // require to be installed. + 'Corrupted', // require but corrupted. Need to redownload. +] as const + +export type InstallationStateTuple = typeof ALL_INSTALLATION_STATE +export type InstallationState = InstallationStateTuple[number] + /** * Represents a base extension. * This class should be extended by any class that represents an extension. @@ -33,4 +49,32 @@ export abstract class BaseExtension implements ExtensionType { * Any cleanup logic for the extension should be put here. */ abstract onUnload(): void + + /** + * The compatibility of the extension. + * This is used to check if the extension is compatible with the current environment. + * @property {Array} platform + */ + compatibility(): Compatibility | undefined { + return undefined + } + + /** + * Determine if the prerequisites for the extension are installed. + * + * @returns {boolean} true if the prerequisites are installed, false otherwise. + */ + async installationState(): Promise { + return 'NotRequired' + } + + /** + * Install the prerequisites for the extension. + * + * @returns {Promise} + */ + // @ts-ignore + async install(...args): Promise { + return + } } diff --git a/core/src/extensions/ai-engines/AIEngine.ts b/core/src/extensions/ai-engines/AIEngine.ts new file mode 100644 index 000000000..608b5c193 --- /dev/null +++ b/core/src/extensions/ai-engines/AIEngine.ts @@ -0,0 +1,60 @@ +import { getJanDataFolderPath, joinPath } from '../../core' +import { events } from '../../events' +import { BaseExtension } from '../../extension' +import { fs } from '../../fs' +import { Model, ModelEvent } from '../../types' + +/** + * Base AIEngine + * Applicable to all AI Engines + */ +export abstract class AIEngine extends BaseExtension { + // The inference engine + abstract provider: string + // The model folder + modelFolder: string = 'models' + + abstract models(): Promise + + /** + * On extension load, subscribe to events. + */ + onLoad() { + this.prePopulateModels() + } + + /** + * Pre-populate models to App Data Folder + */ + prePopulateModels(): Promise { + return this.models().then((models) => { + const prePoluateOperations = models.map((model) => + getJanDataFolderPath() + .then((janDataFolder) => + // Attempt to create the model folder + joinPath([janDataFolder, this.modelFolder, model.id]).then((path) => + fs + .mkdirSync(path) + .catch() + .then(() => path) + ) + ) + .then((path) => joinPath([path, 'model.json'])) + .then((path) => { + // Do not overwite existing model.json + return fs.existsSync(path).then((exist: any) => { + if (!exist) return fs.writeFileSync(path, JSON.stringify(model, null, 2)) + }) + }) + .catch((e: Error) => { + console.error('Error', e) + }) + ) + Promise.all(prePoluateOperations).then(() => + // Emit event to update models + // So the UI can update the models list + events.emit(ModelEvent.OnModelsUpdate, {}) + ) + }) + } +} diff --git a/core/src/extensions/ai-engines/LocalOAIEngine.ts b/core/src/extensions/ai-engines/LocalOAIEngine.ts new file mode 100644 index 000000000..79dbcbf5e --- /dev/null +++ b/core/src/extensions/ai-engines/LocalOAIEngine.ts @@ -0,0 +1,63 @@ +import { executeOnMain, getJanDataFolderPath, joinPath } from '../../core' +import { events } from '../../events' +import { Model, ModelEvent } from '../../types' +import { OAIEngine } from './OAIEngine' + +/** + * Base OAI Local Inference Provider + * Added the implementation of loading and unloading model (applicable to local inference providers) + */ +export abstract class LocalOAIEngine extends OAIEngine { + // The inference engine + loadModelFunctionName: string = 'loadModel' + unloadModelFunctionName: string = 'unloadModel' + isRunning: boolean = false + + /** + * On extension load, subscribe to events. + */ + onLoad() { + super.onLoad() + // These events are applicable to local inference providers + events.on(ModelEvent.OnModelInit, (model: Model) => this.onModelInit(model)) + events.on(ModelEvent.OnModelStop, (model: Model) => this.onModelStop(model)) + } + + /** + * Load the model. + */ + async onModelInit(model: Model) { + if (model.engine.toString() !== this.provider) return + + const modelFolder = await joinPath([await getJanDataFolderPath(), this.modelFolder, model.id]) + + const res = await executeOnMain(this.nodeModule, this.loadModelFunctionName, { + modelFolder, + model, + }) + + if (res?.error) { + events.emit(ModelEvent.OnModelFail, { + ...model, + error: res.error, + }) + return + } else { + this.loadedModel = model + events.emit(ModelEvent.OnModelReady, model) + this.isRunning = true + } + } + /** + * Stops the model. + */ + onModelStop(model: Model) { + if (model.engine?.toString() !== this.provider) return + + this.isRunning = false + + executeOnMain(this.nodeModule, this.unloadModelFunctionName).then(() => { + events.emit(ModelEvent.OnModelStopped, {}) + }) + } +} diff --git a/core/src/extensions/ai-engines/OAIEngine.ts b/core/src/extensions/ai-engines/OAIEngine.ts new file mode 100644 index 000000000..3e583c9b9 --- /dev/null +++ b/core/src/extensions/ai-engines/OAIEngine.ts @@ -0,0 +1,116 @@ +import { requestInference } from './helpers/sse' +import { ulid } from 'ulid' +import { AIEngine } from './AIEngine' +import { + ChatCompletionRole, + ContentType, + InferenceEvent, + MessageEvent, + MessageRequest, + MessageRequestType, + MessageStatus, + Model, + ModelInfo, + ThreadContent, + ThreadMessage, +} from '../../types' +import { events } from '../../events' + +/** + * Base OAI Inference Provider + * Applicable to all OAI compatible inference providers + */ +export abstract class OAIEngine extends AIEngine { + // The inference engine + abstract inferenceUrl: string + abstract nodeModule: string + + // Controller to handle stop requests + controller = new AbortController() + isCancelled = false + + // The loaded model instance + loadedModel: Model | undefined + + /** + * On extension load, subscribe to events. + */ + onLoad() { + super.onLoad() + events.on(MessageEvent.OnMessageSent, (data: MessageRequest) => this.inference(data)) + events.on(InferenceEvent.OnInferenceStopped, () => this.onInferenceStopped()) + } + + /** + * On extension unload + */ + onUnload(): void {} + + /* + * Inference request + */ + inference(data: MessageRequest) { + if (data.model?.engine?.toString() !== this.provider) return + + const timestamp = Date.now() + const message: ThreadMessage = { + id: ulid(), + thread_id: data.threadId, + type: data.type, + assistant_id: data.assistantId, + role: ChatCompletionRole.Assistant, + content: [], + status: MessageStatus.Pending, + created: timestamp, + updated: timestamp, + object: 'thread.message', + } + + if (data.type !== MessageRequestType.Summary) { + events.emit(MessageEvent.OnMessageResponse, message) + } + + this.isCancelled = false + this.controller = new AbortController() + + const model: ModelInfo = { + ...(this.loadedModel ? this.loadedModel : {}), + ...data.model, + } + + requestInference(this.inferenceUrl, data.messages ?? [], model, this.controller).subscribe({ + next: (content: any) => { + const messageContent: ThreadContent = { + type: ContentType.Text, + text: { + value: content.trim(), + annotations: [], + }, + } + message.content = [messageContent] + events.emit(MessageEvent.OnMessageUpdate, message) + }, + complete: async () => { + message.status = message.content.length ? MessageStatus.Ready : MessageStatus.Error + events.emit(MessageEvent.OnMessageUpdate, message) + }, + error: async (err: any) => { + if (this.isCancelled || message.content.length) { + message.status = MessageStatus.Stopped + events.emit(MessageEvent.OnMessageUpdate, message) + return + } + message.status = MessageStatus.Error + events.emit(MessageEvent.OnMessageUpdate, message) + }, + }) + } + + /** + * Stops the inference. + */ + onInferenceStopped() { + this.isCancelled = true + this.controller?.abort() + } +} diff --git a/core/src/extensions/ai-engines/helpers/sse.ts b/core/src/extensions/ai-engines/helpers/sse.ts new file mode 100644 index 000000000..3d810d934 --- /dev/null +++ b/core/src/extensions/ai-engines/helpers/sse.ts @@ -0,0 +1,67 @@ +import { Observable } from 'rxjs' +import { ModelRuntimeParams } from '../../../types' +/** + * Sends a request to the inference server to generate a response based on the recent messages. + * @param recentMessages - An array of recent messages to use as context for the inference. + * @returns An Observable that emits the generated response as a string. + */ +export function requestInference( + inferenceUrl: string, + recentMessages: any[], + model: { + id: string + parameters: ModelRuntimeParams + }, + controller?: AbortController +): Observable { + return new Observable((subscriber) => { + const requestBody = JSON.stringify({ + messages: recentMessages, + model: model.id, + stream: true, + ...model.parameters, + }) + fetch(inferenceUrl, { + method: 'POST', + headers: { + 'Content-Type': 'application/json', + 'Access-Control-Allow-Origin': '*', + 'Accept': model.parameters.stream ? 'text/event-stream' : 'application/json', + }, + body: requestBody, + signal: controller?.signal, + }) + .then(async (response) => { + if (model.parameters.stream === false) { + const data = await response.json() + subscriber.next(data.choices[0]?.message?.content ?? '') + } else { + const stream = response.body + const decoder = new TextDecoder('utf-8') + const reader = stream?.getReader() + let content = '' + + while (true && reader) { + const { done, value } = await reader.read() + if (done) { + break + } + const text = decoder.decode(value) + const lines = text.trim().split('\n') + for (const line of lines) { + if (line.startsWith('data: ') && !line.includes('data: [DONE]')) { + const data = JSON.parse(line.replace('data: ', '')) + content += data.choices[0]?.delta?.content ?? '' + if (content.startsWith('assistant: ')) { + content = content.replace('assistant: ', '') + } + subscriber.next(content) + } + } + } + } + subscriber.complete() + }) + .catch((err) => subscriber.error(err)) + }) +} diff --git a/core/src/extensions/ai-engines/index.ts b/core/src/extensions/ai-engines/index.ts new file mode 100644 index 000000000..f4da62a7c --- /dev/null +++ b/core/src/extensions/ai-engines/index.ts @@ -0,0 +1,3 @@ +export * from './AIEngine' +export * from './OAIEngine' +export * from './LocalOAIEngine' diff --git a/core/src/extensions/index.ts b/core/src/extensions/index.ts index c6834482c..c049f3b3a 100644 --- a/core/src/extensions/index.ts +++ b/core/src/extensions/index.ts @@ -28,3 +28,8 @@ export { ModelExtension } from './model' * Hugging Face extension for converting HF models to GGUF. */ export { HuggingFaceExtension } from './huggingface' + +/** + * Base AI Engines. + */ +export * from './ai-engines' diff --git a/core/src/extensions/model.ts b/core/src/extensions/model.ts index 79202398b..33eec0afc 100644 --- a/core/src/extensions/model.ts +++ b/core/src/extensions/model.ts @@ -1,5 +1,5 @@ import { BaseExtension, ExtensionTypeEnum } from '../extension' -import { ImportingModel, Model, ModelInterface, OptionType } from '../index' +import { GpuSetting, ImportingModel, Model, ModelInterface, OptionType } from '../index' /** * Model extension for managing models. @@ -14,6 +14,7 @@ export abstract class ModelExtension extends BaseExtension implements ModelInter abstract downloadModel( model: Model, + gpuSettings?: GpuSetting, network?: { proxy: string; ignoreSSL?: boolean } ): Promise abstract cancelModelDownload(modelId: string): Promise diff --git a/core/src/extensions/monitoring.ts b/core/src/extensions/monitoring.ts index ba193f0f4..8d61580fc 100644 --- a/core/src/extensions/monitoring.ts +++ b/core/src/extensions/monitoring.ts @@ -1,5 +1,5 @@ import { BaseExtension, ExtensionTypeEnum } from '../extension' -import { MonitoringInterface } from '../index' +import { GpuSetting, MonitoringInterface } from '../index' /** * Monitoring extension for system monitoring. @@ -13,6 +13,7 @@ export abstract class MonitoringExtension extends BaseExtension implements Monit return ExtensionTypeEnum.SystemMonitoring } + abstract getGpuSetting(): Promise abstract getResourcesInfo(): Promise abstract getCurrentLoad(): Promise } diff --git a/core/src/node/api/processors/download.ts b/core/src/node/api/processors/download.ts index 4ddeff160..8e8e08f2f 100644 --- a/core/src/node/api/processors/download.ts +++ b/core/src/node/api/processors/download.ts @@ -5,7 +5,7 @@ import { getJanDataFolderPath } from '../../helper' import { DownloadManager } from '../../helper/download' import { createWriteStream, renameSync } from 'fs' import { Processor } from './Processor' -import { DownloadState } from '../../../types' +import { DownloadRequest, DownloadState, NetworkConfig } from '../../../types' export class Downloader implements Processor { observer?: Function @@ -20,24 +20,27 @@ export class Downloader implements Processor { return func(this.observer, ...args) } - downloadFile(observer: any, url: string, localPath: string, network: any) { + downloadFile(observer: any, downloadRequest: DownloadRequest, network?: NetworkConfig) { const request = require('request') const progress = require('request-progress') const strictSSL = !network?.ignoreSSL const proxy = network?.proxy?.startsWith('http') ? network.proxy : undefined + + const { localPath, url } = downloadRequest + let normalizedPath = localPath if (typeof localPath === 'string') { - localPath = normalizeFilePath(localPath) + normalizedPath = normalizeFilePath(localPath) } - const array = localPath.split(sep) + const array = normalizedPath.split(sep) const fileName = array.pop() ?? '' const modelId = array.pop() ?? '' - const destination = resolve(getJanDataFolderPath(), localPath) + const destination = resolve(getJanDataFolderPath(), normalizedPath) const rq = request({ url, strictSSL, proxy }) // Put request to download manager instance - DownloadManager.instance.setRequest(localPath, rq) + DownloadManager.instance.setRequest(normalizedPath, rq) // Downloading file to a temp file first const downloadingTempFile = `${destination}.download` @@ -56,16 +59,25 @@ export class Downloader implements Processor { total: 0, transferred: 0, }, + children: [], downloadState: 'downloading', + extensionId: downloadRequest.extensionId, + downloadType: downloadRequest.downloadType, + localPath: normalizedPath, } DownloadManager.instance.downloadProgressMap[modelId] = initialDownloadState + if (downloadRequest.downloadType === 'extension') { + observer?.(DownloadEvent.onFileDownloadUpdate, initialDownloadState) + } + progress(rq, {}) .on('progress', (state: any) => { + const currentDownloadState = DownloadManager.instance.downloadProgressMap[modelId] const downloadState: DownloadState = { + ...currentDownloadState, ...state, - modelId, - fileName, + fileName: fileName, downloadState: 'downloading', } console.debug('progress: ', downloadState) @@ -76,22 +88,22 @@ export class Downloader implements Processor { const currentDownloadState = DownloadManager.instance.downloadProgressMap[modelId] const downloadState: DownloadState = { ...currentDownloadState, + fileName: fileName, error: error.message, downloadState: 'error', } - if (currentDownloadState) { - DownloadManager.instance.downloadProgressMap[modelId] = downloadState - } observer?.(DownloadEvent.onFileDownloadError, downloadState) + DownloadManager.instance.downloadProgressMap[modelId] = downloadState }) .on('end', () => { const currentDownloadState = DownloadManager.instance.downloadProgressMap[modelId] - if (currentDownloadState && DownloadManager.instance.networkRequests[localPath]) { + if (currentDownloadState && DownloadManager.instance.networkRequests[normalizedPath]) { // Finished downloading, rename temp file to actual file renameSync(downloadingTempFile, destination) const downloadState: DownloadState = { ...currentDownloadState, + fileName: fileName, downloadState: 'end', } observer?.(DownloadEvent.onFileDownloadSuccess, downloadState) diff --git a/core/src/node/api/restful/helper/builder.ts b/core/src/node/api/restful/helper/builder.ts index 7001c0c76..6b9bbb3a8 100644 --- a/core/src/node/api/restful/helper/builder.ts +++ b/core/src/node/api/restful/helper/builder.ts @@ -1,7 +1,16 @@ -import fs from 'fs' +import { + existsSync, + readdirSync, + readFileSync, + writeFileSync, + mkdirSync, + appendFileSync, + createWriteStream, + rmdirSync, +} from 'fs' import { JanApiRouteConfiguration, RouteConfiguration } from './configuration' import { join } from 'path' -import { ContentType, MessageStatus, Model, ThreadMessage } from '../../../../index' +import { ContentType, MessageStatus, Model, ThreadMessage } from '../../../../types' import { getEngineConfiguration, getJanDataFolderPath } from '../../../helper' import { DEFAULT_CHAT_COMPLETION_URL } from './consts' @@ -9,12 +18,12 @@ import { DEFAULT_CHAT_COMPLETION_URL } from './consts' export const getBuilder = async (configuration: RouteConfiguration) => { const directoryPath = join(getJanDataFolderPath(), configuration.dirName) try { - if (!fs.existsSync(directoryPath)) { + if (!existsSync(directoryPath)) { console.debug('model folder not found') return [] } - const files: string[] = fs.readdirSync(directoryPath) + const files: string[] = readdirSync(directoryPath) const allDirectories: string[] = [] for (const file of files) { @@ -46,8 +55,8 @@ export const getBuilder = async (configuration: RouteConfiguration) => { } const readModelMetadata = (path: string): string | undefined => { - if (fs.existsSync(path)) { - return fs.readFileSync(path, 'utf-8') + if (existsSync(path)) { + return readFileSync(path, 'utf-8') } else { return undefined } @@ -81,7 +90,7 @@ export const deleteBuilder = async (configuration: RouteConfiguration, id: strin } const objectPath = join(directoryPath, id) - fs.rmdirSync(objectPath, { recursive: true }) + rmdirSync(objectPath, { recursive: true }) return { id: id, object: configuration.delete.object, @@ -96,20 +105,19 @@ export const getMessages = async (threadId: string): Promise => const threadDirPath = join(getJanDataFolderPath(), 'threads', threadId) const messageFile = 'messages.jsonl' try { - const files: string[] = fs.readdirSync(threadDirPath) + const files: string[] = readdirSync(threadDirPath) if (!files.includes(messageFile)) { console.error(`${threadDirPath} not contains message file`) return [] } const messageFilePath = join(threadDirPath, messageFile) - if (!fs.existsSync(messageFilePath)) { + if (!existsSync(messageFilePath)) { console.debug('message file not found') return [] } - const lines = fs - .readFileSync(messageFilePath, 'utf-8') + const lines = readFileSync(messageFilePath, 'utf-8') .toString() .split('\n') .filter((line: any) => line !== '') @@ -157,11 +165,11 @@ export const createThread = async (thread: any) => { const threadDirPath = join(getJanDataFolderPath(), 'threads', updatedThread.id) const threadJsonPath = join(threadDirPath, threadMetadataFileName) - if (!fs.existsSync(threadDirPath)) { - fs.mkdirSync(threadDirPath) + if (!existsSync(threadDirPath)) { + mkdirSync(threadDirPath) } - await fs.writeFileSync(threadJsonPath, JSON.stringify(updatedThread, null, 2)) + await writeFileSync(threadJsonPath, JSON.stringify(updatedThread, null, 2)) return updatedThread } catch (err) { return { @@ -191,7 +199,7 @@ export const updateThread = async (threadId: string, thread: any) => { const threadDirPath = join(getJanDataFolderPath(), 'threads', updatedThread.id) const threadJsonPath = join(threadDirPath, threadMetadataFileName) - await fs.writeFileSync(threadJsonPath, JSON.stringify(updatedThread, null, 2)) + await writeFileSync(threadJsonPath, JSON.stringify(updatedThread, null, 2)) return updatedThread } catch (err) { return { @@ -233,10 +241,10 @@ export const createMessage = async (threadId: string, message: any) => { const threadDirPath = join(getJanDataFolderPath(), 'threads', threadId) const threadMessagePath = join(threadDirPath, threadMessagesFileName) - if (!fs.existsSync(threadDirPath)) { - fs.mkdirSync(threadDirPath) + if (!existsSync(threadDirPath)) { + mkdirSync(threadDirPath) } - fs.appendFileSync(threadMessagePath, JSON.stringify(threadMessage) + '\n') + appendFileSync(threadMessagePath, JSON.stringify(threadMessage) + '\n') return threadMessage } catch (err) { return { @@ -259,8 +267,8 @@ export const downloadModel = async ( } const directoryPath = join(getJanDataFolderPath(), 'models', modelId) - if (!fs.existsSync(directoryPath)) { - fs.mkdirSync(directoryPath) + if (!existsSync(directoryPath)) { + mkdirSync(directoryPath) } // path to model binary @@ -281,7 +289,7 @@ export const downloadModel = async ( .on('end', function () { console.debug('end') }) - .pipe(fs.createWriteStream(modelBinaryPath)) + .pipe(createWriteStream(modelBinaryPath)) } return { diff --git a/core/src/types/file/index.ts b/core/src/types/file/index.ts index cc7274a28..d941987ef 100644 --- a/core/src/types/file/index.ts +++ b/core/src/types/file/index.ts @@ -4,16 +4,43 @@ export type FileStat = { } export type DownloadState = { - modelId: string + modelId: string // TODO: change to download id fileName: string time: DownloadTime speed: number - percent: number + percent: number size: DownloadSize - children?: DownloadState[] - error?: string downloadState: 'downloading' | 'error' | 'end' + children?: DownloadState[] + + error?: string + extensionId?: string + downloadType?: DownloadType + localPath?: string +} + +export type DownloadType = 'model' | 'extension' + +export type DownloadRequest = { + /** + * The URL to download the file from. + */ + url: string + + /** + * The local path to save the file to. + */ + localPath: string + + /** + * The extension ID of the extension that initiated the download. + * + * Can be extension name. + */ + extensionId?: string + + downloadType?: DownloadType } type DownloadTime = { diff --git a/core/src/types/miscellaneous/fileDownloadRequest.ts b/core/src/types/miscellaneous/fileDownloadRequest.ts new file mode 100644 index 000000000..83131aa71 --- /dev/null +++ b/core/src/types/miscellaneous/fileDownloadRequest.ts @@ -0,0 +1,8 @@ +export type FileDownloadRequest = { + downloadId: string + url: string + localPath: string + fileName: string + displayName: string + metadata: Record +} diff --git a/core/src/types/miscellaneous/index.ts b/core/src/types/miscellaneous/index.ts index e9c205a73..b4ef68ab6 100644 --- a/core/src/types/miscellaneous/index.ts +++ b/core/src/types/miscellaneous/index.ts @@ -1,3 +1,5 @@ export * from './systemResourceInfo' export * from './promptTemplate' export * from './appUpdate' +export * from './fileDownloadRequest' +export * from './networkConfig' \ No newline at end of file diff --git a/core/src/types/miscellaneous/networkConfig.ts b/core/src/types/miscellaneous/networkConfig.ts new file mode 100644 index 000000000..2d27f4223 --- /dev/null +++ b/core/src/types/miscellaneous/networkConfig.ts @@ -0,0 +1,4 @@ +export type NetworkConfig = { + proxy?: string + ignoreSSL?: boolean +} diff --git a/core/src/types/miscellaneous/systemResourceInfo.ts b/core/src/types/miscellaneous/systemResourceInfo.ts index 1472cda47..f7dd4a82b 100644 --- a/core/src/types/miscellaneous/systemResourceInfo.ts +++ b/core/src/types/miscellaneous/systemResourceInfo.ts @@ -2,3 +2,31 @@ export type SystemResourceInfo = { numCpuPhysicalCore: number memAvailable: number } + +export type RunMode = 'cpu' | 'gpu' + +export type GpuSetting = { + notify: boolean + run_mode: RunMode + nvidia_driver: { + exist: boolean + version: string + } + cuda: { + exist: boolean + version: string + } + gpus: GpuSettingInfo[] + gpu_highest_vram: string + gpus_in_use: string[] + is_initial: boolean + // TODO: This needs to be set based on user toggle in settings + vulkan: boolean +} + +export type GpuSettingInfo = { + id: string + vram: string + name: string + arch?: string +} diff --git a/core/src/types/model/modelEntity.ts b/core/src/types/model/modelEntity.ts index 11d3e0526..74568686b 100644 --- a/core/src/types/model/modelEntity.ts +++ b/core/src/types/model/modelEntity.ts @@ -19,6 +19,7 @@ export enum InferenceEngine { nitro = 'nitro', openai = 'openai', triton_trtllm = 'triton_trtllm', + nitro_tensorrt_llm = 'nitro-tensorrt-llm', tool_retrieval_enabled = 'tool_retrieval_enabled', } diff --git a/core/src/types/model/modelInterface.ts b/core/src/types/model/modelInterface.ts index 93d5867ee..639c7c8d3 100644 --- a/core/src/types/model/modelInterface.ts +++ b/core/src/types/model/modelInterface.ts @@ -1,3 +1,4 @@ +import { GpuSetting } from '../miscellaneous' import { Model } from './modelEntity' /** @@ -10,7 +11,11 @@ export interface ModelInterface { * @param network - Optional object to specify proxy/whether to ignore SSL certificates. * @returns A Promise that resolves when the model has been downloaded. */ - downloadModel(model: Model, network?: { ignoreSSL?: boolean; proxy?: string }): Promise + downloadModel( + model: Model, + gpuSettings?: GpuSetting, + network?: { ignoreSSL?: boolean; proxy?: string } + ): Promise /** * Cancels the download of a specific model. diff --git a/core/src/types/monitoring/index.ts b/core/src/types/monitoring/index.ts index 5828dae8b..b96c518fd 100644 --- a/core/src/types/monitoring/index.ts +++ b/core/src/types/monitoring/index.ts @@ -1 +1,2 @@ export * from './monitoringInterface' +export * from './resourceInfo' diff --git a/core/src/types/monitoring/resourceInfo.ts b/core/src/types/monitoring/resourceInfo.ts new file mode 100644 index 000000000..b19da5462 --- /dev/null +++ b/core/src/types/monitoring/resourceInfo.ts @@ -0,0 +1,6 @@ +export type ResourceInfo = { + mem: { + totalMemory: number + usedMemory: number + } +} diff --git a/core/tsconfig.json b/core/tsconfig.json index b112079d2..daeb7eeff 100644 --- a/core/tsconfig.json +++ b/core/tsconfig.json @@ -13,7 +13,7 @@ "declarationDir": "dist/types", "outDir": "dist/lib", "importHelpers": true, - "types": ["@types/jest"] + "types": ["@types/jest"], }, - "include": ["src"] + "include": ["src"], } diff --git a/extensions/huggingface-extension/src/index.ts b/extensions/huggingface-extension/src/index.ts index d8f755080..88292ce58 100644 --- a/extensions/huggingface-extension/src/index.ts +++ b/extensions/huggingface-extension/src/index.ts @@ -13,6 +13,7 @@ import { events, DownloadEvent, log, + DownloadRequest, } from '@janhq/core' import { ggufMetadata } from 'hyllama' @@ -148,7 +149,11 @@ export default class JanHuggingFaceExtension extends HuggingFaceExtension { if (this.interrupted) return if (!(await fs.existsSync(localPath))) { - downloadFile(url, localPath, network) + const downloadRequest: DownloadRequest = { + url, + localPath, + } + downloadFile(downloadRequest, network) filePaths.push(filePath) } } diff --git a/extensions/inference-nitro-extension/download.bat b/extensions/inference-nitro-extension/download.bat index 2ef3165c1..bb8c4ffdc 100644 --- a/extensions/inference-nitro-extension/download.bat +++ b/extensions/inference-nitro-extension/download.bat @@ -1,3 +1,3 @@ @echo off set /p NITRO_VERSION=<./bin/version.txt -.\node_modules\.bin\download https://github.com/janhq/nitro/releases/download/v%NITRO_VERSION%/nitro-%NITRO_VERSION%-win-amd64-cuda-12-0.tar.gz -e --strip 1 -o ./bin/win-cuda-12-0 && .\node_modules\.bin\download https://github.com/janhq/nitro/releases/download/v%NITRO_VERSION%/nitro-%NITRO_VERSION%-win-amd64-cuda-11-7.tar.gz -e --strip 1 -o ./bin/win-cuda-11-7 && .\node_modules\.bin\download https://github.com/janhq/nitro/releases/download/v%NITRO_VERSION%/nitro-%NITRO_VERSION%-win-amd64.tar.gz -e --strip 1 -o ./bin/win-cpu && .\node_modules\.bin\download https://github.com/janhq/nitro/releases/download/v%NITRO_VERSION%/nitro-%NITRO_VERSION%-win-amd64-vulkan.tar.gz -e --strip 1 -o ./bin/win-vulkan && .\node_modules\.bin\download https://delta.jan.ai/vulkaninfoSDK.exe -o ./bin +.\node_modules\.bin\download https://github.com/janhq/nitro/releases/download/v%NITRO_VERSION%/nitro-%NITRO_VERSION%-win-amd64-cuda-12-0.tar.gz -e --strip 1 -o ./bin/win-cuda-12-0 && .\node_modules\.bin\download https://github.com/janhq/nitro/releases/download/v%NITRO_VERSION%/nitro-%NITRO_VERSION%-win-amd64-cuda-11-7.tar.gz -e --strip 1 -o ./bin/win-cuda-11-7 && .\node_modules\.bin\download https://github.com/janhq/nitro/releases/download/v%NITRO_VERSION%/nitro-%NITRO_VERSION%-win-amd64.tar.gz -e --strip 1 -o ./bin/win-cpu && .\node_modules\.bin\download https://github.com/janhq/nitro/releases/download/v%NITRO_VERSION%/nitro-%NITRO_VERSION%-win-amd64-vulkan.tar.gz -e --strip 1 -o ./bin/win-vulkan diff --git a/extensions/inference-nitro-extension/package.json b/extensions/inference-nitro-extension/package.json index e6365ad92..dd5798764 100644 --- a/extensions/inference-nitro-extension/package.json +++ b/extensions/inference-nitro-extension/package.json @@ -8,7 +8,7 @@ "license": "AGPL-3.0", "scripts": { "build": "tsc --module commonjs && rollup -c rollup.config.ts", - "downloadnitro:linux": "NITRO_VERSION=$(cat ./bin/version.txt) && download https://github.com/janhq/nitro/releases/download/v${NITRO_VERSION}/nitro-${NITRO_VERSION}-linux-amd64.tar.gz -e --strip 1 -o ./bin/linux-cpu && chmod +x ./bin/linux-cpu/nitro && download https://github.com/janhq/nitro/releases/download/v${NITRO_VERSION}/nitro-${NITRO_VERSION}-linux-amd64-cuda-12-0.tar.gz -e --strip 1 -o ./bin/linux-cuda-12-0 && chmod +x ./bin/linux-cuda-12-0/nitro && download https://github.com/janhq/nitro/releases/download/v${NITRO_VERSION}/nitro-${NITRO_VERSION}-linux-amd64-cuda-11-7.tar.gz -e --strip 1 -o ./bin/linux-cuda-11-7 && chmod +x ./bin/linux-cuda-11-7/nitro && download https://github.com/janhq/nitro/releases/download/v${NITRO_VERSION}/nitro-${NITRO_VERSION}-linux-amd64-vulkan.tar.gz -e --strip 1 -o ./bin/linux-vulkan && chmod +x ./bin/linux-vulkan/nitro && download https://delta.jan.ai/vulkaninfo -o ./bin && chmod +x ./bin/vulkaninfo", + "downloadnitro:linux": "NITRO_VERSION=$(cat ./bin/version.txt) && download https://github.com/janhq/nitro/releases/download/v${NITRO_VERSION}/nitro-${NITRO_VERSION}-linux-amd64.tar.gz -e --strip 1 -o ./bin/linux-cpu && chmod +x ./bin/linux-cpu/nitro && download https://github.com/janhq/nitro/releases/download/v${NITRO_VERSION}/nitro-${NITRO_VERSION}-linux-amd64-cuda-12-0.tar.gz -e --strip 1 -o ./bin/linux-cuda-12-0 && chmod +x ./bin/linux-cuda-12-0/nitro && download https://github.com/janhq/nitro/releases/download/v${NITRO_VERSION}/nitro-${NITRO_VERSION}-linux-amd64-cuda-11-7.tar.gz -e --strip 1 -o ./bin/linux-cuda-11-7 && chmod +x ./bin/linux-cuda-11-7/nitro && download https://github.com/janhq/nitro/releases/download/v${NITRO_VERSION}/nitro-${NITRO_VERSION}-linux-amd64-vulkan.tar.gz -e --strip 1 -o ./bin/linux-vulkan && chmod +x ./bin/linux-vulkan/nitro", "downloadnitro:darwin": "NITRO_VERSION=$(cat ./bin/version.txt) && download https://github.com/janhq/nitro/releases/download/v${NITRO_VERSION}/nitro-${NITRO_VERSION}-mac-arm64.tar.gz -e --strip 1 -o ./bin/mac-arm64 && chmod +x ./bin/mac-arm64/nitro && download https://github.com/janhq/nitro/releases/download/v${NITRO_VERSION}/nitro-${NITRO_VERSION}-mac-amd64.tar.gz -e --strip 1 -o ./bin/mac-x64 && chmod +x ./bin/mac-x64/nitro", "downloadnitro:win32": "download.bat", "downloadnitro": "run-script-os", diff --git a/extensions/inference-nitro-extension/src/index.ts b/extensions/inference-nitro-extension/src/index.ts index 979b4cfac..70244a5d9 100644 --- a/extensions/inference-nitro-extension/src/index.ts +++ b/extensions/inference-nitro-extension/src/index.ts @@ -108,9 +108,6 @@ export default class JanInferenceNitroExtension extends InferenceExtension { events.on(InferenceEvent.OnInferenceStopped, () => this.onInferenceStopped() ) - - // Attempt to fetch nvidia info - await executeOnMain(NODE, 'updateNvidiaInfo', {}) } /** diff --git a/extensions/inference-nitro-extension/src/node/accelerator.ts b/extensions/inference-nitro-extension/src/node/accelerator.ts deleted file mode 100644 index 1ffdbc5bd..000000000 --- a/extensions/inference-nitro-extension/src/node/accelerator.ts +++ /dev/null @@ -1,237 +0,0 @@ -import { writeFileSync, existsSync, readFileSync } from 'fs' -import { exec, spawn } from 'child_process' -import path from 'path' -import { getJanDataFolderPath, log } from '@janhq/core/node' - -/** - * Default GPU settings - * TODO: This needs to be refactored to support multiple accelerators - **/ -const DEFALT_SETTINGS = { - notify: true, - run_mode: 'cpu', - nvidia_driver: { - exist: false, - version: '', - }, - cuda: { - exist: false, - version: '', - }, - gpus: [], - gpu_highest_vram: '', - gpus_in_use: [], - is_initial: true, - // TODO: This needs to be set based on user toggle in settings - vulkan: false -} - -/** - * Path to the settings file - **/ -export const GPU_INFO_FILE = path.join( - getJanDataFolderPath(), - 'settings', - 'settings.json' -) - -/** - * Current nitro process - */ -let nitroProcessInfo: NitroProcessInfo | undefined = undefined - -/** - * Nitro process info - */ -export interface NitroProcessInfo { - isRunning: boolean -} - -/** - * This will retrive GPU informations and persist settings.json - * Will be called when the extension is loaded to turn on GPU acceleration if supported - */ -export async function updateNvidiaInfo() { - if (process.platform !== 'darwin') { - let data - try { - data = JSON.parse(readFileSync(GPU_INFO_FILE, 'utf-8')) - } catch (error) { - data = DEFALT_SETTINGS - writeFileSync(GPU_INFO_FILE, JSON.stringify(data, null, 2)) - } - updateNvidiaDriverInfo() - updateGpuInfo() - } -} - -/** - * Retrieve current nitro process - */ -export const getNitroProcessInfo = (subprocess: any): NitroProcessInfo => { - nitroProcessInfo = { - isRunning: subprocess != null, - } - return nitroProcessInfo -} - -/** - * Validate nvidia and cuda for linux and windows - */ -export async function updateNvidiaDriverInfo(): Promise { - exec( - 'nvidia-smi --query-gpu=driver_version --format=csv,noheader', - (error, stdout) => { - let data = JSON.parse(readFileSync(GPU_INFO_FILE, 'utf-8')) - - if (!error) { - const firstLine = stdout.split('\n')[0].trim() - data['nvidia_driver'].exist = true - data['nvidia_driver'].version = firstLine - } else { - data['nvidia_driver'].exist = false - } - - writeFileSync(GPU_INFO_FILE, JSON.stringify(data, null, 2)) - Promise.resolve() - } - ) -} - -/** - * Check if file exists in paths - */ -export function checkFileExistenceInPaths( - file: string, - paths: string[] -): boolean { - return paths.some((p) => existsSync(path.join(p, file))) -} - -/** - * Validate cuda for linux and windows - */ -export function updateCudaExistence( - data: Record = DEFALT_SETTINGS -): Record { - let filesCuda12: string[] - let filesCuda11: string[] - let paths: string[] - let cudaVersion: string = '' - - if (process.platform === 'win32') { - filesCuda12 = ['cublas64_12.dll', 'cudart64_12.dll', 'cublasLt64_12.dll'] - filesCuda11 = ['cublas64_11.dll', 'cudart64_11.dll', 'cublasLt64_11.dll'] - paths = process.env.PATH ? process.env.PATH.split(path.delimiter) : [] - } else { - filesCuda12 = ['libcudart.so.12', 'libcublas.so.12', 'libcublasLt.so.12'] - filesCuda11 = ['libcudart.so.11.0', 'libcublas.so.11', 'libcublasLt.so.11'] - paths = process.env.LD_LIBRARY_PATH - ? process.env.LD_LIBRARY_PATH.split(path.delimiter) - : [] - paths.push('/usr/lib/x86_64-linux-gnu/') - } - - let cudaExists = filesCuda12.every( - (file) => existsSync(file) || checkFileExistenceInPaths(file, paths) - ) - - if (!cudaExists) { - cudaExists = filesCuda11.every( - (file) => existsSync(file) || checkFileExistenceInPaths(file, paths) - ) - if (cudaExists) { - cudaVersion = '11' - } - } else { - cudaVersion = '12' - } - - data['cuda'].exist = cudaExists - data['cuda'].version = cudaVersion - console.debug(data['is_initial'], data['gpus_in_use']) - if (cudaExists && data['is_initial'] && data['gpus_in_use'].length > 0) { - data.run_mode = 'gpu' - } - data.is_initial = false - return data -} - -/** - * Get GPU information - */ -export async function updateGpuInfo(): Promise { - let data = JSON.parse(readFileSync(GPU_INFO_FILE, 'utf-8')) - - // Cuda - if (data['vulkan'] === true) { - // Vulkan - exec( - process.platform === 'win32' - ? `${__dirname}\\..\\bin\\vulkaninfoSDK.exe --summary` - : `${__dirname}/../bin/vulkaninfo --summary`, - (error, stdout) => { - if (!error) { - const output = stdout.toString() - log(output) - const gpuRegex = /GPU(\d+):(?:[\s\S]*?)deviceName\s*=\s*(.*)/g - - let gpus = [] - let match - while ((match = gpuRegex.exec(output)) !== null) { - const id = match[1] - const name = match[2] - gpus.push({ id, vram: 0, name }) - } - data.gpus = gpus - - if (!data['gpus_in_use'] || data['gpus_in_use'].length === 0) { - data.gpus_in_use = [data.gpus.length > 1 ? '1' : '0'] - } - - data = updateCudaExistence(data) - writeFileSync(GPU_INFO_FILE, JSON.stringify(data, null, 2)) - } - Promise.resolve() - } - ) - } else { - exec( - 'nvidia-smi --query-gpu=index,memory.total,name --format=csv,noheader,nounits', - (error, stdout) => { - if (!error) { - log(stdout) - // Get GPU info and gpu has higher memory first - let highestVram = 0 - let highestVramId = '0' - let gpus = stdout - .trim() - .split('\n') - .map((line) => { - let [id, vram, name] = line.split(', ') - vram = vram.replace(/\r/g, '') - if (parseFloat(vram) > highestVram) { - highestVram = parseFloat(vram) - highestVramId = id - } - return { id, vram, name } - }) - - data.gpus = gpus - data.gpu_highest_vram = highestVramId - } else { - data.gpus = [] - data.gpu_highest_vram = '' - } - - if (!data['gpus_in_use'] || data['gpus_in_use'].length === 0) { - data.gpus_in_use = [data['gpu_highest_vram']] - } - - data = updateCudaExistence(data) - writeFileSync(GPU_INFO_FILE, JSON.stringify(data, null, 2)) - Promise.resolve() - } - ) - } -} diff --git a/extensions/inference-nitro-extension/src/node/execute.ts b/extensions/inference-nitro-extension/src/node/execute.ts index f9a668507..8bcc75ae4 100644 --- a/extensions/inference-nitro-extension/src/node/execute.ts +++ b/extensions/inference-nitro-extension/src/node/execute.ts @@ -1,12 +1,19 @@ +import { getJanDataFolderPath } from '@janhq/core/node' import { readFileSync } from 'fs' import * as path from 'path' -import { GPU_INFO_FILE } from './accelerator' export interface NitroExecutableOptions { executablePath: string cudaVisibleDevices: string vkVisibleDevices: string } + +export const GPU_INFO_FILE = path.join( + getJanDataFolderPath(), + 'settings', + 'settings.json' +) + /** * Find which executable file to run based on the current platform. * @returns The name of the executable file to run. diff --git a/extensions/inference-nitro-extension/src/node/index.ts b/extensions/inference-nitro-extension/src/node/index.ts index 9b2684a6c..c57eb262d 100644 --- a/extensions/inference-nitro-extension/src/node/index.ts +++ b/extensions/inference-nitro-extension/src/node/index.ts @@ -4,7 +4,6 @@ import { ChildProcessWithoutNullStreams, spawn } from 'child_process' import tcpPortUsed from 'tcp-port-used' import fetchRT from 'fetch-retry' import { log, getSystemResourceInfo } from '@janhq/core/node' -import { getNitroProcessInfo, updateNvidiaInfo } from './accelerator' import { Model, InferenceEngine, @@ -385,11 +384,26 @@ function dispose() { killSubprocess() } +/** + * Nitro process info + */ +export interface NitroProcessInfo { + isRunning: boolean +} + +/** + * Retrieve current nitro process + */ +const getCurrentNitroProcessInfo = (): NitroProcessInfo => { + return { + isRunning: subprocess != null, + } +} + export default { runModel, stopModel, killSubprocess, dispose, - updateNvidiaInfo, - getCurrentNitroProcessInfo: () => getNitroProcessInfo(subprocess), + getCurrentNitroProcessInfo, } diff --git a/extensions/model-extension/src/index.ts b/extensions/model-extension/src/index.ts index fb1f26885..d05e7d07f 100644 --- a/extensions/model-extension/src/index.ts +++ b/extensions/model-extension/src/index.ts @@ -17,6 +17,8 @@ import { ImportingModel, LocalImportModelEvent, baseName, + GpuSetting, + DownloadRequest, } from '@janhq/core' import { extractFileName } from './helpers/path' @@ -29,10 +31,14 @@ export default class JanModelExtension extends ModelExtension { private static readonly _modelMetadataFileName = 'model.json' private static readonly _supportedModelFormat = '.gguf' private static readonly _incompletedModelFileName = '.download' - private static readonly _offlineInferenceEngine = InferenceEngine.nitro - + private static readonly _offlineInferenceEngine = [ + InferenceEngine.nitro, + InferenceEngine.nitro_tensorrt_llm, + ] + private static readonly _tensorRtEngineFormat = '.engine' private static readonly _configDirName = 'config' private static readonly _defaultModelFileName = 'default-model.json' + private static readonly _supportedGpuArch = ['turing', 'ampere', 'ada'] /** * Called when the extension is loaded. @@ -89,12 +95,52 @@ export default class JanModelExtension extends ModelExtension { */ async downloadModel( model: Model, + gpuSettings?: GpuSetting, network?: { ignoreSSL?: boolean; proxy?: string } ): Promise { // create corresponding directory const modelDirPath = await joinPath([JanModelExtension._homeDir, model.id]) if (!(await fs.existsSync(modelDirPath))) await fs.mkdirSync(modelDirPath) + if (model.engine === InferenceEngine.nitro_tensorrt_llm) { + if (!gpuSettings || gpuSettings.gpus.length === 0) { + console.error('No GPU found. Please check your GPU setting.') + return + } + const firstGpu = gpuSettings.gpus[0] + if (!firstGpu.name.toLowerCase().includes('nvidia')) { + console.error('No Nvidia GPU found. Please check your GPU setting.') + return + } + const gpuArch = firstGpu.arch + if (gpuArch === undefined) { + console.error( + 'No GPU architecture found. Please check your GPU setting.' + ) + return + } + + if (!JanModelExtension._supportedGpuArch.includes(gpuArch)) { + console.error( + `Your GPU: ${firstGpu} is not supported. Only 20xx, 30xx, 40xx series are supported.` + ) + return + } + + const os = 'windows' // TODO: remove this hard coded value + + const newSources = model.sources.map((source) => { + const newSource = { ...source } + newSource.url = newSource.url + .replace(//g, os) + .replace(//g, gpuArch) + return newSource + }) + model.sources = newSources + } + + console.debug(`Download sources: ${JSON.stringify(model.sources)}`) + if (model.sources.length > 1) { // path to model binaries for (const source of model.sources) { @@ -105,8 +151,11 @@ export default class JanModelExtension extends ModelExtension { if (source.filename) { path = await joinPath([modelDirPath, source.filename]) } - - downloadFile(source.url, path, network) + const downloadRequest: DownloadRequest = { + url: source.url, + localPath: path, + } + downloadFile(downloadRequest, network) } // TODO: handle multiple binaries for web later } else { @@ -115,7 +164,11 @@ export default class JanModelExtension extends ModelExtension { JanModelExtension._supportedModelFormat ) const path = await joinPath([modelDirPath, fileName]) - downloadFile(model.sources[0]?.url, path, network) + const downloadRequest: DownloadRequest = { + url: model.sources[0]?.url, + localPath: path, + } + downloadFile(downloadRequest, network) if (window && window.core?.api && window.core.api.baseApiUrl) { this.startPollingDownloadProgress(model.id) @@ -238,7 +291,7 @@ export default class JanModelExtension extends ModelExtension { async getDownloadedModels(): Promise { return await this.getModelsMetadata( async (modelDir: string, model: Model) => { - if (model.engine !== JanModelExtension._offlineInferenceEngine) + if (!JanModelExtension._offlineInferenceEngine.includes(model.engine)) return true // model binaries (sources) are absolute path & exist @@ -247,22 +300,32 @@ export default class JanModelExtension extends ModelExtension { ) if (existFiles.every((exist) => exist)) return true - return await fs + const result = await fs .readdirSync(await joinPath([JanModelExtension._homeDir, modelDir])) .then((files: string[]) => { // Model binary exists in the directory // Model binary name can match model ID or be a .gguf file and not be an incompleted model file return ( files.includes(modelDir) || - files.filter( - (file) => + files.filter((file) => { + if ( + file.endsWith(JanModelExtension._incompletedModelFileName) + ) { + return false + } + return ( file .toLowerCase() - .includes(JanModelExtension._supportedModelFormat) && - !file.endsWith(JanModelExtension._incompletedModelFileName) - )?.length >= model.sources.length + .includes(JanModelExtension._supportedModelFormat) || + file + .toLowerCase() + .includes(JanModelExtension._tensorRtEngineFormat) + ) + })?.length > 0 // TODO: NamH find better way (can use basename to check the file name with source url) ) }) + + return result } ) } diff --git a/extensions/monitoring-extension/bin/.gitkeep b/extensions/monitoring-extension/bin/.gitkeep new file mode 100644 index 000000000..e69de29bb diff --git a/extensions/monitoring-extension/download.bat b/extensions/monitoring-extension/download.bat new file mode 100644 index 000000000..f1cf8b7ea --- /dev/null +++ b/extensions/monitoring-extension/download.bat @@ -0,0 +1,2 @@ +@echo off +.\node_modules\.bin\download https://delta.jan.ai/vulkaninfoSDK.exe -o ./bin \ No newline at end of file diff --git a/extensions/monitoring-extension/package.json b/extensions/monitoring-extension/package.json index 582f7cd7b..73d28ab37 100644 --- a/extensions/monitoring-extension/package.json +++ b/extensions/monitoring-extension/package.json @@ -3,21 +3,40 @@ "version": "1.0.10", "description": "This extension provides system health and OS level data", "main": "dist/index.js", - "module": "dist/module.js", + "node": "dist/node/index.cjs.js", "author": "Jan ", "license": "AGPL-3.0", "scripts": { - "build": "tsc -b . && webpack --config webpack.config.js", + "build": "tsc --module commonjs && rollup -c rollup.config.ts && npm run download-artifacts", + "download-artifacts": "run-script-os && cpx \"bin/**\" \"dist/bin\"", + "download-artifacts:darwin": "echo 'No artifacts to download for darwin'", + "download-artifacts:win32": "download.bat", + "download-artifacts:linux": "download https://delta.jan.ai/vulkaninfo -o ./bin && chmod +x ./bin/vulkaninfo", "build:publish": "rimraf *.tgz --glob && npm run build && npm pack && cpx *.tgz ../../pre-install" }, + "exports": { + ".": "./dist/index.js", + "./main": "./dist/node/index.cjs.js" + }, "devDependencies": { + "@rollup/plugin-commonjs": "^25.0.7", + "@rollup/plugin-json": "^6.1.0", + "@rollup/plugin-node-resolve": "^15.2.3", + "@types/node": "^20.11.4", + "@types/node-os-utils": "^1.3.4", + "run-script-os": "^1.1.6", + "cpx": "^1.5.0", "rimraf": "^3.0.2", - "webpack": "^5.88.2", - "webpack-cli": "^5.1.4", - "ts-loader": "^9.5.0" + "rollup": "^2.38.5", + "rollup-plugin-define": "^1.0.1", + "rollup-plugin-sourcemaps": "^0.6.3", + "rollup-plugin-typescript2": "^0.36.0", + "typescript": "^5.3.3", + "download-cli": "^1.1.1" }, "dependencies": { "@janhq/core": "file:../../core", + "@rollup/plugin-replace": "^5.0.5", "node-os-utils": "^1.3.7" }, "files": [ diff --git a/extensions/monitoring-extension/rollup.config.ts b/extensions/monitoring-extension/rollup.config.ts new file mode 100644 index 000000000..1b7a40bad --- /dev/null +++ b/extensions/monitoring-extension/rollup.config.ts @@ -0,0 +1,68 @@ +import resolve from '@rollup/plugin-node-resolve' +import commonjs from '@rollup/plugin-commonjs' +import sourceMaps from 'rollup-plugin-sourcemaps' +import typescript from 'rollup-plugin-typescript2' +import json from '@rollup/plugin-json' +import replace from '@rollup/plugin-replace' +const packageJson = require('./package.json') + +export default [ + { + input: `src/index.ts`, + output: [{ file: packageJson.main, format: 'es', sourcemap: true }], + // Indicate here external modules you don't wanna include in your bundle (i.e.: 'lodash') + external: [], + watch: { + include: 'src/**', + }, + plugins: [ + replace({ + NODE: JSON.stringify(`${packageJson.name}/${packageJson.node}`), + }), + // Allow json resolution + json(), + // Compile TypeScript files + typescript({ useTsconfigDeclarationDir: true }), + // Compile TypeScript files + // Allow bundling cjs modules (unlike webpack, rollup doesn't understand cjs) + commonjs(), + // Allow node_modules resolution, so you can use 'external' to control + // which external modules to include in the bundle + // https://github.com/rollup/rollup-plugin-node-resolve#usage + resolve({ + extensions: ['.js', '.ts', '.svelte'], + }), + + // Resolve source maps to the original source + sourceMaps(), + ], + }, + { + input: `src/node/index.ts`, + output: [ + { file: 'dist/node/index.cjs.js', format: 'cjs', sourcemap: true }, + ], + // Indicate here external modules you don't wanna include in your bundle (i.e.: 'lodash') + external: ['@janhq/core/node'], + watch: { + include: 'src/node/**', + }, + plugins: [ + // Allow json resolution + json(), + // Compile TypeScript files + typescript({ useTsconfigDeclarationDir: true }), + // Allow bundling cjs modules (unlike webpack, rollup doesn't understand cjs) + commonjs(), + // Allow node_modules resolution, so you can use 'external' to control + // which external modules to include in the bundle + // https://github.com/rollup/rollup-plugin-node-resolve#usage + resolve({ + extensions: ['.ts', '.js', '.json'], + }), + + // Resolve source maps to the original source + sourceMaps(), + ], + }, +] diff --git a/extensions/monitoring-extension/src/@types/global.d.ts b/extensions/monitoring-extension/src/@types/global.d.ts index 8106353cf..dfa96a0b1 100644 --- a/extensions/monitoring-extension/src/@types/global.d.ts +++ b/extensions/monitoring-extension/src/@types/global.d.ts @@ -1 +1,18 @@ -declare const MODULE: string +declare const NODE: string + +type CpuGpuInfo = { + cpu: { + usage: number + } + gpu: GpuInfo[] +} + +type GpuInfo = { + id: string + name: string + temperature: string + utilization: string + memoryTotal: string + memoryFree: string + memoryUtilization: string +} diff --git a/extensions/monitoring-extension/src/index.ts b/extensions/monitoring-extension/src/index.ts index ce9b2fc14..c7f53455d 100644 --- a/extensions/monitoring-extension/src/index.ts +++ b/extensions/monitoring-extension/src/index.ts @@ -1,4 +1,4 @@ -import { MonitoringExtension, executeOnMain } from '@janhq/core' +import { GpuSetting, MonitoringExtension, executeOnMain } from '@janhq/core' /** * JanMonitoringExtension is a extension that provides system monitoring functionality. @@ -8,19 +8,30 @@ export default class JanMonitoringExtension extends MonitoringExtension { /** * Called when the extension is loaded. */ - async onLoad() {} + async onLoad() { + // Attempt to fetch nvidia info + await executeOnMain(NODE, 'updateNvidiaInfo') + } /** * Called when the extension is unloaded. */ onUnload(): void {} + /** + * Returns the GPU configuration. + * @returns A Promise that resolves to an object containing the GPU configuration. + */ + async getGpuSetting(): Promise { + return executeOnMain(NODE, 'getGpuConfig') + } + /** * Returns information about the system resources. * @returns A Promise that resolves to an object containing information about the system resources. */ getResourcesInfo(): Promise { - return executeOnMain(MODULE, 'getResourcesInfo') + return executeOnMain(NODE, 'getResourcesInfo') } /** @@ -28,6 +39,6 @@ export default class JanMonitoringExtension extends MonitoringExtension { * @returns A Promise that resolves to an object containing information about the current system load. */ getCurrentLoad(): Promise { - return executeOnMain(MODULE, 'getCurrentLoad') + return executeOnMain(NODE, 'getCurrentLoad') } } diff --git a/extensions/monitoring-extension/src/module.ts b/extensions/monitoring-extension/src/module.ts deleted file mode 100644 index 27781a5d6..000000000 --- a/extensions/monitoring-extension/src/module.ts +++ /dev/null @@ -1,92 +0,0 @@ -const nodeOsUtils = require('node-os-utils') -const getJanDataFolderPath = require('@janhq/core/node').getJanDataFolderPath -const path = require('path') -const { readFileSync } = require('fs') -const exec = require('child_process').exec - -const NVIDIA_INFO_FILE = path.join( - getJanDataFolderPath(), - 'settings', - 'settings.json' -) - -const getResourcesInfo = () => - new Promise((resolve) => { - nodeOsUtils.mem.used().then((ramUsedInfo) => { - const totalMemory = ramUsedInfo.totalMemMb * 1024 * 1024 - const usedMemory = ramUsedInfo.usedMemMb * 1024 * 1024 - const response = { - mem: { - totalMemory, - usedMemory, - }, - } - resolve(response) - }) - }) - -const getCurrentLoad = () => - new Promise((resolve, reject) => { - nodeOsUtils.cpu.usage().then((cpuPercentage) => { - let data = { - run_mode: 'cpu', - gpus_in_use: [], - } - if (process.platform !== 'darwin') { - data = JSON.parse(readFileSync(NVIDIA_INFO_FILE, 'utf-8')) - } - if (data.run_mode === 'gpu' && data.gpus_in_use.length > 0) { - const gpuIds = data['gpus_in_use'].join(',') - if (gpuIds !== '' && data['vulkan'] !== true) { - exec( - `nvidia-smi --query-gpu=index,name,temperature.gpu,utilization.gpu,memory.total,memory.free,utilization.memory --format=csv,noheader,nounits --id=${gpuIds}`, - (error, stdout, _) => { - if (error) { - console.error(`exec error: ${error}`) - reject(error) - return - } - const gpuInfo = stdout - .trim() - .split('\n') - .map((line) => { - const [ - id, - name, - temperature, - utilization, - memoryTotal, - memoryFree, - memoryUtilization, - ] = line.split(', ').map((item) => item.replace(/\r/g, '')) - return { - id, - name, - temperature, - utilization, - memoryTotal, - memoryFree, - memoryUtilization, - } - }) - resolve({ - cpu: { usage: cpuPercentage }, - gpu: gpuInfo, - }) - } - ) - } else { - // Handle the case where gpuIds is empty - resolve({ cpu: { usage: cpuPercentage }, gpu: [] }) - } - } else { - // Handle the case where run_mode is not 'gpu' or no GPUs are in use - resolve({ cpu: { usage: cpuPercentage }, gpu: [] }) - } - }) - }) - -module.exports = { - getResourcesInfo, - getCurrentLoad, -} diff --git a/extensions/monitoring-extension/src/node/index.ts b/extensions/monitoring-extension/src/node/index.ts new file mode 100644 index 000000000..1d65704de --- /dev/null +++ b/extensions/monitoring-extension/src/node/index.ts @@ -0,0 +1,317 @@ +import { GpuSetting, GpuSettingInfo, ResourceInfo } from '@janhq/core' +import { getJanDataFolderPath, log } from '@janhq/core/node' +import { mem, cpu } from 'node-os-utils' +import { exec } from 'child_process' +import { writeFileSync, existsSync, readFileSync } from 'fs' +import path from 'path' + +/** + * Path to the settings file + **/ +export const GPU_INFO_FILE = path.join( + getJanDataFolderPath(), + 'settings', + 'settings.json' +) + +/** + * Default GPU settings + * TODO: This needs to be refactored to support multiple accelerators + **/ +const DEFAULT_SETTINGS: GpuSetting = { + notify: true, + run_mode: 'cpu', + nvidia_driver: { + exist: false, + version: '', + }, + cuda: { + exist: false, + version: '', + }, + gpus: [], + gpu_highest_vram: '', + gpus_in_use: [], + is_initial: true, + // TODO: This needs to be set based on user toggle in settings + vulkan: false, +} + +export const getGpuConfig = async (): Promise => { + if (process.platform === 'darwin') return undefined + return JSON.parse(readFileSync(GPU_INFO_FILE, 'utf-8')) +} + +export const getResourcesInfo = async (): Promise => { + const ramUsedInfo = await mem.used() + const totalMemory = ramUsedInfo.totalMemMb * 1024 * 1024 + const usedMemory = ramUsedInfo.usedMemMb * 1024 * 1024 + + const resourceInfo: ResourceInfo = { + mem: { + totalMemory, + usedMemory, + }, + } + + return resourceInfo +} + +export const getCurrentLoad = () => + new Promise(async (resolve, reject) => { + const cpuPercentage = await cpu.usage() + let data = { + run_mode: 'cpu', + gpus_in_use: [], + } + + if (process.platform !== 'darwin') { + data = JSON.parse(readFileSync(GPU_INFO_FILE, 'utf-8')) + } + + if (data.run_mode === 'gpu' && data.gpus_in_use.length > 0) { + const gpuIds = data.gpus_in_use.join(',') + if (gpuIds !== '' && data['vulkan'] !== true) { + exec( + `nvidia-smi --query-gpu=index,name,temperature.gpu,utilization.gpu,memory.total,memory.free,utilization.memory --format=csv,noheader,nounits --id=${gpuIds}`, + (error, stdout, _) => { + if (error) { + console.error(`exec error: ${error}`) + throw new Error(error.message) + } + const gpuInfo: GpuInfo[] = stdout + .trim() + .split('\n') + .map((line) => { + const [ + id, + name, + temperature, + utilization, + memoryTotal, + memoryFree, + memoryUtilization, + ] = line.split(', ').map((item) => item.replace(/\r/g, '')) + return { + id, + name, + temperature, + utilization, + memoryTotal, + memoryFree, + memoryUtilization, + } + }) + + resolve({ + cpu: { usage: cpuPercentage }, + gpu: gpuInfo, + }) + } + ) + } else { + // Handle the case where gpuIds is empty + resolve({ + cpu: { usage: cpuPercentage }, + gpu: [], + }) + } + } else { + // Handle the case where run_mode is not 'gpu' or no GPUs are in use + resolve({ + cpu: { usage: cpuPercentage }, + gpu: [], + }) + } + }) + +/** + * This will retrive GPU informations and persist settings.json + * Will be called when the extension is loaded to turn on GPU acceleration if supported + */ +export const updateNvidiaInfo = async () => { + // ignore if macos + if (process.platform === 'darwin') return + + try { + JSON.parse(readFileSync(GPU_INFO_FILE, 'utf-8')) + } catch (error) { + writeFileSync(GPU_INFO_FILE, JSON.stringify(DEFAULT_SETTINGS, null, 2)) + } + + await updateNvidiaDriverInfo() + await updateGpuInfo() +} + +const updateNvidiaDriverInfo = async () => + new Promise((resolve, reject) => { + exec( + 'nvidia-smi --query-gpu=driver_version --format=csv,noheader', + (error, stdout) => { + const data: GpuSetting = JSON.parse( + readFileSync(GPU_INFO_FILE, 'utf-8') + ) + + if (!error) { + const firstLine = stdout.split('\n')[0].trim() + data.nvidia_driver.exist = true + data.nvidia_driver.version = firstLine + } else { + data.nvidia_driver.exist = false + } + + writeFileSync(GPU_INFO_FILE, JSON.stringify(data, null, 2)) + resolve({}) + } + ) + }) + +const getGpuArch = (gpuName: string): string => { + if (!gpuName.toLowerCase().includes('nvidia')) return 'unknown' + + if (gpuName.includes('20')) return 'turing' + else if (gpuName.includes('30')) return 'ampere' + else if (gpuName.includes('40')) return 'ada' + else return 'unknown' +} + +const updateGpuInfo = async () => + new Promise((resolve, reject) => { + let data: GpuSetting = JSON.parse(readFileSync(GPU_INFO_FILE, 'utf-8')) + + // Cuda + if (data.vulkan === true) { + // Vulkan + exec( + process.platform === 'win32' + ? `${__dirname}\\..\\bin\\vulkaninfoSDK.exe --summary` + : `${__dirname}/../bin/vulkaninfo --summary`, + (error, stdout) => { + if (!error) { + const output = stdout.toString() + + log(output) + const gpuRegex = /GPU(\d+):(?:[\s\S]*?)deviceName\s*=\s*(.*)/g + + const gpus: GpuSettingInfo[] = [] + let match + while ((match = gpuRegex.exec(output)) !== null) { + const id = match[1] + const name = match[2] + const arch = getGpuArch(name) + gpus.push({ id, vram: '0', name, arch }) + } + data.gpus = gpus + + if (!data.gpus_in_use || data.gpus_in_use.length === 0) { + data.gpus_in_use = [data.gpus.length > 1 ? '1' : '0'] + } + + data = updateCudaExistence(data) + writeFileSync(GPU_INFO_FILE, JSON.stringify(data, null, 2)) + resolve({}) + } else { + reject(error) + } + } + ) + } else { + exec( + 'nvidia-smi --query-gpu=index,memory.total,name --format=csv,noheader,nounits', + (error, stdout) => { + if (!error) { + log(stdout) + // Get GPU info and gpu has higher memory first + let highestVram = 0 + let highestVramId = '0' + const gpus: GpuSettingInfo[] = stdout + .trim() + .split('\n') + .map((line) => { + let [id, vram, name] = line.split(', ') + const arch = getGpuArch(name) + vram = vram.replace(/\r/g, '') + if (parseFloat(vram) > highestVram) { + highestVram = parseFloat(vram) + highestVramId = id + } + return { id, vram, name, arch } + }) + + data.gpus = gpus + data.gpu_highest_vram = highestVramId + } else { + data.gpus = [] + data.gpu_highest_vram = '' + } + + if (!data.gpus_in_use || data.gpus_in_use.length === 0) { + data.gpus_in_use = [data.gpu_highest_vram] + } + + data = updateCudaExistence(data) + writeFileSync(GPU_INFO_FILE, JSON.stringify(data, null, 2)) + resolve({}) + } + ) + } + }) + +/** + * Check if file exists in paths + */ +const checkFileExistenceInPaths = (file: string, paths: string[]): boolean => { + return paths.some((p) => existsSync(path.join(p, file))) +} + +/** + * Validate cuda for linux and windows + */ +const updateCudaExistence = ( + data: GpuSetting = DEFAULT_SETTINGS +): GpuSetting => { + let filesCuda12: string[] + let filesCuda11: string[] + let paths: string[] + let cudaVersion: string = '' + + if (process.platform === 'win32') { + filesCuda12 = ['cublas64_12.dll', 'cudart64_12.dll', 'cublasLt64_12.dll'] + filesCuda11 = ['cublas64_11.dll', 'cudart64_11.dll', 'cublasLt64_11.dll'] + paths = process.env.PATH ? process.env.PATH.split(path.delimiter) : [] + } else { + filesCuda12 = ['libcudart.so.12', 'libcublas.so.12', 'libcublasLt.so.12'] + filesCuda11 = ['libcudart.so.11.0', 'libcublas.so.11', 'libcublasLt.so.11'] + paths = process.env.LD_LIBRARY_PATH + ? process.env.LD_LIBRARY_PATH.split(path.delimiter) + : [] + paths.push('/usr/lib/x86_64-linux-gnu/') + } + + let cudaExists = filesCuda12.every( + (file) => existsSync(file) || checkFileExistenceInPaths(file, paths) + ) + + if (!cudaExists) { + cudaExists = filesCuda11.every( + (file) => existsSync(file) || checkFileExistenceInPaths(file, paths) + ) + if (cudaExists) { + cudaVersion = '11' + } + } else { + cudaVersion = '12' + } + + data.cuda.exist = cudaExists + data.cuda.version = cudaVersion + + console.debug(data.is_initial, data.gpus_in_use) + + if (cudaExists && data.is_initial && data.gpus_in_use.length > 0) { + data.run_mode = 'gpu' + } + + data.is_initial = false + return data +} diff --git a/extensions/monitoring-extension/webpack.config.js b/extensions/monitoring-extension/webpack.config.js deleted file mode 100644 index c8c3a34f7..000000000 --- a/extensions/monitoring-extension/webpack.config.js +++ /dev/null @@ -1,35 +0,0 @@ -const path = require('path') -const webpack = require('webpack') -const packageJson = require('./package.json') - -module.exports = { - experiments: { outputModule: true }, - entry: './src/index.ts', // Adjust the entry point to match your project's main file - mode: 'production', - module: { - rules: [ - { - test: /\.tsx?$/, - use: 'ts-loader', - exclude: /node_modules/, - }, - ], - }, - output: { - filename: 'index.js', // Adjust the output file name as needed - path: path.resolve(__dirname, 'dist'), - library: { type: 'module' }, // Specify ESM output format - }, - plugins: [ - new webpack.DefinePlugin({ - MODULE: JSON.stringify(`${packageJson.name}/${packageJson.module}`), - }), - ], - resolve: { - extensions: ['.ts', '.js'], - }, - optimization: { - minimize: false, - }, - // Add loaders and other configuration as needed for your project -} diff --git a/extensions/tensorrt-llm-extension/README.md b/extensions/tensorrt-llm-extension/README.md new file mode 100644 index 000000000..34a670516 --- /dev/null +++ b/extensions/tensorrt-llm-extension/README.md @@ -0,0 +1,79 @@ +# Tensorrt-LLM Extension + +Created using Jan extension example + +# Create a Jan Extension using Typescript + +Use this template to bootstrap the creation of a TypeScript Jan extension. 🚀 + +## Create Your Own Extension + +To create your own extension, you can use this repository as a template! Just follow the below instructions: + +1. Click the Use this template button at the top of the repository +2. Select Create a new repository +3. Select an owner and name for your new repository +4. Click Create repository +5. Clone your new repository + +## Initial Setup + +After you've cloned the repository to your local machine or codespace, you'll need to perform some initial setup steps before you can develop your extension. + +> [!NOTE] +> +> You'll need to have a reasonably modern version of +> [Node.js](https://nodejs.org) handy. If you are using a version manager like +> [`nodenv`](https://github.com/nodenv/nodenv) or +> [`nvm`](https://github.com/nvm-sh/nvm), you can run `nodenv install` in the +> root of your repository to install the version specified in +> [`package.json`](./package.json). Otherwise, 20.x or later should work! + +1. :hammer_and_wrench: Install the dependencies + + ```bash + npm install + ``` + +1. :building_construction: Package the TypeScript for distribution + + ```bash + npm run bundle + ``` + +1. :white_check_mark: Check your artifact + + There will be a tgz file in your extension directory now + +## Update the Extension Metadata + +The [`package.json`](package.json) file defines metadata about your extension, such as +extension name, main entry, description and version. + +When you copy this repository, update `package.json` with the name, description for your extension. + +## Update the Extension Code + +The [`src/`](./src/) directory is the heart of your extension! This contains the +source code that will be run when your extension functions are invoked. You can replace the +contents of this directory with your own code. + +There are a few things to keep in mind when writing your extension code: + +- Most Jan Extension functions are processed asynchronously. + In `index.ts`, you will see that the extension function will return a `Promise`. + + ```typescript + import { events, MessageEvent, MessageRequest } from '@janhq/core' + + function onStart(): Promise { + return events.on(MessageEvent.OnMessageSent, (data: MessageRequest) => + this.inference(data) + ) + } + ``` + + For more information about the Jan Extension Core module, see the + [documentation](https://github.com/janhq/jan/blob/main/core/README.md). + +So, what are you waiting for? Go ahead and start customizing your extension! diff --git a/extensions/tensorrt-llm-extension/models.json b/extensions/tensorrt-llm-extension/models.json new file mode 100644 index 000000000..bc6a78256 --- /dev/null +++ b/extensions/tensorrt-llm-extension/models.json @@ -0,0 +1,49 @@ +[ + { + "sources": [ + { + "filename": "config.json", + "url": "https://delta.jan.ai/dist/models///LlamaCorn-1.1B-Chat-fp16/config.json" + }, + { + "filename": "rank0.engine", + "url": "https://delta.jan.ai/dist/models///LlamaCorn-1.1B-Chat-fp16/rank0.engine" + }, + { + "filename": "tokenizer.model", + "url": "https://delta.jan.ai/dist/models///LlamaCorn-1.1B-Chat-fp16/tokenizer.model" + }, + { + "filename": "special_tokens_map.json", + "url": "https://delta.jan.ai/dist/models///LlamaCorn-1.1B-Chat-fp16/special_tokens_map.json" + }, + { + "filename": "tokenizer.json", + "url": "https://delta.jan.ai/dist/models///LlamaCorn-1.1B-Chat-fp16/tokenizer.json" + }, + { + "filename": "tokenizer_config.json", + "url": "https://delta.jan.ai/dist/models///LlamaCorn-1.1B-Chat-fp16/tokenizer_config.json" + } + ], + "id": "llamacorn-1.1b-chat-fp16", + "object": "model", + "name": "LlamaCorn 1.1B Chat FP16", + "version": "1.0", + "description": "LlamaCorn is a refined version of TinyLlama-1.1B, optimized for conversational quality, running on consumer devices through TensorRT-LLM", + "format": "TensorRT-LLM", + "settings": { + "ctx_len": 2048 + }, + "parameters": { + "stream": true, + "max_tokens": 4096 + }, + "metadata": { + "author": "LLama", + "tags": ["TensorRT-LLM", "1B", "Finetuned"], + "size": 2151000000 + }, + "engine": "nitro-tensorrt-llm" + } +] diff --git a/extensions/tensorrt-llm-extension/package.json b/extensions/tensorrt-llm-extension/package.json new file mode 100644 index 000000000..01ff3e2c6 --- /dev/null +++ b/extensions/tensorrt-llm-extension/package.json @@ -0,0 +1,75 @@ +{ + "name": "@janhq/tensorrt-llm-extension", + "version": "0.0.2", + "description": "Enables accelerated inference leveraging Nvidia's TensorRT-LLM for optimal GPU hardware optimizations. Compatible with models in TensorRT-LLM format. Requires Nvidia GPU driver and CUDA Toolkit installation.", + "main": "dist/index.js", + "node": "dist/node/index.cjs.js", + "author": "Jan ", + "license": "AGPL-3.0", + "config": { + "host": "127.0.0.1", + "port": "3928" + }, + "compatibility": { + "platform": [ + "win32", + "linux" + ], + "app": [ + "0.1.0" + ] + }, + "scripts": { + "build": "tsc --module commonjs && rollup -c rollup.config.ts", + "build:publish:win32": "rimraf *.tgz --glob && npm run build && cpx \"bin/**\" \"dist/bin\" && npm pack && cpx *.tgz ../../pre-install", + "build:publish:linux": "rimraf *.tgz --glob && npm run build && cpx \"bin/**\" \"dist/bin\" && npm pack && cpx *.tgz ../../pre-install", + "build:publish:darwin": "rimraf *.tgz --glob && npm run build && cpx \"bin/**\" \"dist/bin\" && npm pack && cpx *.tgz ../../pre-install", + "build:publish": "run-script-os" + }, + "exports": { + ".": "./dist/index.js", + "./main": "./dist/node/index.cjs.js" + }, + "devDependencies": { + "@rollup/plugin-commonjs": "^25.0.7", + "@rollup/plugin-json": "^6.1.0", + "@rollup/plugin-node-resolve": "^15.2.3", + "@rollup/plugin-replace": "^5.0.5", + "@types/node": "^20.11.4", + "@types/os-utils": "^0.0.4", + "@types/tcp-port-used": "^1.0.4", + "@types/decompress": "4.2.7", + "cpx": "^1.5.0", + "download-cli": "^1.1.1", + "rimraf": "^3.0.2", + "rollup": "^2.38.5", + "rollup-plugin-define": "^1.0.1", + "rollup-plugin-sourcemaps": "^0.6.3", + "rollup-plugin-typescript2": "^0.36.0", + "run-script-os": "^1.1.6", + "typescript": "^5.2.2" + }, + "dependencies": { + "@janhq/core": "file:../../core", + "decompress": "^4.2.1", + "fetch-retry": "^5.0.6", + "path-browserify": "^1.0.1", + "rxjs": "^7.8.1", + "tcp-port-used": "^1.0.2", + "ulid": "^2.3.0" + }, + "engines": { + "node": ">=18.0.0" + }, + "files": [ + "dist/*", + "package.json", + "README.md" + ], + "bundleDependencies": [ + "tcp-port-used", + "fetch-retry", + "decompress", + "@janhq/core" + ] +} diff --git a/extensions/tensorrt-llm-extension/rollup.config.ts b/extensions/tensorrt-llm-extension/rollup.config.ts new file mode 100644 index 000000000..33e45823b --- /dev/null +++ b/extensions/tensorrt-llm-extension/rollup.config.ts @@ -0,0 +1,73 @@ +import resolve from '@rollup/plugin-node-resolve' +import commonjs from '@rollup/plugin-commonjs' +import sourceMaps from 'rollup-plugin-sourcemaps' +import typescript from 'rollup-plugin-typescript2' +import json from '@rollup/plugin-json' +import replace from '@rollup/plugin-replace' +const packageJson = require('./package.json') + +export default [ + { + input: `src/index.ts`, + output: [{ file: packageJson.main, format: 'es', sourcemap: true }], + watch: { + include: 'src/**', + }, + plugins: [ + replace({ + EXTENSION_NAME: JSON.stringify(packageJson.name), + TENSORRT_VERSION: JSON.stringify('0.1.5'), + DOWNLOAD_RUNNER_URL: + process.platform === 'darwin' || process.platform === 'win32' + ? JSON.stringify( + 'https://github.com/janhq/nitro-tensorrt-llm/releases/download/windows-v/nitro-windows-v-amd64-tensorrt-llm-.tar.gz' + ) + : JSON.stringify( + 'https://github.com/janhq/nitro-tensorrt-llm/releases/download/linux-v/nitro-linux-v-amd64-tensorrt-llm-.tar.gz' + ), + NODE: JSON.stringify(`${packageJson.name}/${packageJson.node}`), + INFERENCE_URL: JSON.stringify( + process.env.INFERENCE_URL || + `${packageJson.config?.protocol ?? 'http'}://${packageJson.config?.host}:${packageJson.config?.port}/v1/chat/completions` + ), + COMPATIBILITY: JSON.stringify(packageJson.compatibility), + }), + json(), + typescript({ useTsconfigDeclarationDir: true }), + commonjs(), + resolve({ + extensions: ['.js', '.ts', '.svelte'], + }), + sourceMaps(), + ], + }, + { + input: `src/node/index.ts`, + output: [ + { file: 'dist/node/index.cjs.js', format: 'cjs', sourcemap: true }, + ], + external: ['@janhq/core/node'], + watch: { + include: 'src/node/**', + }, + plugins: [ + replace({ + LOAD_MODEL_URL: JSON.stringify( + `${packageJson.config?.protocol ?? 'http'}://${packageJson.config?.host}:${packageJson.config?.port}/inferences/tensorrtllm/loadmodel` + ), + TERMINATE_ENGINE_URL: JSON.stringify( + `${packageJson.config?.protocol ?? 'http'}://${packageJson.config?.host}:${packageJson.config?.port}/inferences/processmanager/destroy` + ), + ENGINE_HOST: JSON.stringify(packageJson.config?.host ?? '127.0.0.1'), + ENGINE_PORT: JSON.stringify(packageJson.config?.port ?? '3928'), + }), + json(), + typescript({ useTsconfigDeclarationDir: true }), + commonjs(), + resolve({ + extensions: ['.ts', '.js', '.json'], + }), + sourceMaps(), + ], + }, +] diff --git a/extensions/tensorrt-llm-extension/src/@types/global.d.ts b/extensions/tensorrt-llm-extension/src/@types/global.d.ts new file mode 100644 index 000000000..905e86380 --- /dev/null +++ b/extensions/tensorrt-llm-extension/src/@types/global.d.ts @@ -0,0 +1,10 @@ +declare const NODE: string +declare const INFERENCE_URL: string +declare const LOAD_MODEL_URL: string +declare const TERMINATE_ENGINE_URL: string +declare const ENGINE_HOST: string +declare const ENGINE_PORT: string +declare const DOWNLOAD_RUNNER_URL: string +declare const TENSORRT_VERSION: string +declare const COMPATIBILITY: object +declare const EXTENSION_NAME: string diff --git a/extensions/tensorrt-llm-extension/src/index.ts b/extensions/tensorrt-llm-extension/src/index.ts new file mode 100644 index 000000000..076951c3f --- /dev/null +++ b/extensions/tensorrt-llm-extension/src/index.ts @@ -0,0 +1,147 @@ +/** + * @module tensorrt-llm-extension/src/index + */ + +import { + Compatibility, + DownloadEvent, + DownloadRequest, + DownloadState, + GpuSetting, + InstallationState, + Model, + baseName, + downloadFile, + events, + executeOnMain, + joinPath, + showToast, + systemInformations, + LocalOAIEngine, + fs, +} from '@janhq/core' +import models from '../models.json' + +/** + * TensorRTLLMExtension - Implementation of LocalOAIEngine + * @extends BaseOAILocalInferenceProvider + * Provide pre-populated models for TensorRTLLM + */ +export default class TensorRTLLMExtension extends LocalOAIEngine { + /** + * Override custom function name for loading and unloading model + * Which are implemented from node module + */ + override provider = 'nitro-tensorrt-llm' + override inferenceUrl = INFERENCE_URL + override nodeModule = NODE + + private supportedGpuArch = ['turing', 'ampere', 'ada'] + + compatibility() { + return COMPATIBILITY as unknown as Compatibility + } + /** + * models implemented by the extension + * define pre-populated models + */ + async models(): Promise { + if ((await this.installationState()) === 'Installed') + return models as unknown as Model[] + return [] + } + + override async install(): Promise { + const info = await systemInformations() + console.debug( + `TensorRTLLMExtension installing pre-requisites... ${JSON.stringify(info)}` + ) + const gpuSetting: GpuSetting | undefined = info.gpuSetting + if (gpuSetting === undefined || gpuSetting.gpus.length === 0) { + console.error('No GPU setting found. Please check your GPU setting.') + return + } + + // TODO: we only check for the first graphics card. Need to refactor this later. + const firstGpu = gpuSetting.gpus[0] + if (!firstGpu.name.toLowerCase().includes('nvidia')) { + console.error('No Nvidia GPU found. Please check your GPU setting.') + return + } + + if (firstGpu.arch === undefined) { + console.error('No GPU architecture found. Please check your GPU setting.') + return + } + + if (!this.supportedGpuArch.includes(firstGpu.arch)) { + console.error( + `Your GPU: ${firstGpu} is not supported. Only 20xx, 30xx, 40xx series are supported.` + ) + return + } + + const binaryFolderPath = await executeOnMain( + this.nodeModule, + 'binaryFolder' + ) + if (!(await fs.existsSync(binaryFolderPath))) { + await fs.mkdirSync(binaryFolderPath) + } + + const placeholderUrl = DOWNLOAD_RUNNER_URL + const tensorrtVersion = TENSORRT_VERSION + + const url = placeholderUrl + .replace(//g, tensorrtVersion) + .replace(//g, firstGpu.arch) + + const tarball = await baseName(url) + + const tarballFullPath = await joinPath([binaryFolderPath, tarball]) + const downloadRequest: DownloadRequest = { + url, + localPath: tarballFullPath, + extensionId: EXTENSION_NAME, + downloadType: 'extension', + } + downloadFile(downloadRequest) + + // TODO: wrap this into a Promise + const onFileDownloadSuccess = async (state: DownloadState) => { + // if other download, ignore + if (state.fileName !== tarball) return + events.off(DownloadEvent.onFileDownloadSuccess, onFileDownloadSuccess) + await executeOnMain(this.nodeModule, 'decompressRunner', tarballFullPath) + events.emit(DownloadEvent.onFileUnzipSuccess, state) + + // Prepopulate models as soon as it's ready + this.prePopulateModels().then(() => { + showToast( + 'Extension installed successfully.', + 'New models are added to Model Hub.' + ) + }) + } + events.on(DownloadEvent.onFileDownloadSuccess, onFileDownloadSuccess) + } + + override async installationState(): Promise { + // For now, we just check the executable of nitro x tensor rt + const isNitroExecutableAvailable = await executeOnMain( + this.nodeModule, + 'isNitroExecutableAvailable' + ) + + return isNitroExecutableAvailable ? 'Installed' : 'NotInstalled' + } + + override onInferenceStopped() { + if (!this.isRunning) return + showToast( + 'Unable to Stop Inference', + 'The model does not support stopping inference.' + ) + return Promise.resolve() + } +} diff --git a/extensions/tensorrt-llm-extension/src/node/index.ts b/extensions/tensorrt-llm-extension/src/node/index.ts new file mode 100644 index 000000000..252468fc1 --- /dev/null +++ b/extensions/tensorrt-llm-extension/src/node/index.ts @@ -0,0 +1,191 @@ +import path from 'path' +import { ChildProcessWithoutNullStreams, spawn } from 'child_process' +import tcpPortUsed from 'tcp-port-used' +import fetchRT from 'fetch-retry' +import { log } from '@janhq/core/node' +import { existsSync } from 'fs' +import decompress from 'decompress' + +// Polyfill fetch with retry +const fetchRetry = fetchRT(fetch) + +/** + * The response object for model init operation. + */ +interface ModelLoadParams { + engine_path: string + ctx_len: number +} + +// The subprocess instance for Engine +let subprocess: ChildProcessWithoutNullStreams | undefined = undefined + +/** + * Initializes a engine subprocess to load a machine learning model. + * @param params - The model load settings. + */ +async function loadModel(params: any): Promise<{ error: Error | undefined }> { + // modelFolder is the absolute path to the running model folder + // e.g. ~/jan/models/llama-2 + let modelFolder = params.modelFolder + + const settings: ModelLoadParams = { + engine_path: modelFolder, + ctx_len: params.model.settings.ctx_len ?? 2048, + } + return runEngineAndLoadModel(settings) +} + +/** + * Stops a Engine subprocess. + */ +function unloadModel(): Promise { + const controller = new AbortController() + setTimeout(() => controller.abort(), 5000) + debugLog(`Request to kill engine`) + + subprocess?.kill() + return fetch(TERMINATE_ENGINE_URL, { + method: 'DELETE', + signal: controller.signal, + }) + .then(() => { + subprocess = undefined + }) + .catch(() => {}) // Do nothing with this attempt + .then(() => tcpPortUsed.waitUntilFree(parseInt(ENGINE_PORT), 300, 5000)) // Wait for port available + .then(() => debugLog(`Engine process is terminated`)) + .catch((err) => { + debugLog( + `Could not kill running process on port ${ENGINE_PORT}. Might be another process running on the same port? ${err}` + ) + throw 'PORT_NOT_AVAILABLE' + }) +} +/** + * 1. Spawn engine process + * 2. Load model into engine subprocess + * @returns + */ +async function runEngineAndLoadModel(settings: ModelLoadParams) { + return unloadModel() + .then(runEngine) + .then(() => loadModelRequest(settings)) + .catch((err) => { + // TODO: Broadcast error so app could display proper error message + debugLog(`${err}`, 'Error') + return { error: err } + }) +} + +/** + * Loads a LLM model into the Engine subprocess by sending a HTTP POST request. + */ +function loadModelRequest( + settings: ModelLoadParams +): Promise<{ error: Error | undefined }> { + debugLog(`Loading model with params ${JSON.stringify(settings)}`) + return fetchRetry(LOAD_MODEL_URL, { + method: 'POST', + headers: { + 'Content-Type': 'application/json', + }, + body: JSON.stringify(settings), + retries: 3, + retryDelay: 500, + }) + .then((res) => { + debugLog(`Load model success with response ${JSON.stringify(res)}`) + return Promise.resolve({ error: undefined }) + }) + .catch((err) => { + debugLog(`Load model failed with error ${err}`, 'Error') + return Promise.resolve({ error: err }) + }) +} + +/** + * Spawns engine subprocess. + */ +function runEngine(): Promise { + debugLog(`Spawning engine subprocess...`) + + return new Promise((resolve, reject) => { + // Current directory by default + let binaryFolder = path.join(__dirname, '..', 'bin') + // Binary path + const binary = path.join( + binaryFolder, + process.platform === 'win32' ? 'nitro.exe' : 'nitro' + ) + + const args: string[] = ['1', ENGINE_HOST, ENGINE_PORT] + // Execute the binary + debugLog(`Spawn nitro at path: ${binary}, and args: ${args}`) + subprocess = spawn(binary, args, { + cwd: binaryFolder, + env: { + ...process.env, + }, + }) + + // Handle subprocess output + subprocess.stdout.on('data', (data: any) => { + debugLog(`${data}`) + }) + + subprocess.stderr.on('data', (data: any) => { + debugLog(`${data}`) + }) + + subprocess.on('close', (code: any) => { + debugLog(`Engine exited with code: ${code}`) + subprocess = undefined + reject(`child process exited with code ${code}`) + }) + + tcpPortUsed.waitUntilUsed(parseInt(ENGINE_PORT), 300, 30000).then(() => { + debugLog(`Engine is ready`) + resolve() + }) + }) +} + +function debugLog(message: string, level: string = 'Debug') { + log(`[TENSORRT_LLM_NITRO]::${level}:${message}`) +} + +const binaryFolder = async (): Promise => { + return path.join(__dirname, '..', 'bin') +} + +const decompressRunner = async (zipPath: string) => { + const output = path.join(__dirname, '..', 'bin') + console.debug(`Decompressing ${zipPath} to ${output}...`) + try { + const files = await decompress(zipPath, output) + console.debug('Decompress finished!', files) + } catch (err) { + console.error(`Decompress ${zipPath} failed: ${err}`) + } +} + +const isNitroExecutableAvailable = async (): Promise => { + const binary = path.join( + __dirname, + '..', + 'bin', + process.platform === 'win32' ? 'nitro.exe' : 'nitro' + ) + + return existsSync(binary) +} + +export default { + binaryFolder, + decompressRunner, + loadModel, + unloadModel, + dispose: unloadModel, + isNitroExecutableAvailable, +} diff --git a/extensions/tensorrt-llm-extension/tsconfig.json b/extensions/tensorrt-llm-extension/tsconfig.json new file mode 100644 index 000000000..478a05728 --- /dev/null +++ b/extensions/tensorrt-llm-extension/tsconfig.json @@ -0,0 +1,20 @@ +{ + "compilerOptions": { + "moduleResolution": "node", + "target": "es5", + "module": "ES2020", + "lib": ["es2015", "es2016", "es2017", "dom"], + "strict": true, + "sourceMap": true, + "declaration": true, + "allowSyntheticDefaultImports": true, + "experimentalDecorators": true, + "emitDecoratorMetadata": true, + "declarationDir": "dist/types", + "outDir": "dist", + "importHelpers": true, + "resolveJsonModule": true, + "typeRoots": ["node_modules/@types"] + }, + "include": ["src"] +} diff --git a/web/containers/DropdownListSidebar/index.tsx b/web/containers/DropdownListSidebar/index.tsx index c05d26e51..ef7532063 100644 --- a/web/containers/DropdownListSidebar/index.tsx +++ b/web/containers/DropdownListSidebar/index.tsx @@ -73,8 +73,9 @@ const DropdownListSidebar = ({ const [copyId, setCopyId] = useState('') + // TODO: Update filter condition for the local model const localModel = downloadedModels.filter( - (model) => model.engine === InferenceEngine.nitro + (model) => model.engine !== InferenceEngine.openai ) const remoteModel = downloadedModels.filter( (model) => model.engine === InferenceEngine.openai diff --git a/web/containers/Layout/BottomBar/InstallingExtension/InstallingExtensionModal.tsx b/web/containers/Layout/BottomBar/InstallingExtension/InstallingExtensionModal.tsx new file mode 100644 index 000000000..d46764e88 --- /dev/null +++ b/web/containers/Layout/BottomBar/InstallingExtension/InstallingExtensionModal.tsx @@ -0,0 +1,87 @@ +import { useCallback, useEffect } from 'react' + +import { abortDownload } from '@janhq/core' +import { + Button, + Modal, + ModalContent, + ModalHeader, + ModalTitle, + Progress, +} from '@janhq/uikit' +import { atom, useAtom, useAtomValue } from 'jotai' + +import { + formatDownloadPercentage, + formatExtensionsName, +} from '@/utils/converter' + +import { + InstallingExtensionState, + installingExtensionAtom, +} from '@/helpers/atoms/Extension.atom' + +export const showInstallingExtensionModalAtom = atom(false) + +const InstallingExtensionModal: React.FC = () => { + const [showInstallingExtensionModal, setShowInstallingExtensionModal] = + useAtom(showInstallingExtensionModalAtom) + const installingExtensions = useAtomValue(installingExtensionAtom) + + useEffect(() => { + if (installingExtensions.length === 0) { + setShowInstallingExtensionModal(false) + } + }, [installingExtensions, setShowInstallingExtensionModal]) + + const onAbortInstallingExtensionClick = useCallback( + (item: InstallingExtensionState) => { + if (item.localPath) { + abortDownload(item.localPath) + } + }, + [] + ) + + return ( + setShowInstallingExtensionModal(false)} + > + + + Installing Extension + + {Object.values(installingExtensions).map((item) => ( +
+ +
+
+

+ {formatExtensionsName(item.extensionId)} +

+ {formatDownloadPercentage(item.percentage)} +
+ +
+
+ ))} + + + ) +} + +export default InstallingExtensionModal diff --git a/web/containers/Layout/BottomBar/InstallingExtension/index.tsx b/web/containers/Layout/BottomBar/InstallingExtension/index.tsx new file mode 100644 index 000000000..05e803881 --- /dev/null +++ b/web/containers/Layout/BottomBar/InstallingExtension/index.tsx @@ -0,0 +1,52 @@ +import { Fragment, useCallback } from 'react' + +import { Progress } from '@janhq/uikit' +import { useAtomValue, useSetAtom } from 'jotai' + +import { showInstallingExtensionModalAtom } from './InstallingExtensionModal' + +import { installingExtensionAtom } from '@/helpers/atoms/Extension.atom' + +const InstallingExtension: React.FC = () => { + const installingExtensions = useAtomValue(installingExtensionAtom) + const setShowInstallingExtensionModal = useSetAtom( + showInstallingExtensionModalAtom + ) + const shouldShowInstalling = installingExtensions.length > 0 + + let totalPercentage = 0 + let totalExtensions = 0 + for (const installation of installingExtensions) { + totalPercentage += installation.percentage + totalExtensions++ + } + const progress = (totalPercentage / totalExtensions) * 100 + + const onClick = useCallback(() => { + setShowInstallingExtensionModal(true) + }, [setShowInstallingExtensionModal]) + + return ( + + {shouldShowInstalling ? ( +
+

+ Installing Extension +

+ +
+ + + {progress.toFixed(2)}% + +
+
+ ) : null} + + ) +} + +export default InstallingExtension diff --git a/web/containers/Layout/BottomBar/index.tsx b/web/containers/Layout/BottomBar/index.tsx index 2373ac3d4..3683d23db 100644 --- a/web/containers/Layout/BottomBar/index.tsx +++ b/web/containers/Layout/BottomBar/index.tsx @@ -16,6 +16,7 @@ import ProgressBar from '@/containers/ProgressBar' import { appDownloadProgress } from '@/containers/Providers/Jotai' import ImportingModelState from './ImportingModelState' +import InstallingExtension from './InstallingExtension' import SystemMonitor from './SystemMonitor' import UpdatedFailedModal from './UpdateFailedModal' @@ -46,6 +47,7 @@ const BottomBar = () => { +
diff --git a/web/containers/Layout/index.tsx b/web/containers/Layout/index.tsx index 7e3ad38ab..fb08bc6ac 100644 --- a/web/containers/Layout/index.tsx +++ b/web/containers/Layout/index.tsx @@ -22,6 +22,8 @@ import ImportModelOptionModal from '@/screens/Settings/ImportModelOptionModal' import ImportingModelModal from '@/screens/Settings/ImportingModelModal' import SelectingModelModal from '@/screens/Settings/SelectingModelModal' +import InstallingExtensionModal from './BottomBar/InstallingExtension/InstallingExtensionModal' + import { mainViewStateAtom } from '@/helpers/atoms/App.atom' const BaseLayout = (props: PropsWithChildren) => { @@ -68,6 +70,7 @@ const BaseLayout = (props: PropsWithChildren) => { {importModelStage === 'IMPORTING_MODEL' && } {importModelStage === 'EDIT_MODEL_INFO' && } {importModelStage === 'CONFIRM_CANCEL' && } +
) } diff --git a/web/containers/Providers/EventListener.tsx b/web/containers/Providers/EventListener.tsx index bfc87917b..20fc6dde2 100644 --- a/web/containers/Providers/EventListener.tsx +++ b/web/containers/Providers/EventListener.tsx @@ -7,6 +7,10 @@ import { useSetAtom } from 'jotai' import { setDownloadStateAtom } from '@/hooks/useDownloadState' +import { formatExtensionsName } from '@/utils/converter' + +import { toaster } from '../Toast' + import AppUpdateListener from './AppUpdateListener' import ClipboardListener from './ClipboardListener' import EventHandler from './EventHandler' @@ -14,46 +18,89 @@ import EventHandler from './EventHandler' import ModelImportListener from './ModelImportListener' import QuickAskListener from './QuickAskListener' +import { + InstallingExtensionState, + removeInstallingExtensionAtom, + setInstallingExtensionAtom, +} from '@/helpers/atoms/Extension.atom' + const EventListenerWrapper = ({ children }: PropsWithChildren) => { const setDownloadState = useSetAtom(setDownloadStateAtom) + const setInstallingExtension = useSetAtom(setInstallingExtensionAtom) + const removeInstallingExtension = useSetAtom(removeInstallingExtensionAtom) const onFileDownloadUpdate = useCallback( async (state: DownloadState) => { console.debug('onFileDownloadUpdate', state) - setDownloadState(state) + if (state.downloadType === 'extension') { + const installingExtensionState: InstallingExtensionState = { + extensionId: state.extensionId!, + percentage: state.percent, + localPath: state.localPath, + } + setInstallingExtension(state.extensionId!, installingExtensionState) + } else { + setDownloadState(state) + } }, - [setDownloadState] + [setDownloadState, setInstallingExtension] ) const onFileDownloadError = useCallback( (state: DownloadState) => { console.debug('onFileDownloadError', state) - setDownloadState(state) + if (state.downloadType === 'extension') { + removeInstallingExtension(state.extensionId!) + } else { + setDownloadState(state) + } }, - [setDownloadState] + [setDownloadState, removeInstallingExtension] ) const onFileDownloadSuccess = useCallback( (state: DownloadState) => { console.debug('onFileDownloadSuccess', state) - setDownloadState(state) + if (state.downloadType !== 'extension') { + setDownloadState(state) + } }, [setDownloadState] ) + const onFileUnzipSuccess = useCallback( + (state: DownloadState) => { + console.debug('onFileUnzipSuccess', state) + toaster({ + title: 'Success', + description: `Install ${formatExtensionsName(state.extensionId!)} successfully.`, + type: 'success', + }) + removeInstallingExtension(state.extensionId!) + }, + [removeInstallingExtension] + ) + useEffect(() => { console.debug('EventListenerWrapper: registering event listeners...') events.on(DownloadEvent.onFileDownloadUpdate, onFileDownloadUpdate) events.on(DownloadEvent.onFileDownloadError, onFileDownloadError) events.on(DownloadEvent.onFileDownloadSuccess, onFileDownloadSuccess) + events.on(DownloadEvent.onFileUnzipSuccess, onFileUnzipSuccess) return () => { console.debug('EventListenerWrapper: unregistering event listeners...') events.off(DownloadEvent.onFileDownloadUpdate, onFileDownloadUpdate) events.off(DownloadEvent.onFileDownloadError, onFileDownloadError) events.off(DownloadEvent.onFileDownloadSuccess, onFileDownloadSuccess) + events.off(DownloadEvent.onFileUnzipSuccess, onFileUnzipSuccess) } - }, [onFileDownloadUpdate, onFileDownloadError, onFileDownloadSuccess]) + }, [ + onFileDownloadUpdate, + onFileDownloadError, + onFileDownloadSuccess, + onFileUnzipSuccess, + ]) return ( diff --git a/web/extension/ExtensionManager.ts b/web/extension/ExtensionManager.ts index 1259021f7..c976010c6 100644 --- a/web/extension/ExtensionManager.ts +++ b/web/extension/ExtensionManager.ts @@ -23,7 +23,9 @@ export class ExtensionManager { * @param type - The type of the extension to retrieve. * @returns The extension, if found. */ - get(type: ExtensionTypeEnum): T | undefined { + get( + type: ExtensionTypeEnum | string + ): T | undefined { return this.extensions.get(type) as T | undefined } diff --git a/web/helpers/atoms/Extension.atom.ts b/web/helpers/atoms/Extension.atom.ts new file mode 100644 index 000000000..7af755e35 --- /dev/null +++ b/web/helpers/atoms/Extension.atom.ts @@ -0,0 +1,40 @@ +import { atom } from 'jotai' + +type ExtensionId = string + +export type InstallingExtensionState = { + extensionId: ExtensionId + percentage: number + localPath?: string +} + +export const installingExtensionAtom = atom([]) + +export const setInstallingExtensionAtom = atom( + null, + (get, set, extensionId: string, state: InstallingExtensionState) => { + const current = get(installingExtensionAtom) + + const isExists = current.some((e) => e.extensionId === extensionId) + if (isExists) { + const newCurrent = current.map((e) => { + if (e.extensionId === extensionId) { + return state + } + return e + }) + set(installingExtensionAtom, newCurrent) + } else { + set(installingExtensionAtom, [...current, state]) + } + } +) + +export const removeInstallingExtensionAtom = atom( + null, + (get, set, extensionId: string) => { + const current = get(installingExtensionAtom) + const newCurrent = current.filter((e) => e.extensionId !== extensionId) + set(installingExtensionAtom, newCurrent) + } +) diff --git a/web/hooks/useActiveModel.ts b/web/hooks/useActiveModel.ts index 600e10783..e6c519f9f 100644 --- a/web/hooks/useActiveModel.ts +++ b/web/hooks/useActiveModel.ts @@ -40,6 +40,16 @@ export function useActiveModel() { console.debug(`Model ${modelId} is already initialized. Ignore..`) return } + + let model = downloadedModelsRef?.current.find((e) => e.id === modelId) + + // Switch between engines + if (model && activeModel && activeModel.engine !== model.engine) { + stopModel() + // TODO: Refactor inference provider would address this + await new Promise((res) => setTimeout(res, 1000)) + } + // TODO: incase we have multiple assistants, the configuration will be from assistant setLoadModelError(undefined) @@ -47,8 +57,6 @@ export function useActiveModel() { setStateModel({ state: 'start', loading: true, model: modelId }) - let model = downloadedModelsRef?.current.find((e) => e.id === modelId) - if (!model) { toaster({ title: `Model ${modelId} not found!`, diff --git a/web/hooks/useDownloadModel.ts b/web/hooks/useDownloadModel.ts index 9f6334c71..d0d13d93b 100644 --- a/web/hooks/useDownloadModel.ts +++ b/web/hooks/useDownloadModel.ts @@ -8,12 +8,15 @@ import { joinPath, ModelArtifact, DownloadState, + GpuSetting, } from '@janhq/core' import { useAtomValue, useSetAtom } from 'jotai' import { setDownloadStateAtom } from './useDownloadState' +import useGpuSetting from './useGpuSetting' + import { extensionManager } from '@/extension/ExtensionManager' import { ignoreSslAtom, @@ -29,6 +32,8 @@ export default function useDownloadModel() { const setDownloadState = useSetAtom(setDownloadStateAtom) const addDownloadingModel = useSetAtom(addDownloadingModelAtom) + const { getGpuSettings } = useGpuSetting() + const downloadModel = useCallback( async (model: Model) => { const childProgresses: DownloadState[] = model.sources.map( @@ -68,10 +73,22 @@ export default function useDownloadModel() { }) addDownloadingModel(model) - - await localDownloadModel(model, ignoreSSL, proxyEnabled ? proxy : '') + const gpuSettings = await getGpuSettings() + await localDownloadModel( + model, + ignoreSSL, + proxyEnabled ? proxy : '', + gpuSettings + ) }, - [ignoreSSL, proxy, proxyEnabled, addDownloadingModel, setDownloadState] + [ + ignoreSSL, + proxy, + proxyEnabled, + getGpuSettings, + addDownloadingModel, + setDownloadState, + ] ) const abortModelDownload = useCallback(async (model: Model) => { @@ -90,8 +107,9 @@ export default function useDownloadModel() { const localDownloadModel = async ( model: Model, ignoreSSL: boolean, - proxy: string + proxy: string, + gpuSettings?: GpuSetting ) => extensionManager .get(ExtensionTypeEnum.Model) - ?.downloadModel(model, { ignoreSSL, proxy }) + ?.downloadModel(model, gpuSettings, { ignoreSSL, proxy }) diff --git a/web/hooks/useDownloadState.ts b/web/hooks/useDownloadState.ts index 06de9bef6..03a8883cb 100644 --- a/web/hooks/useDownloadState.ts +++ b/web/hooks/useDownloadState.ts @@ -18,123 +18,129 @@ export const modelDownloadStateAtom = atom>({}) export const setDownloadStateAtom = atom( null, (get, set, state: DownloadState) => { - const currentState = { ...get(modelDownloadStateAtom) } + try { + const currentState = { ...get(modelDownloadStateAtom) } - if (state.downloadState === 'end') { - const modelDownloadState = currentState[state.modelId] + if (state.downloadState === 'end') { + const modelDownloadState = currentState[state.modelId] - const updatedChildren: DownloadState[] = - modelDownloadState.children!.filter( - (m) => m.fileName !== state.fileName + const updatedChildren: DownloadState[] = ( + modelDownloadState.children ?? [] + ).filter((m) => m.fileName !== state.fileName) + updatedChildren.push(state) + modelDownloadState.children = updatedChildren + currentState[state.modelId] = modelDownloadState + + const isAllChildrenDownloadEnd = modelDownloadState.children?.every( + (m) => m.downloadState === 'end' ) - updatedChildren.push(state) - modelDownloadState.children = updatedChildren - currentState[state.modelId] = modelDownloadState - const isAllChildrenDownloadEnd = modelDownloadState.children?.every( - (m) => m.downloadState === 'end' - ) + if (isAllChildrenDownloadEnd) { + // download successfully + delete currentState[state.modelId] + set(removeDownloadingModelAtom, state.modelId) - if (isAllChildrenDownloadEnd) { - // download successfully + const model = get(configuredModelsAtom).find( + (e) => e.id === state.modelId + ) + if (model) set(downloadedModelsAtom, (prev) => [...prev, model]) + toaster({ + title: 'Download Completed', + description: `Download ${state.modelId} completed`, + type: 'success', + }) + } + } else if (state.downloadState === 'error') { + // download error delete currentState[state.modelId] set(removeDownloadingModelAtom, state.modelId) - - const model = get(configuredModelsAtom).find( - (e) => e.id === state.modelId - ) - if (model) set(downloadedModelsAtom, (prev) => [...prev, model]) - toaster({ - title: 'Download Completed', - description: `Download ${state.modelId} completed`, - type: 'success', - }) - } - } else if (state.downloadState === 'error') { - // download error - delete currentState[state.modelId] - set(removeDownloadingModelAtom, state.modelId) - if (state.error === 'aborted') { - toaster({ - title: 'Cancel Download', - description: `Model ${state.modelId} download cancelled`, - type: 'warning', - }) - } else { - let error = state.error - if ( - typeof error?.includes === 'function' && - state.error?.includes('certificate') - ) { - error += - '. To fix enable "Ignore SSL Certificates" in Advanced settings.' + if (state.error === 'aborted') { + toaster({ + title: 'Cancel Download', + description: `Model ${state.modelId} download cancelled`, + type: 'warning', + }) + } else { + let error = state.error + if ( + typeof error?.includes === 'function' && + state.error?.includes('certificate') + ) { + error += + '. To fix enable "Ignore SSL Certificates" in Advanced settings.' + } + toaster({ + title: 'Download Failed', + description: `Model ${state.modelId} download failed: ${error}`, + type: 'error', + }) + } + } else { + // download in progress + if (state.size.total === 0) { + // this is initial state, just set the state + currentState[state.modelId] = state + set(modelDownloadStateAtom, currentState) + return } - toaster({ - title: 'Download Failed', - description: `Model ${state.modelId} download failed: ${error}`, - type: 'error', - }) - } - } else { - // download in progress - if (state.size.total === 0) { - // this is initial state, just set the state - currentState[state.modelId] = state - set(modelDownloadStateAtom, currentState) - return - } - const modelDownloadState = currentState[state.modelId] - if (!modelDownloadState) { - console.debug('setDownloadStateAtom: modelDownloadState not found') - return - } + const modelDownloadState = currentState[state.modelId] + if (!modelDownloadState) { + console.debug('setDownloadStateAtom: modelDownloadState not found') + return + } - // delete the children if the filename is matched and replace the new state - const updatedChildren: DownloadState[] = - modelDownloadState.children!.filter( - (m) => m.fileName !== state.fileName + // delete the children if the filename is matched and replace the new state + const updatedChildren: DownloadState[] = ( + modelDownloadState.children ?? [] + ).filter((m) => m.fileName !== state.fileName) + + updatedChildren.push(state) + + // re-calculate the overall progress if we have all the children download data + const isAnyChildDownloadNotReady = updatedChildren.some( + (m) => + m.size.total === 0 && + !modelDownloadState.children?.some( + (e) => e.fileName === m.fileName && e.downloadState === 'end' + ) && + modelDownloadState.children?.some((e) => e.fileName === m.fileName) ) - updatedChildren.push(state) + modelDownloadState.children = updatedChildren + if (isAnyChildDownloadNotReady) { + // just update the children + currentState[state.modelId] = modelDownloadState + set(modelDownloadStateAtom, currentState) + return + } - // re-calculate the overall progress if we have all the children download data - const isAnyChildDownloadNotReady = updatedChildren.some( - (m) => m.size.total === 0 - ) + const parentTotalSize = modelDownloadState.size.total + if (parentTotalSize === 0) { + // calculate the total size of the parent by sum all children total size + const totalSize = updatedChildren.reduce( + (acc, m) => acc + m.size.total, + 0 + ) - modelDownloadState.children = updatedChildren + modelDownloadState.size.total = totalSize + } - if (isAnyChildDownloadNotReady) { - // just update the children - currentState[state.modelId] = modelDownloadState - set(modelDownloadStateAtom, currentState) - - return - } - - const parentTotalSize = modelDownloadState.size.total - if (parentTotalSize === 0) { - // calculate the total size of the parent by sum all children total size - const totalSize = updatedChildren.reduce( - (acc, m) => acc + m.size.total, + // calculate the total transferred size by sum all children transferred size + const transferredSize = updatedChildren.reduce( + (acc, m) => acc + m.size.transferred, 0 ) - - modelDownloadState.size.total = totalSize + modelDownloadState.size.transferred = transferredSize + modelDownloadState.percent = + parentTotalSize === 0 ? 0 : transferredSize / parentTotalSize + currentState[state.modelId] = modelDownloadState } - // calculate the total transferred size by sum all children transferred size - const transferredSize = updatedChildren.reduce( - (acc, m) => acc + m.size.transferred, - 0 - ) - modelDownloadState.size.transferred = transferredSize - modelDownloadState.percent = - parentTotalSize === 0 ? 0 : transferredSize / parentTotalSize - currentState[state.modelId] = modelDownloadState + set(modelDownloadStateAtom, currentState) + } catch (e) { + console.debug('setDownloadStateAtom: state', state) + console.debug('setDownloadStateAtom: error', e) } - - set(modelDownloadStateAtom, currentState) } ) diff --git a/web/hooks/useGpuSetting.ts b/web/hooks/useGpuSetting.ts new file mode 100644 index 000000000..36f51ed57 --- /dev/null +++ b/web/hooks/useGpuSetting.ts @@ -0,0 +1,21 @@ +import { useCallback } from 'react' + +import { ExtensionTypeEnum, MonitoringExtension } from '@janhq/core' + +import { extensionManager } from '@/extension' + +export default function useGpuSetting() { + const getGpuSettings = useCallback(async () => { + const gpuSetting = await extensionManager + ?.get(ExtensionTypeEnum.SystemMonitoring) + ?.getGpuSetting() + + if (!gpuSetting) { + console.debug('No GPU setting found') + return undefined + } + return gpuSetting + }, []) + + return { getGpuSettings } +} diff --git a/web/next.config.js b/web/next.config.js index a4b3e6d43..48ea0703e 100644 --- a/web/next.config.js +++ b/web/next.config.js @@ -38,6 +38,7 @@ const nextConfig = { isMac: process.platform === 'darwin', isWindows: process.platform === 'win32', isLinux: process.platform === 'linux', + PLATFORM: JSON.stringify(process.platform), }), ] return config diff --git a/web/screens/ExploreModels/ExploreModelItem/index.tsx b/web/screens/ExploreModels/ExploreModelItem/index.tsx index 9cdfbc01a..e8887e70f 100644 --- a/web/screens/ExploreModels/ExploreModelItem/index.tsx +++ b/web/screens/ExploreModels/ExploreModelItem/index.tsx @@ -3,6 +3,8 @@ import { useState } from 'react' import { Model } from '@janhq/core' import { Badge } from '@janhq/uikit' +import { twMerge } from 'tailwind-merge' + import ExploreModelItemHeader from '@/screens/ExploreModels/ExploreModelItemHeader' type Props = { @@ -75,7 +77,16 @@ const ExploreModelItem: React.FC = ({ model }) => { Format -

{model.format}

+

+ {model.format} +

diff --git a/web/screens/ExploreModels/ExploreModelItemHeader/index.tsx b/web/screens/ExploreModels/ExploreModelItemHeader/index.tsx index 38e7f65a6..581a628ba 100644 --- a/web/screens/ExploreModels/ExploreModelItemHeader/index.tsx +++ b/web/screens/ExploreModels/ExploreModelItemHeader/index.tsx @@ -152,6 +152,7 @@ const ExploreModelItemHeader: React.FC = ({ model, onClick, open }) => {
{model.name} +
@@ -172,4 +173,21 @@ const ExploreModelItemHeader: React.FC = ({ model, onClick, open }) => { ) } +type EngineBadgeProps = { + engine: string +} + +const EngineBadge: React.FC = ({ engine }) => { + switch (engine) { + case 'nitro-tensorrt-llm': + return ( +
+ TensorRT-LLM +
+ ) + default: + return null + } +} + export default ExploreModelItemHeader diff --git a/web/screens/Settings/CoreExtensions/TensorRtExtensionItem.tsx b/web/screens/Settings/CoreExtensions/TensorRtExtensionItem.tsx new file mode 100644 index 000000000..e4c8ae04f --- /dev/null +++ b/web/screens/Settings/CoreExtensions/TensorRtExtensionItem.tsx @@ -0,0 +1,225 @@ +import { useCallback, useEffect, useState } from 'react' + +import { + Compatibility, + GpuSetting, + InstallationState, + abortDownload, + systemInformations, +} from '@janhq/core' +import { + Button, + Progress, + Tooltip, + TooltipArrow, + TooltipContent, + TooltipPortal, + TooltipTrigger, +} from '@janhq/uikit' + +import { InfoCircledIcon } from '@radix-ui/react-icons' +import { useAtomValue } from 'jotai' + +import { extensionManager } from '@/extension' +import Extension from '@/extension/Extension' +import { installingExtensionAtom } from '@/helpers/atoms/Extension.atom' + +type Props = { + item: Extension +} + +const TensorRtExtensionItem: React.FC = ({ item }) => { + const [compatibility, setCompatibility] = useState( + undefined + ) + const [installState, setInstallState] = + useState('NotRequired') + const installingExtensions = useAtomValue(installingExtensionAtom) + const [isGpuSupported, setIsGpuSupported] = useState(false) + + const isInstalling = installingExtensions.some( + (e) => e.extensionId === item.name + ) + + const progress = isInstalling + ? installingExtensions.find((e) => e.extensionId === item.name) + ?.percentage ?? -1 + : -1 + + useEffect(() => { + const getSystemInfos = async () => { + const info = await systemInformations() + if (!info) { + setIsGpuSupported(false) + return + } + + const gpuSettings: GpuSetting | undefined = info.gpuSetting + if (!gpuSettings || gpuSettings.gpus.length === 0) { + setIsGpuSupported(false) + return + } + + const arch = gpuSettings.gpus[0].arch + if (!arch) { + setIsGpuSupported(false) + return + } + + const supportedGpuArch = ['turing', 'ampere', 'ada'] + setIsGpuSupported(supportedGpuArch.includes(arch)) + } + getSystemInfos() + }, []) + + useEffect(() => { + const getExtensionInstallationState = async () => { + const extension = extensionManager.get(item.name ?? '') + if (!extension) return + + if (typeof extension?.installationState === 'function') { + const installState = await extension.installationState() + setInstallState(installState) + } + } + + getExtensionInstallationState() + }, [item.name, isInstalling]) + + useEffect(() => { + const extension = extensionManager.get(item.name ?? '') + if (!extension) return + setCompatibility(extension.compatibility()) + }, [setCompatibility, item.name]) + + const onInstallClick = useCallback(async () => { + const extension = extensionManager.get(item.name ?? '') + if (!extension) return + + await extension.install() + }, [item.name]) + + const onCancelInstallingClick = () => { + const extension = installingExtensions.find( + (e) => e.extensionId === item.name + ) + if (extension?.localPath) { + abortDownload(extension.localPath) + } + } + + return ( +
+
+
+
+ TensorRT-LLM Extension +
+

+ v{item.version} +

+
+

+ {item.description} +

+
+ {(!compatibility || compatibility['platform']?.includes(PLATFORM)) && + isGpuSupported ? ( +
+ +
+ ) : ( +
+
+ Incompatible{' '} + + + + + + + {compatibility ? ( + + Only available on{' '} + {compatibility?.platform + ?.map((e: string) => + e === 'win32' + ? 'Windows' + : e === 'linux' + ? 'Linux' + : 'MacOS' + ) + .join(', ')} + + ) : ( + + Your GPUs are not compatible with this extension + + )} + + + + +
+
+ )} +
+ ) +} + +type InstallStateProps = { + installProgress: number + installState: InstallationState + onInstallClick: () => void + onCancelClick: () => void +} + +const InstallStateIndicator: React.FC = ({ + installProgress, + installState, + onInstallClick, + onCancelClick, +}) => { + // TODO: NamH support dark mode for this + if (installProgress !== -1) { + const progress = installProgress * 100 + return ( +
+ +
+ + + {progress.toFixed(0)}% + +
+
+ ) + } + + // TODO: NamH check for dark mode here + switch (installState) { + case 'Installed': + return ( +
+ Installed +
+ ) + case 'NotInstalled': + return ( + + ) + default: + return
+ } +} + +export default TensorRtExtensionItem diff --git a/web/screens/Settings/CoreExtensions/index.tsx b/web/screens/Settings/CoreExtensions/index.tsx index 8c9f92d7a..f5b66abeb 100644 --- a/web/screens/Settings/CoreExtensions/index.tsx +++ b/web/screens/Settings/CoreExtensions/index.tsx @@ -4,13 +4,18 @@ import React, { useState, useEffect, useRef } from 'react' import { Button, ScrollArea } from '@janhq/uikit' +import Loader from '@/containers/Loader' + import { formatExtensionsName } from '@/utils/converter' +import TensorRtExtensionItem from './TensorRtExtensionItem' + import { extensionManager } from '@/extension' import Extension from '@/extension/Extension' const ExtensionCatalog = () => { const [activeExtensions, setActiveExtensions] = useState([]) + const [showLoading, setShowLoading] = useState(false) const fileInputRef = useRef(null) /** * Fetches the active extensions and their preferences from the `extensions` and `preferences` modules. @@ -63,65 +68,76 @@ const ExtensionCatalog = () => { const handleFileChange = (event: React.ChangeEvent) => { const file = event.target.files?.[0] if (file) { + setShowLoading(true) install(event) } } return ( - -
- {activeExtensions.map((item, i) => { - return ( -
-
-
-
- {formatExtensionsName(item.name ?? item.description ?? '')} -
-

- v{item.version} + <> + +

+ {activeExtensions.map((item, i) => { + // TODO: this is bad code, rewrite it + if (item.name === '@janhq/tensorrt-llm-extension') { + return + } + + return ( +
+
+
+
+ {formatExtensionsName( + item.name ?? item.description ?? '' + )} +
+

+ v{item.version} +

+
+

+ {item.description}

-

- {item.description} -

+ ) + })} + {/* Manual Installation */} +
+
+
+
+ Manual Installation +
+
+

+ Select a extension file to install (.tgz) +

- ) - })} - {/* Manual Installation */} -
-
-
-
- Manual Installation -
+
+ +
-

- Select a extension file to install (.tgz) -

-
-
- -
-
- + + {showLoading && } + ) } diff --git a/web/services/appService.ts b/web/services/appService.ts new file mode 100644 index 000000000..9327d55c3 --- /dev/null +++ b/web/services/appService.ts @@ -0,0 +1,24 @@ +import { ExtensionTypeEnum, MonitoringExtension } from '@janhq/core' + +import { toaster } from '@/containers/Toast' + +import { extensionManager } from '@/extension' + +export const appService = { + systemInformations: async () => { + const gpuSetting = await extensionManager + ?.get(ExtensionTypeEnum.SystemMonitoring) + ?.getGpuSetting() + + return { + gpuSetting, + // TODO: Other system information + } + }, + showToast: (title: string, description: string) => { + toaster({ + title, + description: description, + }) + }, +} diff --git a/web/services/coreService.ts b/web/services/coreService.ts index c010c6cec..a483cc452 100644 --- a/web/services/coreService.ts +++ b/web/services/coreService.ts @@ -1,5 +1,7 @@ +import { appService } from './appService' import { EventEmitter } from './eventsService' import { restAPI } from './restService' + export const setupCoreServices = () => { if (typeof window === 'undefined') { console.debug('undefine', window) @@ -10,7 +12,10 @@ export const setupCoreServices = () => { if (!window.core) { window.core = { events: new EventEmitter(), - api: window.electronAPI ?? restAPI, + api: { + ...(window.electronAPI ? window.electronAPI : restAPI), + ...appService, + }, } } } diff --git a/web/types/index.d.ts b/web/types/index.d.ts index 833c3e2bd..ed83e0d14 100644 --- a/web/types/index.d.ts +++ b/web/types/index.d.ts @@ -11,6 +11,7 @@ declare global { declare const isMac: boolean declare const isWindows: boolean declare const isLinux: boolean + declare const PLATFORM: string interface Core { api: APIFunctions events: EventEmitter From 561b1dd94a07aec5289236410f75f6be2e055d14 Mon Sep 17 00:00:00 2001 From: Service Account Date: Thu, 14 Mar 2024 07:27:24 +0000 Subject: [PATCH 26/37] janhq/jan: Update README.md with nightly build artifact URL --- README.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index bc206f4eb..b25a917d7 100644 --- a/README.md +++ b/README.md @@ -76,31 +76,31 @@ Jan is an open-source ChatGPT alternative that runs 100% offline on your compute Experimental (Nightly Build) - + jan.exe - + Intel - + M1/M2 - + jan.deb - + jan.AppImage From 03a0978f5345e394522f21f1638d0d62f89d2796 Mon Sep 17 00:00:00 2001 From: NamH Date: Thu, 14 Mar 2024 15:58:42 +0700 Subject: [PATCH 27/37] fix: some costmetic issues: badges corner, recommended for tensorrt models(#2346) Signed-off-by: hiro Co-authored-by: hiro --- uikit/src/badge/styles.scss | 2 +- web/containers/DropdownListSidebar/index.tsx | 2 +- web/screens/ExploreModels/ExploreModelItemHeader/index.tsx | 7 ++++--- 3 files changed, 6 insertions(+), 5 deletions(-) diff --git a/uikit/src/badge/styles.scss b/uikit/src/badge/styles.scss index b777892d7..4788f65be 100644 --- a/uikit/src/badge/styles.scss +++ b/uikit/src/badge/styles.scss @@ -1,5 +1,5 @@ .badge { - @apply focus:ring-ring border-border inline-flex items-center rounded-full border px-2.5 py-0.5 text-xs font-semibold transition-colors focus:outline-none focus:ring-2 focus:ring-offset-2; + @apply focus:ring-ring border-border inline-flex items-center rounded-md border px-2.5 py-0.5 text-xs font-semibold transition-colors focus:outline-none focus:ring-2 focus:ring-offset-2; &-primary { @apply border-transparent bg-blue-100 text-blue-600; diff --git a/web/containers/DropdownListSidebar/index.tsx b/web/containers/DropdownListSidebar/index.tsx index ef7532063..70651a4d4 100644 --- a/web/containers/DropdownListSidebar/index.tsx +++ b/web/containers/DropdownListSidebar/index.tsx @@ -294,7 +294,7 @@ const DropdownListSidebar = ({ {toGibibytes(x.metadata.size)} - {x.engine == InferenceEngine.nitro && ( + {x.metadata.size && ( )}
diff --git a/web/screens/ExploreModels/ExploreModelItemHeader/index.tsx b/web/screens/ExploreModels/ExploreModelItemHeader/index.tsx index 581a628ba..ef75f57f5 100644 --- a/web/screens/ExploreModels/ExploreModelItemHeader/index.tsx +++ b/web/screens/ExploreModels/ExploreModelItemHeader/index.tsx @@ -180,10 +180,11 @@ type EngineBadgeProps = { const EngineBadge: React.FC = ({ engine }) => { switch (engine) { case 'nitro-tensorrt-llm': + const title = 'TensorRT-LLM' return ( -
- TensorRT-LLM -
+ + {title} + ) default: return null From d9c3852997eb42159701fefa5265f7ba3dd74ea3 Mon Sep 17 00:00:00 2001 From: NamH Date: Thu, 14 Mar 2024 17:16:14 +0700 Subject: [PATCH 28/37] fix: ts error when declar var in case (#2348) Signed-off-by: James Co-authored-by: James --- web/screens/ExploreModels/ExploreModelItemHeader/index.tsx | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/web/screens/ExploreModels/ExploreModelItemHeader/index.tsx b/web/screens/ExploreModels/ExploreModelItemHeader/index.tsx index ef75f57f5..465e69fa6 100644 --- a/web/screens/ExploreModels/ExploreModelItemHeader/index.tsx +++ b/web/screens/ExploreModels/ExploreModelItemHeader/index.tsx @@ -178,9 +178,10 @@ type EngineBadgeProps = { } const EngineBadge: React.FC = ({ engine }) => { + const title = 'TensorRT-LLM' + switch (engine) { case 'nitro-tensorrt-llm': - const title = 'TensorRT-LLM' return ( {title} From 8120ad2a038e0b88333e5044db96b43dd73d4c4e Mon Sep 17 00:00:00 2001 From: Faisal Amir Date: Thu, 14 Mar 2024 18:37:54 +0700 Subject: [PATCH 29/37] fix: badge or progress tensorRtExtensionItem (#2349) --- .../Settings/CoreExtensions/TensorRtExtensionItem.tsx | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/web/screens/Settings/CoreExtensions/TensorRtExtensionItem.tsx b/web/screens/Settings/CoreExtensions/TensorRtExtensionItem.tsx index e4c8ae04f..b363f79be 100644 --- a/web/screens/Settings/CoreExtensions/TensorRtExtensionItem.tsx +++ b/web/screens/Settings/CoreExtensions/TensorRtExtensionItem.tsx @@ -123,6 +123,7 @@ const TensorRtExtensionItem: React.FC = ({ item }) => { {item.description}

+ {(!compatibility || compatibility['platform']?.includes(PLATFORM)) && isGpuSupported ? (
@@ -185,15 +186,14 @@ const InstallStateIndicator: React.FC = ({ onInstallClick, onCancelClick, }) => { - // TODO: NamH support dark mode for this if (installProgress !== -1) { const progress = installProgress * 100 return ( -
+
-
+
{progress.toFixed(0)}% From aab8ee84afd02f6e3da3ead0d9d5c27bcd1fe50f Mon Sep 17 00:00:00 2001 From: Louis Date: Thu, 14 Mar 2024 19:04:31 +0700 Subject: [PATCH 30/37] fix: disable rag & stream settings from tensorrt model.json (#2351) --- extensions/tensorrt-llm-extension/models.json | 4 ++-- extensions/tensorrt-llm-extension/src/index.ts | 7 +++++++ web/screens/Chat/ChatInput/index.tsx | 7 ++----- 3 files changed, 11 insertions(+), 7 deletions(-) diff --git a/extensions/tensorrt-llm-extension/models.json b/extensions/tensorrt-llm-extension/models.json index bc6a78256..31bb11a9e 100644 --- a/extensions/tensorrt-llm-extension/models.json +++ b/extensions/tensorrt-llm-extension/models.json @@ -33,10 +33,10 @@ "description": "LlamaCorn is a refined version of TinyLlama-1.1B, optimized for conversational quality, running on consumer devices through TensorRT-LLM", "format": "TensorRT-LLM", "settings": { - "ctx_len": 2048 + "ctx_len": 2048, + "text_model": false }, "parameters": { - "stream": true, "max_tokens": 4096 }, "metadata": { diff --git a/extensions/tensorrt-llm-extension/src/index.ts b/extensions/tensorrt-llm-extension/src/index.ts index 076951c3f..e3014b447 100644 --- a/extensions/tensorrt-llm-extension/src/index.ts +++ b/extensions/tensorrt-llm-extension/src/index.ts @@ -19,6 +19,7 @@ import { systemInformations, LocalOAIEngine, fs, + MessageRequest, } from '@janhq/core' import models from '../models.json' @@ -144,4 +145,10 @@ export default class TensorRTLLMExtension extends LocalOAIEngine { ) return Promise.resolve() } + + inference(data: MessageRequest): void { + // TensorRT LLM Extension supports streaming only + if (data.model) data.model.parameters.stream = true + super.inference(data) + } } diff --git a/web/screens/Chat/ChatInput/index.tsx b/web/screens/Chat/ChatInput/index.tsx index c90a12cd2..8707e8bcd 100644 --- a/web/screens/Chat/ChatInput/index.tsx +++ b/web/screens/Chat/ChatInput/index.tsx @@ -244,16 +244,13 @@ const ChatInput: React.FC = () => {
  • { if ( - !activeThread?.assistants[0].model.settings - .vision_model || activeThread?.assistants[0].model.settings .text_model !== false ) { From 0415786e2063bbffa2c625be6ea261803b0c5dd7 Mon Sep 17 00:00:00 2001 From: Louis Date: Thu, 14 Mar 2024 19:28:38 +0700 Subject: [PATCH 31/37] fix: app does not recognize GPU first launch (#2350) --- .../monitoring-extension/src/node/index.ts | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) diff --git a/extensions/monitoring-extension/src/node/index.ts b/extensions/monitoring-extension/src/node/index.ts index 1d65704de..25f151112 100644 --- a/extensions/monitoring-extension/src/node/index.ts +++ b/extensions/monitoring-extension/src/node/index.ts @@ -2,17 +2,17 @@ import { GpuSetting, GpuSettingInfo, ResourceInfo } from '@janhq/core' import { getJanDataFolderPath, log } from '@janhq/core/node' import { mem, cpu } from 'node-os-utils' import { exec } from 'child_process' -import { writeFileSync, existsSync, readFileSync } from 'fs' +import { writeFileSync, existsSync, readFileSync, mkdirSync } from 'fs' import path from 'path' +/** + * Path to the settings directory + **/ +export const SETTINGS_DIR = path.join(getJanDataFolderPath(), 'settings') /** * Path to the settings file **/ -export const GPU_INFO_FILE = path.join( - getJanDataFolderPath(), - 'settings', - 'settings.json' -) +export const GPU_INFO_FILE = path.join(SETTINGS_DIR, 'settings.json') /** * Default GPU settings @@ -136,6 +136,11 @@ export const updateNvidiaInfo = async () => { try { JSON.parse(readFileSync(GPU_INFO_FILE, 'utf-8')) } catch (error) { + if (!existsSync(SETTINGS_DIR)) { + mkdirSync(SETTINGS_DIR, { + recursive: true, + }) + } writeFileSync(GPU_INFO_FILE, JSON.stringify(DEFAULT_SETTINGS, null, 2)) } From f878555598563e10ecda4498b08f5cad249a01e5 Mon Sep 17 00:00:00 2001 From: Nicole Zhu Date: Thu, 14 Mar 2024 20:30:37 +0800 Subject: [PATCH 32/37] docs: trt-llm extension guides --- docs/docs/guides/inference/README.mdx | 8 ++ docs/docs/guides/inference/image.png | Bin 0 -> 27275 bytes docs/docs/guides/inference/llama-cpp.md | 11 +++ docs/docs/guides/inference/tensorrt-llm.md | 83 +++++++++++++++++++++ docs/sidebars.js | 13 ++++ 5 files changed, 115 insertions(+) create mode 100644 docs/docs/guides/inference/README.mdx create mode 100644 docs/docs/guides/inference/image.png create mode 100644 docs/docs/guides/inference/llama-cpp.md create mode 100644 docs/docs/guides/inference/tensorrt-llm.md diff --git a/docs/docs/guides/inference/README.mdx b/docs/docs/guides/inference/README.mdx new file mode 100644 index 000000000..289fd8241 --- /dev/null +++ b/docs/docs/guides/inference/README.mdx @@ -0,0 +1,8 @@ +--- +title: Extensions +slug: /guides/inference/ +--- + +import DocCardList from "@theme/DocCardList"; + + diff --git a/docs/docs/guides/inference/image.png b/docs/docs/guides/inference/image.png new file mode 100644 index 0000000000000000000000000000000000000000..5f1f7104eb85703538d73fff97bdeee0019e4d33 GIT binary patch literal 27275 zcmc$`XH-*N*EUL1P^2g#(vc!aFG{Z>y-F_zP(*r_0HF%pqEe+xM`_ZF^eQSXp-2fu z2ti1IP$eNkLV&ZmpZER7c)#Z_(oE+eS=~4X2KYebs$rx-M%I`@OL}?<_iZT<+y z2*t?9#?V7wZ&N6itbF6)TP3dWt@4Njt6N_er%ec8=1}E}c?DfrqNR0ESC1ywM&C(U zhyRQZl)I0x?=wHBS7qk3&CJQPWVSs$c^IhuLR z$+(V-$}T)lf41CB7w?=bl&zxCumrS3-xE>p5K!YYE#i|jemxJCMgh8F8P`#v z7uUMS3m=F?7<(yr%ZgQeFPTPV)`wd*V#@Q%{bMFs&PW@pWznxJfQz=uY$?gNL}ne7 zE>%xo=7mf1axOn3{t|P{8_X63*>(TbSH#2<`zLGdZNx&BoFh3h1TXkG>o;Z5fT(S@<#yA{3`o@&j$DaiCepU<7_@?Nhr5 zGOy^OZ8Gjs#db}XaVT2qF)+l`BlcVQ<2fo7D;;JMeO`XTdTFKaTSUv>}2FkP$~+NTYp7&qiU`)J9JLDpjZ4<J-kujSc z?T;-Lcg%Z*cYY68>Y_`zB|ptj72;;ew5LxtC`hl~-n+#zbd1yQD>(*l-AA+x1*I2& zXKe=>hq&xTL7mOVgo&uaIG}$Yp>yITLGn_qG=BH{(=R4+6%LOK+7=^E4m;f@zhJ3fkvF^6i60ZQZ7Q7y&w<{c z9Ytz1QJ2K(;9x4;N0;1vXT@b|)#?*Yv7$-hJLR6z*n`c!JQJ8}xK5;uKm?5H-!OVDy=75MM}YY zKbwKTv2O$%dP|M;n_>EX+fsmm-4QUQ-PT3%p1m&k+#xbLIH)Shbj*1o2LxOYcTgFd-4YU$a2( zDF$ywdev^C`N!A7t+9 z85A1Z{c~lQclcQ@OuWm>ZJC=etmEN(3)W~jn-JM`_+22R+_H)76^>KRKuT>rf^;xc zasNiK_Hto3b_9d(*^ltem6$$&h6W}$$!FQVVPfocW|x%Yn!eF~BV>a-7=`XTUF^#& z(oHLUFxGy#g2#=I=Th{|Q=zEC67I62$%kxW5!P*+cyFt9fcmqTQ1fw>}5vs46D+s80NrjnEJ@NAYsX?6y z*tbFzUwLTzKn!@K$E5c9G4)ZW+c%G|zrU`F7Axt#k@)kOeuI_)+jLtjVdg1Uht)LP zw%P5mLTjhny(DmQZ}7XI?d8MpyFmCqS>wtdq?4s8 zKO93%>UOl3sM(p+cWzZ3DzeHh`F#)Y0z(q`Y8$D{sA_7IPW1_e&5X@yw|CcPADf&D*eLs(BZFzW5 z+INY0(vIsB=!lqpE1K|*`x7O(Hc_*$!5dAgeM-1dndv~PPO5K$1%YO{aRiw^>{B<4 z0awmYK+~x|$W1q>9()#ge`~e~d7v`gs2oFTTscBr!MqsOas&|tWW)hKc`@Ap{A5aX z51!TNW_#&WGVI;gB!f-T1L3o4LgXAN#!&9Yn>0T@WDTiEiACQ*Ft#^-CbHD2i*QH# zmeNgfXHM7#B?nF^j)v!`MaA7Olr(*V|2=84gdPVSDGSH*B%`OIG6|RV+FezyWaX$) zlKU^~k9DR-X2JN{KE&Z|T?@d_dza5=L~(aaQDGFDxCymBo9+;Q1+8FXwbR4ZcY=kw zm=_@2>|(oAe4rK1Fpa+!GPpaDt-s4-S~ieT9MY`g32e`mz7Oq;%dFh)td|S<{sU~k zlJfA<6>Sj;^G?y99$QJKY5Z3lX6H&*{GSvT9oU@d1Cf(!cg+~c5@v2Y%mR)xbH1H_ zk^V^iSL6TK#N+xeatF5W|8DR}Pyfvb1|{WzPX7P6th1G{p2q@gjn5kweY!G>+^U@N zn!FbVryfbmXU;Hf6a+h5WE-WwiQDxOYM|E>t6}Bv<1Nd`l>6K~Sv!#=9I^MPM8XB8zL{R{-ZL|5br6J_Ni5ZexNUh|&y zJ@FC|7=wW%xn&zsVDr!SV9XYD%SjuICre8&bl609myQJd>8Gwx3;|NDdv)DdDApX` zW$)8bHVT!uP;gf$^09K|=Fgd~iAlmBle`df%LpV^9TrYn_eJ^G-EQts8(DgX$rD#M~GmFAuFBpzWvCf zuJ2rV-~6VHVZV=Pry&?w6zB|5lEy;6bMeaFKsuF|MutzmaDV8G# zs*$jt$&I6)qAtZC#JYcdVBTO5)Jh56jEE?!g751Z^!k*>p5}$urSHe0G(;XJYv+mU zgxUrb){}xn(RbIsOPZz4*7)*JPkri&C+idTCKowlfS3GL<%!J_b8rg-Pu~J$qriMcUF+u+;tmQyyU90w zRmrJ)Oxtr{%+1m!WMfm-$L=(h45XP8~mcx7%W))^rVpZKoaA|?26^4f|HEeImr8Mj) zJC3+wGr*knyU-@ci@PVcV6apAed z+=c&7;In=Giyqkfpgtn-Ywg4}`iW2G1 zxQ`4GpE=ljnPQQCa={M${;s0iHoNw)WVynAO_9K|VFuLBA4_TuKXt&mhZb!*XNC8* zRsH;6ZA_3uT+z8Q9fZqtB!G--jMX(iHEUw*C3rD6e3X)U%36+d-&qN@EtTf8{@}Vx z@f>8Sva$T&1bH5*VCC3q8thaRim!GAB{L65L=j*>&Qs&aamYspmh+sgb*=XX(js@; zx5~9yYO{D?!H_`W=>?TVyL`IVxBvXqWGG`I1~L9vcX>_lml1V&P{>8rtA?L3hXvPy z?YOC0K`o*)Gi_(nkv+(cm!*x8>_h}O~P^3(I(r`V^tC26AP2Xg>Qn2-VhJ=;o zP{|Oqh@)*XorBr12%T`cCWA^duzIPO6mwN`pjTwsYA?k>^Pum-Q@o8t48)^~N{~A_Kb%c=OTsBJ+d{u`ynng51&}E+#H*=@Kt#WdQC8|#&y{K$%1N) z=MMMDe&4w%=zHsEiU2idps9kajgD66%Zp_A=f*Y2+Bow36vUmx1@(_fz+O>?9KC$60w0 z%mMjEP3h?VClV7gi(sH!%5cVuS*Hfs={|g*hCR9{w1LpRZ}xkxV1YY|bO5XolZCAM zXdQG0Up-F0@fTFq_HNU)L+KQy$?E&-i}iM`*giidmj*jy))f4Tx>YvA?lToz$5|z7 zBg$vB>&!Cy!BoU8o(QVA$K#8L%Z}&7xi?}Jb)%?gaSb2Y7h&oO^&&b|&O`RgmL=w zLTle@ya=>$Z>D9ml@F=Cb5v_ATLf9p3GnjjZwS3`JiB-EV+C0uaUBbW%lYq;OYJ%W zo4zz?qFfK)27-iSkhj>VNcJRs7W`^!cE4s<`?NGFlPTLn-4xp2P=^-wZGnpP3YV|d z84#N$-xWl6iND;g`C_bKCDA9ESt4OJZK%z@%HXqbGEugC*>Z!#Zk6*x(9mUQl)RUz`Ck|2Q$QNl)SFqW?MdVlq3f z%0zmkYYohc)XsQA5 zcRn4Qva>O=mxz56gw_aO4qzq^p(-FG-SzKy;M-C~Yh3x~gIWQmV>|8Y>)kxxwIHxP zeXdW9OK~_rS`;#F$+Y1xyGl^a9Rzx>Rz|*+bdS7ym|LM^0&g2OG%`hgG;f8G_F0{eN_#&-FO@!IKVm-&*4SSVPzpPAxH4ZH6a)(s=lG79cwkvgoUj-ci7F-xB5q3iWz(xu2vOh z?6(||(Bv3AwP?JMEE32!Vxm5how}9jkZFV;%(!A)S+nr8*ZQRkf}MyJ8wAB1}C?kA|dLu;x3lH_nj(u5d``IM<0zA zisDRBh}`-5GP_kclkg+%`qj)i5o#*n=n5cPv4&X&!TYTQeTN zLL4J3tR?;!&Ie^#BI*Y?d?&y_#YM@+|3650~ zuZHME4Jz}jy|VW%bI)X%JCCh}a|?W}Ru#(2vwo&h-KQlnJ^W4>EjFYljJO%Fbmxi6z2ZK5 zJZswS=PXXBG-ObOda9FHDqduYxb^l`S1!YCvA&mlySOLMY72*xTM_Mx3WpIKCoaK;geI=-K|L!_gsvpi$SWFYx(q2h(FBRS1a(wpMJppQcNw>~V&(-Oi{&}1 z3|7mPdli{q6&Cwk!dXn{{bi={`eF1K1OLr+hnseh6WswjANYRD9n6_hFkj!cK{VkwjP#^>h{7esZ`n2WMy z(Si<_Gzz5l3wSa^EiY?JFF=-H_h(zq;GlvPG6zfV>)TcwuSS$+i*|EMcG?tk2&4@) zHBs*O8n^fZ~-s;I^I$0*5u_VNVrRJ9TR@Z+|HCi)V7NaF%U$ic?q{`(_F=K4fMQk^?=@_e&OeONbou zo%iulqcDd?D9JtYzx1W9jt7aWj+wax4;A>U7so-+CaWKjD~mA~<_ofmF{9txbtwFV z<@Z}_gRdk(n)oI^$;^JeZxK5gB;M2DH5(1CwCb#p**0dgbEG@; z@n!9)f0XxJ*w$XLZqX~kof5l!;pbn>O9i$v!~1wa9HQ7OMJz9r(@on!A6%zzB>wT7Z3HFS89A@8(OdRS= zY#sI4e&u6n(zGxLAAHLZ{StgyB5Jm`*l{PGVD1P3T0e*lgwcXTB%2Aw_LPWANs z7J$H6UNaVCENI3l#bVf<|GQ;~mHb|c<*L7MAX!QmtNR1=iz@&pR{P9Pas1fx_Q7VB zx)T%>*P06eZ*;CeqXWLm8c2mS+?9o$ONYF4f!|h!XJq_~y=DDNr~OYPPAUT6xWr~{ z!wT+?=BNnL0l?qX^9J+<4flUW>B?C|-W2vr=y^=9=lMEtKv1azK3%(3k-U0am+Xz< zSpd`33ou61Q+~^kjm#qb(>GXlmz)3u9mVwj2{+SC8-xOwpyDIRPJoum*pFZJ_6@UF zC^vqeV(l0K_McEWx>g{$1gyRT!F?wn`_i%pANt+1-4q&IN)_mp|9lO6XPNsp0yn9{ zEjDyn44~;a%$4RHuRKhVUw!+mKjzYaohJaHnBUCvYSR0c>tO)}rh;pl;2T+PrCQO^&&wFF4?P4k$O=zOQj1` z`LOf(NZP978QL)(sAKP=1?n$w4Wc;5%TX!xhtdbH0gTlB;q}p%Eo(sD`to3{S3DN2+pOm87A-vAjgF8=WM&u?0qVm z#j#V$YjyVHvzWePlk-CpO;>Dmm#z+OC6=Ll2*we%(})(-Vym}jAr1}*=!!c}+L>Br zuL~1JSF&KHll2mHu;V! zO(Lp7zOye&?nuj^OArJZ260!P7PyrrA$by;kQY5K>!mrxLNHu-L9l+}ujO6HgDvZf z@aZM2U!PO(QU3+RY}^$4$?63#atN}N>vJzdM7-SZIWk^@5Z+CU?5CTkLTn-!Ph>PK z3c@@1G9R>R_0lQbqXutn)dgYe#1UR^qDVtgh9{8G1nP;ePxrq;23efsKR%+SkA zzKH;XsVP=*`d8=-Cef#SNfT0}=>VBP^9uj53NgIWcaaRf5jK-lo$Y!Q^(U4)!#VaZ z5NZKl%)&@zw>!5#r*r>|1Q;^lVXg}9TF3BkAMZK>z*||Y$aHIL_-{F>`k;OFyDAsd z1;OW!Is~XD>6Ccj-V*N_Pzuz1K5ncgYpJf59PN6v?$9==XJ`MCoc-O*P%fAEWIPg- zZ%M0RSd}5gFMFjaNHoFF?%uvvtcU6s@%->dy%uc~x{z-GjJ+^1)wzyw;TlbVq-}Kg zM9LJ}aAUvNale0)|D?%CNvlxXw%X^r5H(5z%%1!YB!JQhgvi|BJh4-Hres=(g)|z? z%Z6LGja7#}DDoR*bp_^0UKd^2ypDmvh!!yx1lnujkx zeX2;C5V0tC9zJ<^N7a?*w)3#k;mfa|6<1zSLK`(i>+K*L`?fEZiBR&StIBE(>1Q@6 z$`bN6lDlry)I>0v5JQ@>6XB2qIKG{X%MD%>@NUbLSOYH!zib&o#9ERZJz8+F(UR7Z z8N%T{rnT*K%|z`qkT7!U&9rG+Q+Nw!8MS<11 z5aRc~LPD~hL>0fQ(~wVXO8EOO{Aq7RVW>y0>X2(=-Kf%3-aaJ&b=xY)e$GQC@=pzS z2OKD9R%dkqSecr)=d2)BL4wvPFX+aj81L&YNrcqw($wxG!a2YZc$|2Ue{gjudheJT z4EeP%r%k(w4V=ql`U679PT)OMSHfcUg%U<3eDc$pU$A!8OsG9RcyQXJDsO7;ul4g5 z^@8&8X9tgJhtN9DI|3m|>hb!r;+Nt`X}8+HH^&>SH7CBhDl48uE$oEN-T&O)&sn{6 zn-L4~Q(qfjF&#_T@=ZPG0=rj^Q*&$%GtsZ3+&O0KL9-J$S`PJUjTdK&o%^=SwQb9a z9@dG9gfe(|PgQ89YsnooSw!G;V1Nakausf49OdyUC))&i3{VA1HP^lv4cE_WAF$S% zKh6@e21m8LMFRp7@OQul6QpFO1k%Y?Wafr{Z*R>uSJ+PKASF5C=^IX2(V{MbQP-~p zxyJ>`fkjbggzr$7mM3$W*{R~&8%UroaQGhMj6M9pWU_PI6zP>P#2uyFvS}5?j=xG{ z1&LR$%k{CCwnQv{HlD5~psJ?lGPU$PlY4va2s0RJXn|BYfCs>b*<~GueRqE)QZN+b?Tl; z`g|%S0(rhuUmZTZH-+!1QsRb&n=-B@KAnc5uJcyomh=VY7U0Er>Q71SO5CwJk!mu+ zdQ(vp&=_hMSb=~w3jS3FsYBydy%>#t0!T#AU7s%W6R^fl^ZrAeu*Oi7#)B~g=|y*H zuSbW#u`~NWTW1PslhA7reKwuNzN!y}OA)cW^LS(j!;nH1!>g7Q9YMH z_*o`IVQy^Ko`C#nVfPa6;-{Zr1sZr@=aNM&v~=67g3ZE$@@~bObfX$sL!k4!gjtw!bB5|NdN(tgqo+p@FnX zv@T4td|(QrwirlSHGh;J( zaQ1rwwGg7TJZI%9l%Z51rjjH3R>z^097@|0geI+J%eH9| zM*IHT^=5E zX>vqd(?GY7eMyUtzL^?kQMJOEIha317A3}0z{8W^BxH1F)7;cGSR-V2A zG4NcDYGGe8Vh8MESwKZX>2W>zDSY2jtOkhFS8p$V zK&F{I&T4lrAHfEwP5A*VUUuJY<;vfzJxfai!Eh0ljEKFNaWF0~@)6R>C+?Gk9J;tO zr4bo86$8D(LGGS^duodve;ekssiS;1CE|5x*&@+uYNlzBs)9hy{6rhIBK7O1&v}m` zfzG5kE`TMV_!1T5e8vrw!2VYL*y7~tx>$Y*!tDoovOC48SJhbyIY~+r$|fwj@AAG{ zi=*$9V;n}~R-U5Qo_@k6I*Bk!zSoXv2!fimgj-2;r0_EouTfQS!e;|ojU4{k)4`>$ zsIDm$`Y@8)@~CYaJ9H9^`tGD)ru7w8P&J{KJ|g z8IIRzo|ZB-JB+?j?c5P*mQ5+^&Abn;k?MEnqKe7W(4MM7E&^=7V{(#s-*XdBx25)n z>RHs6A?Fqp|Ij2K8>m}Cg30YnH%sR9TLc1XY8w=0iz@g~wcmG@VUj}{?{ox`w3Nhp znNnS;7$!fx-U^pA6R~XxzmeaXmkojO3CizY21BY<+P9`G`BhotyETS zsQXd-S4aB7XWyk#M3Ywj1!Q8Wqv)lxsZAi;cWQktIhL?IajK!<{#!-p;L4Y_jdcT_ zqBP$k4qS_wyVy9skhYTKOT`#2+oNf>ztYKp7qcLK zM1EY%+o2$L1Uo+ChS~2e_QO*=xDE<+Qqjle9ek|Y;lm~EPv6gVDyKKUz`)-8VQRKy zID6WRksr45G^?eHu;qm}Hst_3-*<|4Om*#{SN6ExafD%dI41JpAl445wVA zS$a>R2#y4XkAA=#fzm>$xjZ=kPLbh??dkNn-fV>vK(;o2c%0R*I3+zOom1GvzMZ?T zgCLcIcK7li4wg8+Y~O@V3{CmNGsZJz6Fqj9!T>Vymm2`Vz#!FZX2gyw>prkUq&&nk zP@*u&$~qvzRj^C5HSS6YoEyk}J(uVnQ*rpO&~b>$)1aA!nY}>#Ei9 zld7X_M2<%k#aoL0>YlHW(~5)L6x1}cv3KZiIrSD?wk`NnZ_|55mB8`!s|I*HU%{P& zvx6pGqH92~^rk)xR!3jFC(m`y+98=4? z!?BX2?$_@qqwsMbbRD^_UH4kAeUh=^(OYcUi>hrV@V&B;qqsH&^Q;lUodV@i(ct8m z?=Lu}SHWSKINJv7Qj0uLa!oH<=oR$Gg0D9XbzZT2@I`T+T*qyr;)IOtc!W!SI9m{ZW(skci-lS!%xnfK#Zsp;&z?ZD$3HHDN~XZCi#K50n~6@2cn zI<1R5SwsS3Xoo2?_0NVOPnMhOsOTW3KSERKr*!RxWF+aREFqD8Kl$fT#rh6Sg>X_b zzV!7QWH4Wh@jBBpC99<1!b$0&PQQj^$WZ6|!Jru5ECa%`&~?oYXyfu@)#`2CA|tz1 zn#&)=hr)%vP`SecRP1^ixM~;(Y1HOa9JN_PVOc}rIZ>PIas`{uksJ~kl-L}n#@wPB zTEa*Hv7j(9HTbdb>v`4D`gHbB>JFA0+1YbJM}qf4|7<-3#4jhGk1g6o!x*v494Z-#x&v zNWuAq)XVjn)iK|Nl6jvuoC)4-$d4odZvk#%y27ZakeGMofADf3jm1GX_CnUoS}x&u zX4V*hj$b_sJMTG0_@0*>pYBK6M*>3{bnVqGH?3Er+9 zI6Xqxi$-C}ZkAr$TKdypak%x`=l?|A620y9AR_$tw?jaL&2LnC*ao;Mt@Q4HX|HA+ zX3y7(&N3V>8s9yK{U20Xy8Z53f&X4CZT}y+*Z)V#uJY3jKGjZaS+;{{VI;Bg*$8cv zxPwD)_c3t87Lj5X)5*squ&7zrJZWsQ2`&I9VToqxU*V}{S!mNmq2a9y!Em(oPb^ZNX;$k5PG0xZvvR$zPnY|9gl^nGW3TFS2mXS5hRX; z0LJMQTYJ#}Dfex{ln|J4(*Ke`jExE;kZsonHmqw>=@E_U|CehmJ@{ zSpQmS`*n{!JJ0{+RKS;JPMyyNpw9p`{MF?PSEoNEMunaz#RE#D`eOhqvk{Hl>v%>? zWc}~BHIABo!yQ_D>?nu)P#6s5xtM?4ACz_4ftvZ!T1l7s@kpl@LHy&vH(zKevl4IZ!7fpe3R@KBz@S`s2pr zDt_mB*4^vlWKVHr)BDfp%K*KK$%HnK11Mjs|~D$hk{T^CFHeBzV0CNxm7#LFdiKDW7LOE_zq-s~%V&Sv zD&{!IMVMp7!V7Ex<)@;bCAD3X0A9ivwa_OB{k!N}l+{uGFM+lFh$~i}9J0oa1jKSg zleO6CA+9h{Q0qz>P4=Ja)};X0D&RQ7HL#L1Rm~)Y2(TusZzZsQ+5_tN!_bQ}6?5v< zLawMIWK@)*NG+}L+84Lz0XMmR$#EB<<(Jdy)dqb27=9no6}cq^g&Vyostr<6c6_+6 zdB&sR*1^yR#HUxH%2}^|&a!9f-WRu~?D^%l5$E zcYy=STeB=+ObJI~tK*MW7Z9%LYv($Q+g&N>i<(vAPB^3++nehO#ym%CXAQ}&h0W&e zugL*`&5dm?9Pfl3S_)!`D6cIa8#tpe-da0?5->=$SILS-H+qfAg#R3-o;Nb}hsbYN0QC zLH!|1WZ#=0MNvf?VlE&2r;0oUBX&2RtA=x)rSDkl0EoD| zN!D`W{rAM`o4>ZMm2jWBZ3(My1Q2beaz1c&+qEyex{yfbo&(XGT`r)$K0b1x)XMje zze<7h1x^f9?%(Z$9RGfkz!|*M*bb;h;;!0LO&v0iG#bYEciyw0Wuzv{GEvN7;QV82 ze8>!H^|IBUd!(@6jhpfm<=&|Qc4L32$!zAth|(0WDP_~|j-nz!+^fSWTY5ARos6)j z_Nfy9RdDtbtEUDlo4U=gGo!TM{n8&0XqI`yWLN2bk(?e9=f?rqeA<)<^gBKyn7f;4 zEnDMSXd;E!>)LlPmg+p}-cZ5(;{%+qa_7qbks6$_7$hu9KkHfudC`cUwP(mYuwhW0 zn7+I|67-z2Be(zX8tM|x?R7*}_cJvW(x2~d{=(;nj(34crmhd5-LS7oY%&vBQY-UsvcOJedyXLx;x_ z0|Cxq7Dw4t?7rPnrjBTg;E`+ZtxsoBtO z`V%rp?%eK0u-Rg84+sw!yE<3`BX zp;;+$u7XdM*4?QH8pLYCH$7VF-1)tVqR7x+zpLazP05*VC$sqI)1;OMcdv13Zfh*R zT-9*O^nD|5(}Z{y)>vpUN{Iek09DW#z;#SS_}Fj-ynmsl@cSR8xUsJIriWzekx^Ub zMQm2dG%@qsm6EvThKPG_kigox*gFyj;RIA^?kDKzdq6ah0O=#=Uj=kpEZsKuV)yiE zBmMS6LN_YSAih2vErPCyg``%fWoaoSpksHxlfP@qtG9P5g0@k^_=cUs7$2cd3`$d# zlfmy4Q5W2gDlTWDbb=9l^TfFDZ`B8eTK>rj>ijM4PEk%4o&>x{N+a)W6B_9YN`3he<8K1Y#E^{=9Dvm7yJ|fU9p`re zI?Rq!t4BBdpb?7kVuY4;O*Aj|v4w*_qjN7jmOS&a(sO2J-V1{3DUSMecA#nWfHci#+ZTlnz&@~( z_%*ZnjhjO0C00Ns9uzueANEu44u5?mW0hvt7R4e>`xZ7a_L3Oy+js#JJ|!MP|KG}W zTEZYJgePh{ret_Ji2h;0v#xbsqx^U4HuYP7Q@{~po(|~_{h>{C0i-(OVydZz6)qQA zfdBN3T`zy-JrW-9UD^Wg!g_k51i&po1FylW@TVTg9|OYTuYjT#+ecLfRktCCx(c4+7mLb1$V?VK=S zZJ=2{HMbkS1r6TWhn&t|h2?H`o|mQ@+leJL?p9f=bxCCbPpF0}q(3~5?ZYSE7<2l# zw+4j#q&uqA%=f92KL^qTVQ;krJ!;6a2hJUZCOJvW-Vt7_(&3J1=CBu_z*xPGmU;3+ zdUfR{jMr#qlw)On68kkNOY^G7O{bt2Q)9NW?i)L`cGpwlz=#XJmi?C{B8Rsf<)ug* zLs3TGUKIBugWBsVH2CqP#6HXSH(&T;PlTa#B{1a=-gK8*D$pCli+fnMT1CXT;bcry z@>F~l^fl}l9aXXDTZ>9WgM&O{c&3ZrydK(Xl~57nAcr9p6C7Q!(CHqbJ-CQyYf%3JK2_w>=1u|bCW|DeLf*1M&vPlY-BS~aS(j= zhK86|bzyhZ+r*^1#A6PBoyaFwf3Fm_ea7oI8zu|V8{gd1^=&rwx4IpzaepIyyqoNc zv)@<2uWQ9xXhW@0dZh&$lbE-icNn~Ef`hu{=gle~9bH?$URyKiG5S4@vhU+QA7))$ zPfJ%)!*lL=$v2dlcti$`m5rRZmgRxyBP^9g4Jj_Zbviv!(Z309&}gCzr43xf3vv6vk>5jA@p9Cx!lBU-VYzI zncc#~f6J@_IFaslbe8=TdVa*JZCUCE(frL6{vY+Xv%ymuWYgYG_3LhC2uBvTBeOPq z2-PTjcDWprw~~>_FZ9kS$>}wKuEp}He~faY^4WEm^5nIP;p=Hm*0ysixZ_v$EGg@| zM*=f-c+4SienbNQ(K$3|-uR^Akd0cJsNTZ%%_SR@<=|c!e?LH-ikg)o;JrjfENLfG zHJWFaR7jYjkmu5{6?FUfG|!*0B1(8M2J8uMng}-3QPUT;p5lpbiPbrK-^TY2=2I&U(N>6$c-kEC ziQYYarT$kHH_mh83KOt`?ee_(O{TMua4J)0*KBvookH+w&!DXHVvPB(0Rb*G?(1~9Cl$Bpv_x8%hr=B_}&Okpi-N-yl}t@l||OWS}SL|YYoiQ4w6k)^Wg zKa>tFk~sFD=P@ba74fgqyJRZmaRzf#RiwCET*5J1# zDovezFm+pAsL6YTW!>7k-@E;H9D1H-oxApDx;G%}6dJnBd7+J%7w(ylj5?i*>anvu z0Zuz-LCwxKKPXOeV>G#mKRvb|I^4M5L?<+{l6gdxTXJc!qa;OD(Xyd=YlGZlG{xs0 z7zb0^ei7y6(i~oa7n@?GE>?HI2`Pqe+avYsdu}f7Tu-V`e9IKyT+ZLp(6rrsud>JT!Rl4P>JtiM#vS#^{BEmO$U=S!i%VXQB{c+?YD1B z4+-enk!@c7I!5VP+hXO>$UkxS3)6iJ-|itAsDTl2cD?QkyvWvH9@Q2(uK_bRb}pl6 z#ZpTvLv59@Md8KQOmE4%rE;gU7N=cqr6s7_VT$x#fqEXtU5#~}Pi^&3z?2#?ssEh& zkhl3k2PFfM`z7EHWkYBPIoHtErf2vIpSh7IEA3M5vu<)5d^K7~z)UD^*ZkU}&G7%r zJ)7=wXEsScW5^JVOK372Fg(|v{gMY{QT=!2@D257mvrL##SxXz6kS*4SwTmMaslMv zTIq-I-_J7w)VW7Gn;98_GW2avOgbJ9em%Nt4|bo`<9 z!rjeDF>9#x8GB%W+2jau`avATy5eaL*yaA5k%}toXQhQ; z`FEvUrQxmSvKPLCR*ZZ;fn?k~S=((xs@8Bhv7_+nZRiM5dV%H+J8pl>c01duBKAs>&=bqtuANt$7yiqnoor);n5K_2; zW$>n)zMX=pRhwK2>`G~h`g?R!)Fg!_t#P&B?e~HxW@_{Y*SG9HX?Ty5DOI*jYxjRm z?h(!BOv?%4MR8zv9Xb5*BWKR}^Mwiy>wqn4|dMLeO;iTz2MbAMNsC`QdNw^TUb@IJ5M(+bRF>4(fV> zha~6O{>Om%5&(cPaq&m~{Zb@$AwpY4rJIPWJ?_h`kp>3ula}``2D05& zo6bP@1E8F^IQ-6E%X-b2eY17ZM`W3_F2{0l;EOPF{84Kkz|FQ{St^~Xz_D&Rv=V?a z{$JIdXH=70*RCx{Q7KB3jv^{jl_p445$P&Y1R)3;F`|GVLNJ83#VrxA06|(%O11(< zdMB_ow1AXI5rjk_U_bJ{+vJOJLkXVDOp)-&U@bXh4`VY4?i9P zC%-pmbDdVtDFdggBy+l3l>(mPUGu*(w0-(5^y8-M9TN8Ugs#aI4>u#0}S5UUfzkE0=o2 zD_m*D@qFceHw)A?rY3B)eGY{FPB$jg*Rs}6dv*tow>7+PKc0s1L3F8B*=*{;lFf3K zjOkDBGYH~^Xx59t%iT(?%mrmDH&K>Fkah3UQgT#-Yt+^<8ps1Cf}}Kcn;54eMJl9W z020R|F~~EQkt6dgJqJPK*ZX3;`h6lX5b1LtysiRx?SZGXd*$7Fv}^N^dLE5CZu=Ae z0T#H1-vlm4Cot`bVEQ5m*EF43qsE~@wnxd13#7PIN^x+v?vNOy9zwza2D`r+=y>A5 zefa%ya~M(FMGUkzj@EokAVi!6a^LXVH3o)YR3Y?2^FCReyvS*l3A9}>_qxzqx6~?w zTR1cg6az+lbdAT965*-Q&DY>kKKI9gBhs<>oU{ zzEo)d5(OzBJwgUk&sowF#>~za3J(NpfFjW8z)S9Cv@e_xeenc2e)3E*>odFJp9c>L zYW~xi(zU0pY$k4tC@FO#HM36#-!> ze6xc75ujZw>Sc!kCxl>$pDN^2%&^qeI^v~NQwN2KAH|OD(`?WxFm@2eE^YAl*U%Fd zP}~g)ZWXLqhyD%zH2x9Suvq_CSd))e7Bv`refPdGu3MN)uagozyW%Qlz=4&Ro4V;O zU8oGZv~M(NRWaHZximqf8&B>>%B$yPiA;q%XxYe5k&r|6e>$ah%JC}Y05eWh!p2xJUE^kl(km<&ak51 z5sQykhK^W$Oszh5+FsSw#MpdqjA+TV3j1mPBm7;47D3mthx|ZCFBbl_5?S`lImY&P z8A-ziKR{*gR#-EOP+r?Vt^t6t*{cnL9Vwas}%Jjpinuw=ais& zInp88&FNk$e&SJVAR}&vPFsfXHs7>woMuc|YR83m)>WZnLtwTDkORk#N!d z2h=Q0FI7AWI~(@TB#5(1&+eAmjTt*Yes+vuEV*6*p@&>V%*s#$v{NEq4+k#arGXTU zUFTf7o?E5>^?E?lI8Osl)8QtPyKYEUyCQ7ivS4$_gVo%IgX#Ty-9tf=92Z(>!|U7N zG_=#|Wnmpyr2?ll8d?2fVh6-P2brcW#PTU!8S%&XOyV`n=EU@8M4B&9&ozm3(Y7v_ z2c4(1XNQf~x|aQ6)u~yh@<560Feu0q(|C~tA1nKJALPD4)I`1ToScdqycK)Z&8Y)@ zJ0_7+7)FTem$807b!N ztdUz)#DPV-l)AJk`#uI{HIw$zWP#~yvdbk_hJCN!?-bj%iX56NWUfxY%hYAhT&n3S z3Alc$BII?Zgl+8i+k2D`C7Kzs(_U)!QFIm#p7I2mMA2qp#mQaQ*jKkZ5s{e-f6_r1iy5cVjm%CT{L}WuIWp>VhZ1E}YW~&|ciuRcOg@M(9_@h>i<`kB}7&w$jrRF?x z182+55?iolj}`IK4Sj@Evv-dFD#@u!mH0^YJyHXH@P|qP?LYdB@g*9f3xL`Kf}TD} z28FWzJ=IOiB`j-L$DN393VKJ`NyfI78gG!Mf7Q%|m{)672e;^UzagN8_){_l$>CSI zpgQ}+g!N{i|B}?3fsFNmv0Qy=w+sYQjFvE;hI_K~u-|0}<^)c|e9c%#cG|(eg5!kw z{B~K&&yl;wSO#ugZ$lGEI38iOI2(eI{okV|CHEcoF^^E+5)MboyN~@r7Kl9(tK_yA zUNZitXp6YV;oDnEgiA`14s)LcO^hVdmJ??%2fOa8WH?*)iy}U$F2J|ubQw?XbygJ; ze5Vf2^(0=0?4Aa++}pINM};K!vr6GpA0pkZgs@I2oBdJW=F2gw6Hv6$? zJdD!`?r4(EW&9%FF>&ODfAf`{;_JyyTm>u6B9c6lqL|>UD)fa#QZoBF9l$;r`?%)7 zW5t&?f3qGPU7Ig-nA;Rmh9&C}=c|*L#391csL8Fhfwk(p+ z-z7JgY#P+eo(?NxUOLW*-=V^YAH3)$?_fyJ4}L#>MA0X5n3spDrp@6uTYinr*S5dm zJr)coD7l+@VGzQo2d;~+|CDpOr9lC;_G6&GyWGLDPL&{cqs*^S;4<8aZ|5nJkHI=a z$aHg&Wqjli7o_>0gZXhrLSc7~ZV0WDVaIcfl-8K;Ibh;e&wrBt&fRZd9yx0dRuibr z`oH@9cl3{j!_eG1@#cRwf&L%kg#6!wnb?AmJ17#ZUUmPsDUExB#lQk&Urm7^V|PKr zRjE?NMwg&H#y2sv`djt$X)p=GR`Lc!qT8-E-+6#tOrEX(M3PT|@rE(>d1LyqsWl#Z zh7m~*&0od9R<iSn`4=i;@7w=n`bysQ)=&$`Fkd&oW^S zFOVEE;CbDi*jO@J(}<9+>z@V@U9LW1pFA_RqUq^~>bDK!wkd40Us}BqlXILuWI5Yr z>P1p3PdFF58cZ}la2x{}W;f*8Cj||?(YxggQBci3EpO!m5V#Uqep{FQ4MKmkvfxTW zw)&|DQUE)zc6*F`<33TG2#^rJ88Xm+=)+)^@lDh5-e0m1aHLwA`lAoS+_6WgAUXmv zVIBs~fkh?bmjUyj&u=qdR_4LD)dGEk#CGurbt(lySQ_|u5Y;8gcx9N84VL5D9Er+; z;R!{6?-Ew%+~Ymc%q7=OnP`dZ6WUMT>eX7zp)-1J?QcSPKQTt%qp;q4bB-~|$g?lN zjr*0H_+rdyZW6X9tTrx)Lsj1ThZ7LrWI`LeZo3|EMuWQm%2uL281`;IaOoWWr;4`! zAS>qP#dSliuB4kIi#G?GE6UN{4^8$X^&Sn*_9zz69^}6filKMal1O1&^AO<+>LUv7 zdqRG8ot(2XSs@9~?7QaYTXX^lhY&VFV&|PcK(b_Oi~RK^N4qgyJIw3Qq2!wZdO7h1 z+WlH*c@1G(zb@(h_C<4NUeu5|V3Aqx-a6Nk_eO)T$(Vrr1z2Y1hI8^?-t2U~ z^X;ZR3)6l^Sd+AUt&<^xf?YG90h3`j1Q@5s%ak+d-tsfpC{*16@HloJkz>Jz!`6o3 zwW!u-Z`USYDif8^yu12lxUQS`zwINURXt@d{YZFI&n!Cc^qaU|eK>Ayy$=C4md1b`Wwvbj*KA)!2ZB8*_@|v?iuLSfeGoc?mR;U@CM=`iFotY znb-;8$e-tF^-eWARc|1dh$=x}h)68>5i(x2AmjjY4%#N{u0$fwe|UHbfr zI8mUvzbdIpx@@rj>r930`~R#4?OS?tSFZ8|@f3i|dqcfP{@Dd#Fi&J%rzUMYba%kl zPG{%{$0OyRwax4Y&S#g-!1z4z6xvu6cMIb8y~yGH2zI+viqONI?e5ml1eg`s6^3tI zISFyCZpJJD2gWa~YwA(>-q}$q>zgUJHk{glOGqV}JwGE7-FjK_xsXBxIXHu_knE?k zV1R=G>G0}hsc6mEH>{3mh>=CE$dXW?gfo>`)-Dt4%ZpB)-TIQr+at2 zRHnIVPnBf*q~g>A?}J0u@jSOGD{beDSFV9p#eFH$*#5NuwG8QWy(`OBJAF!j#LJ}l z5I@$hfd%$eOIKqK7abDhD*jA}Hy@EJeV*t2O6E7?Cbyk??PdmMD`Le1B$FayV zS%7A8v{0zd5)!s{5kx2&u9>*mDxI*yI=CC3iTwHI33+kxQ`99;Ni1HHmPbtWO@Kh3 z++Hs}+>CDgKs~>$6A%Ja+alul7G{zFJrDTgzZ0>m+B8j1ND?VX-pIO`Yk+Kbw^KSH zr`K(ES(@^g-yX)yaKZf;z>hZ!BSYf5`}2~(xkL5_9k$e7TbJTPCL*~XRQUUM$4t0l zbvu>v!au;N=9}MBVcD87^rbchSd&aceiZQBqyYm||FynVbzD(wUrY1@S4B4=L~cf* z6#wP*dw)?YMy>*aOmYF04z_wt`6%Q~gRw;kGQabX*!rs-HHYa5svj;zt>!p)1?(Px zI_ysY854N-Ef&tyj_C_TZ<&6qZlX-?Pv<3IKSEUcy|I3o&(YrlVlFX1sT)Qjb}c8c zxabE1Is*WEl|GPmm8~C8*0BlXp+q!bZq@X=)!jLi{)zf*a7Tsnbj;D6O5kck0SXBY z&mk8FekEOXn$!%#?5#*g{SLuqe)X#YS6Cvqwa~pjzOK%=2;P^4h#^aElDP_S>k~p0 zHamK0>u~a0-ISr2W^HcP5`+0>DfrBTd8LHy({`MfS*94WsJp9YfFhja0s}mmxCwUV z!pVTOp(6CZtC2yo9p9@<3E*qW|H({qjQw(3N^a{kN@f&NAY1N8Ualk4tj;(ZtD7x<~_fFurjoxPoxV>(~5r4v4S3vzs02(CDHLx+1FMc@zq{{H=`XUXd z?`ny3wVKV{E;5I1bI1j=yb%K#b6?gCNw&~0$~2Y@`w@V%Nwbc6_TA%$K&IMu3Chi) zsVPl4nuN{ZI`}*KmPzf81Zqli%6(umti*#BYfAr^u7{Dh8Tiy}Fdu~S!H(GMBjAI- z9{NA^8UDY=2Xq4^Kz%>Nk4?co`vvVny@7{>fg5|Z->>1}+kwrm(g+}tx}8#Kr5Ax?KY4vtL>D@L~FRu*@spCRszNn7_g-HU6 z7L5ih$p(f2CGexm_Pbq(B?JCd<1iS?_b<)|M%9M{WMlsf@Fs?pf}1K_>bO8UJU3l{ z0PE<>awv%oRg#eT{RBmv_`*{DdJ_1UcM8i}f>%QI9Ictr$vCRd z5Ir+@nJ1?|AzKVj$uX4^*BP@_d0?S4(0)}hG z%zT-DKqUHFSEjyZsKqcydz0QdBY;{6gXqI>>E7xzcu~v&a~AL+W(G950bM;qGdyaE zC^^Kdw_UD&5;_aewZI1F9&zvBV73tfs8k667Hv>~MG^C8rj%EDjJ3s2!kQppSPIyy z3R^!d(Vzv$s{wlQw$x0!?$Puzkmu>$-CF9O%6v)vtOKsGp`Rhdi0AnRiZ@!l-(T^z z@&bfEWV|v54UC=jBqS}D4X5301F|swUxFfdRs?!vGTGEKDKvQp1C(>Bz=BEm8pZ-! zqm}u7{(!ntE@uNbq>66BfBsG#?8TQ-ST<+inkkYYsc9pbhMt^MdmB1#hm%p zm78P#)Q>NJV7cA9?w13VZ+kypml%DyBXvHN>j_TV5BNC8UF4sB){2XOMXsEfyl3am zo#%M;yR_Zf_!T2UrLxvTzEug4&=n#L!BU5K3}Y(DyDaON5*E`kCBYS~PDEB{i@Rgi z!x1=`_A;I>kB3htk{sgm3SJEch2JtJ@^`{9U+da`2aEE&sr zu=h6f<6-3>F5h~up3j3*6Ap~p7$vJhJWJMw|A*Q_fWkHpTrP)Pc#V)eO9wKMt~q57 zu7dX>)<`4npli^_Wa=NqkQ$BgDGGucsJVX3mGxJM5vsAZ%|$>Hk9{YK7_=XdtZAS( zeC{tO@7f5M$EeTE1%4v7E)W@j<{C0kd5oMP?QtOZZdBdQ=;jJ(D(&s)`GCH+7IfVz z)zQa@s?OuFSKJDjp*J@cGl%v1GN59j&^LgvwL&+(EugIUAz`_l$Q}&cf+XD8IdE{9 z>@v_8ND@o2)pHx2sKxQt^-nW;ypPfHlcmxI#{c~WaNcWuktjwx&p0$60;o;o4h_i?|Eg}~3R!lwr z?$GV|BvFy7D3*SK1vst-<~C&B&Q4*7MUfXQ_hwvPjr^{4Ue&h<^>&^byu2-$>s++C zvwSh*A+a0Q_=|6JAJ3gB#|nwTr)R&-X#|GFH76^kO2FlnbtOt8=h$to?qTX6F96#fxpohxE*KudXJd9u&iV1&{z^QdrA zfB$+Q1Rh|GsIV2C;|d|HYUALt?p?l;Kpx{Wlh3r{`}t!BI__BTm>=Ki=Ar1YSOxm7 z!nBK=zs})hfkv*fIC+wwla^N6v~rqF&P$*&A%li` z__~L%etE6zIOIK$QW|dE)cBX8cpcwS2e)~dk2(bjS*qyNxWFpP&Bgd6Uy`RS9@G;3 z7c$DWE2~-4|BtGPX9IEMRND66GHChV2gijEf+B`OGsU@EnRqoYwLSWd&IY-O@H{CZ zhJ*54%G;-B-*Ws@4|ij8LIyVe@yu<4OA3T%0VlZ=SW6M(LwYC2xj8OknrhdtY%(1@ zFMT?ntfBdE?B>JLp@5K16{zw`F6_?G4e|Z{hOi|ximPA7G_Flh`2?BtWW0;){dPm5 z1;!YLoF9!+0}Ny8a)+O%21O7Ae!3eqC8C&5aU=(jY5Dp`Sz0i!rG!8({7d?uw`7*#5(2_Pa^PZ+B9wfb|OTT!*-EN{&GbUta>!MObG zX5lzRt~6k26YBz=qp8xd=dlcDsBEGG2a{jh`ZaN{zDFg1Gf;MusrY=Oy=6{J+s{^& z)24|LCva9V3BH;i)K*6B1>9eJr--5R-I(AY&u=s)Z;2ALhI0GuerZu$Sns| zR0zdpH(6bippqb()hAIJrtjJe@v^Q3yH~Ckt&N>0p+`|n@6sP^GJ0O3 zOQQ<8+*KieHbA?6jZRQkmzPbaIO8*~|0t^;kggLecQXl-za`P`e5iel>E6{3 z9!7ORh)&D2t!X*F87Dii4zE@NIac}d$aG@))gVOsPJdixt{&Dj^pZC8d!JQ|VVo}M$O{?O>W4oE2e)DvqHri7@ zLtnj-OY03y@c#fi&AP5ytW7ZGzMMfja_1&0goZY~;Y%N?OHy}zbt%>@38zFM-q+05 zDHD4*GyP&}3*|%KV+60~SH#e^u-rB_vAn0`1m|~(1z6Hsh!gXxToht&mjENm=(%2h zEWiD0?qP*ThTWzNT}FIU(~D%aOp2%)muMDk+XPhgL??>JAa`m;LL9r*HglCkcYm1U zh@5B-wQQ(pTR2aDgt<(g;A>Chl+8T7#qhTRSNTK-22&0^YO1VC34?M8|Y4#&%sss7pS{TKiC5Q{f($ZX`xuN zRPHVRjni*JeTwsQ7}_V#5C=-r#reyDlUCw&fFRBfYV4okbT1k00CkrDg^m9$(C`f_0; z^3Q9U2QD+()$1gKEF*JVECurm$PH|Y1M(Q{yvKOfw~3w(@J2)Lj1*tjb}2N)>UeF) zRcRiAtxSC8bA*oZl=|&4-WKV9fb%bSMT~1|-tKV>PySN{`~j&QRu=XbKAiK4{VySR Be Extensions +2. Click install next to the TensorRT-LLM Extension +3. Check that files are correctly downloaded + +```sh +ls ~\jan\extensions\@janhq\tensorrt-llm-extension\dist\bin +# Your Extension Folder should now include `nitro.exe`, among other artifacts needed to run TRT-LLM +``` + +## Download a Compatible Model +TensorRT-LLM can only run models in `TensorRT` format. These models, aka "TensorRT Engines", are prebuilt specifically for each target OS+GPU architecture. + +We offer a handful of precompiled models for Ampere and Ada cards that you can immediately download and play with: + +1. Restart the application and go to the Hub +2. Look for models with the `TensorRT-LLM` label in the recommended models list. Click download. This step might take some time. 🙏 + +![image](https://hackmd.io/_uploads/rJewrEgRp.png) + +3. Click use and start chatting! +4. You may need to allow Nitro in your network + +![alt text](image.png) + +:::info +Due to our limited resources, we only prebuilt a few demo models. You can always build your desired models directly on your machine. [Read here](##Build-your-own-TensorRT-models). +::: + +## Configure Settings + +You can customize the default parameters for how Jan runs TensorRT-LLM. + +:::info +coming soon +::: + +## Troubleshooting + +### Incompatible Extension vs Engine versions + +For now, the model versions are pinned to the extension versions. + +### Uninstall Extension + +1. Quit the app +2. Go to Settings > Extensions +3. Delete the entire Extensions folder. +4. Reopen the app, only the default extensions should be restored. + +### Install Nitro-TensorRT-LLM manually + +To manually build the artifacts needed to run the server and TensorRT-LLM, you can reference the source code. [Read here](https://github.com/janhq/nitro-tensorrt-llm?tab=readme-ov-file#quickstart). + +### Build your own TensorRT models + +:::info +coming soon +::: diff --git a/docs/sidebars.js b/docs/sidebars.js index 4c45cadbe..26cb09eb2 100644 --- a/docs/sidebars.js +++ b/docs/sidebars.js @@ -199,6 +199,19 @@ const sidebars = { "guides/models/integrate-remote", ] }, + { + type: "category", + label: "Inference Providers", + className: "head_SubMenu", + link: { + type: 'doc', + id: "guides/inference/README", + }, + items: [ + "guides/inference/llama-cpp", + "guides/inference/tensorrt-llm", + ] + }, { type: "category", label: "Extensions", From b4eff9a1083b0811bfba99b4283421377df6842f Mon Sep 17 00:00:00 2001 From: Nicole Zhu Date: Thu, 14 Mar 2024 20:55:03 +0800 Subject: [PATCH 33/37] docs: fix slugs --- docs/docs/guides/{inference => engines}/README.mdx | 2 +- docs/docs/guides/{inference => engines}/image.png | Bin .../docs/guides/{inference => engines}/llama-cpp.md | 1 + .../guides/{inference => engines}/tensorrt-llm.md | 3 ++- docs/sidebars.js | 8 ++++---- 5 files changed, 8 insertions(+), 6 deletions(-) rename docs/docs/guides/{inference => engines}/README.mdx (73%) rename docs/docs/guides/{inference => engines}/image.png (100%) rename docs/docs/guides/{inference => engines}/llama-cpp.md (89%) rename docs/docs/guides/{inference => engines}/tensorrt-llm.md (95%) diff --git a/docs/docs/guides/inference/README.mdx b/docs/docs/guides/engines/README.mdx similarity index 73% rename from docs/docs/guides/inference/README.mdx rename to docs/docs/guides/engines/README.mdx index 289fd8241..3a7cdcc44 100644 --- a/docs/docs/guides/inference/README.mdx +++ b/docs/docs/guides/engines/README.mdx @@ -1,6 +1,6 @@ --- title: Extensions -slug: /guides/inference/ +slug: /guides/engines --- import DocCardList from "@theme/DocCardList"; diff --git a/docs/docs/guides/inference/image.png b/docs/docs/guides/engines/image.png similarity index 100% rename from docs/docs/guides/inference/image.png rename to docs/docs/guides/engines/image.png diff --git a/docs/docs/guides/inference/llama-cpp.md b/docs/docs/guides/engines/llama-cpp.md similarity index 89% rename from docs/docs/guides/inference/llama-cpp.md rename to docs/docs/guides/engines/llama-cpp.md index 470424dbb..bc485df6c 100644 --- a/docs/docs/guides/inference/llama-cpp.md +++ b/docs/docs/guides/engines/llama-cpp.md @@ -1,5 +1,6 @@ --- title: Llama-CPP Extension +slug: /guides/engines/llama-cpp --- ## Overview diff --git a/docs/docs/guides/inference/tensorrt-llm.md b/docs/docs/guides/engines/tensorrt-llm.md similarity index 95% rename from docs/docs/guides/inference/tensorrt-llm.md rename to docs/docs/guides/engines/tensorrt-llm.md index d795d55e6..177cc0cf4 100644 --- a/docs/docs/guides/inference/tensorrt-llm.md +++ b/docs/docs/guides/engines/tensorrt-llm.md @@ -1,5 +1,6 @@ --- title: TensorRT-LLM Extension +slug: /guides/engines/tensorrt-llm --- Users with Nvidia GPUs can get 20-40% faster* token speeds on their laptop or desktops by using [TensorRT-LLM](https://github.com/NVIDIA/TensorRT-LLM). @@ -48,7 +49,7 @@ We offer a handful of precompiled models for Ampere and Ada cards that you can i ![alt text](image.png) :::info -Due to our limited resources, we only prebuilt a few demo models. You can always build your desired models directly on your machine. [Read here](##Build-your-own-TensorRT-models). +Due to our limited resources, we only prebuilt a few demo models. You can always build your desired models directly on your machine. [Read here](#build-your-own-tensorrt-models). ::: ## Configure Settings diff --git a/docs/sidebars.js b/docs/sidebars.js index 26cb09eb2..8deafeaa1 100644 --- a/docs/sidebars.js +++ b/docs/sidebars.js @@ -201,15 +201,15 @@ const sidebars = { }, { type: "category", - label: "Inference Providers", + label: "AI Engines", className: "head_SubMenu", link: { type: 'doc', - id: "guides/inference/README", + id: "guides/engines/README", }, items: [ - "guides/inference/llama-cpp", - "guides/inference/tensorrt-llm", + "guides/engines/llama-cpp", + "guides/engines/tensorrt-llm", ] }, { From ab73941d4256549b027c28dc0bfd1a27f1c7b1bf Mon Sep 17 00:00:00 2001 From: Service Account Date: Thu, 14 Mar 2024 12:55:07 +0000 Subject: [PATCH 34/37] janhq/jan: Update README.md with nightly build artifact URL --- README.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index b25a917d7..6f6044b30 100644 --- a/README.md +++ b/README.md @@ -76,31 +76,31 @@ Jan is an open-source ChatGPT alternative that runs 100% offline on your compute Experimental (Nightly Build) - + jan.exe - + Intel - + M1/M2 - + jan.deb - + jan.AppImage From 1fe3dff875de61b53420058200741d5eca7fab7b Mon Sep 17 00:00:00 2001 From: Nicole Zhu Date: Thu, 14 Mar 2024 21:08:50 +0800 Subject: [PATCH 35/37] docs: update slugs again --- docs/docs/guides/engines/llama-cpp.md | 12 ------------ docs/docs/guides/{engines => providers}/README.mdx | 4 ++-- docs/docs/guides/{engines => providers}/image.png | Bin docs/docs/guides/providers/llama-cpp.md | 10 ++++++++++ .../guides/{engines => providers}/tensorrt-llm.md | 4 ++-- docs/sidebars.js | 8 ++++---- 6 files changed, 18 insertions(+), 20 deletions(-) delete mode 100644 docs/docs/guides/engines/llama-cpp.md rename docs/docs/guides/{engines => providers}/README.mdx (54%) rename docs/docs/guides/{engines => providers}/image.png (100%) create mode 100644 docs/docs/guides/providers/llama-cpp.md rename docs/docs/guides/{engines => providers}/tensorrt-llm.md (95%) diff --git a/docs/docs/guides/engines/llama-cpp.md b/docs/docs/guides/engines/llama-cpp.md deleted file mode 100644 index bc485df6c..000000000 --- a/docs/docs/guides/engines/llama-cpp.md +++ /dev/null @@ -1,12 +0,0 @@ ---- -title: Llama-CPP Extension -slug: /guides/engines/llama-cpp ---- - -## Overview - -[LlamaCPP](https://github.com/ggerganov/llama.cpp) is the default AI engine downloaded with Jan. It is served through Nitro, a C++ inference server, that handles additional UX and hardware optimizations. - -The source code for Nitro-llama-cpp is [here](https://github.com/janhq/nitro). - -There is no additional setup needed. \ No newline at end of file diff --git a/docs/docs/guides/engines/README.mdx b/docs/docs/guides/providers/README.mdx similarity index 54% rename from docs/docs/guides/engines/README.mdx rename to docs/docs/guides/providers/README.mdx index 3a7cdcc44..aa3bfea1f 100644 --- a/docs/docs/guides/engines/README.mdx +++ b/docs/docs/guides/providers/README.mdx @@ -1,6 +1,6 @@ --- -title: Extensions -slug: /guides/engines +title: Inference Providers +slug: /guides/providers --- import DocCardList from "@theme/DocCardList"; diff --git a/docs/docs/guides/engines/image.png b/docs/docs/guides/providers/image.png similarity index 100% rename from docs/docs/guides/engines/image.png rename to docs/docs/guides/providers/image.png diff --git a/docs/docs/guides/providers/llama-cpp.md b/docs/docs/guides/providers/llama-cpp.md new file mode 100644 index 000000000..3a21e80a7 --- /dev/null +++ b/docs/docs/guides/providers/llama-cpp.md @@ -0,0 +1,10 @@ +--- +title: llama.cpp +slug: /guides/providers/llama-cpp +--- + +## Overview + +[Nitro](https://github.com/janhq/nitro) is an inference server on top of [llama.cpp](https://github.com/ggerganov/llama.cpp). OpenAI-compatible API, queue, & scaling. + +Nitro is the default AI engine downloaded with Jan. There is no additional setup needed. \ No newline at end of file diff --git a/docs/docs/guides/engines/tensorrt-llm.md b/docs/docs/guides/providers/tensorrt-llm.md similarity index 95% rename from docs/docs/guides/engines/tensorrt-llm.md rename to docs/docs/guides/providers/tensorrt-llm.md index 177cc0cf4..4b0edec2a 100644 --- a/docs/docs/guides/engines/tensorrt-llm.md +++ b/docs/docs/guides/providers/tensorrt-llm.md @@ -1,6 +1,6 @@ --- -title: TensorRT-LLM Extension -slug: /guides/engines/tensorrt-llm +title: TensorRT-LLM +slug: /guides/providers/tensorrt-llm --- Users with Nvidia GPUs can get 20-40% faster* token speeds on their laptop or desktops by using [TensorRT-LLM](https://github.com/NVIDIA/TensorRT-LLM). diff --git a/docs/sidebars.js b/docs/sidebars.js index 8deafeaa1..b95e4044f 100644 --- a/docs/sidebars.js +++ b/docs/sidebars.js @@ -201,15 +201,15 @@ const sidebars = { }, { type: "category", - label: "AI Engines", + label: "Inference Providers", className: "head_SubMenu", link: { type: 'doc', - id: "guides/engines/README", + id: "guides/providers/README", }, items: [ - "guides/engines/llama-cpp", - "guides/engines/tensorrt-llm", + "guides/providers/llama-cpp", + "guides/providers/tensorrt-llm", ] }, { From 70fc24f6f1857f19d0d837fe95086a373443f210 Mon Sep 17 00:00:00 2001 From: Nicole Zhu Date: Thu, 14 Mar 2024 21:12:16 +0800 Subject: [PATCH 36/37] docs: nits --- docs/docs/guides/providers/llama-cpp.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/docs/guides/providers/llama-cpp.md b/docs/docs/guides/providers/llama-cpp.md index 3a21e80a7..d2b0daa2a 100644 --- a/docs/docs/guides/providers/llama-cpp.md +++ b/docs/docs/guides/providers/llama-cpp.md @@ -5,6 +5,6 @@ slug: /guides/providers/llama-cpp ## Overview -[Nitro](https://github.com/janhq/nitro) is an inference server on top of [llama.cpp](https://github.com/ggerganov/llama.cpp). OpenAI-compatible API, queue, & scaling. +[Nitro](https://github.com/janhq/nitro) is an inference server on top of [llama.cpp](https://github.com/ggerganov/llama.cpp). It provides an OpenAI-compatible API, queue, & scaling. Nitro is the default AI engine downloaded with Jan. There is no additional setup needed. \ No newline at end of file From 4b6f218639b5c3f5dc0f231ea88af386c0b1c9c0 Mon Sep 17 00:00:00 2001 From: Nicole Zhu Date: Thu, 14 Mar 2024 21:45:31 +0800 Subject: [PATCH 37/37] docs: nits --- docs/docs/guides/providers/tensorrt-llm.md | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) diff --git a/docs/docs/guides/providers/tensorrt-llm.md b/docs/docs/guides/providers/tensorrt-llm.md index 4b0edec2a..52da83b36 100644 --- a/docs/docs/guides/providers/tensorrt-llm.md +++ b/docs/docs/guides/providers/tensorrt-llm.md @@ -3,14 +3,17 @@ title: TensorRT-LLM slug: /guides/providers/tensorrt-llm --- -Users with Nvidia GPUs can get 20-40% faster* token speeds on their laptop or desktops by using [TensorRT-LLM](https://github.com/NVIDIA/TensorRT-LLM). +Users with Nvidia GPUs can get **20-40% faster\* token speeds** on their laptop or desktops by using [TensorRT-LLM](https://github.com/NVIDIA/TensorRT-LLM). The greater implication is that you are running FP16, which is also more accurate than quantized models. This guide walks you through how to install Jan's official [TensorRT-LLM Extension](https://github.com/janhq/nitro-tensorrt-llm). This extension uses [Nitro-TensorRT-LLM](https://github.com/janhq/nitro-tensorrt-llm) as the AI engine, instead of the default [Nitro-Llama-CPP](https://github.com/janhq/nitro). It includes an efficient C++ server to natively execute the [TRT-LLM C++ runtime](https://nvidia.github.io/TensorRT-LLM/gpt_runtime.html). It also comes with additional feature and performance improvements like OpenAI compatibility, tokenizer improvements, and queues. *Compared to using LlamaCPP engine. -:::info +:::warning This feature is only available for Windows users. Linux is coming soon. + +Additionally, we only prebuilt a few demo models. You can always build your desired models directly on your machine. [Read here](#build-your-own-tensorrt-models). + ::: ## Requirements @@ -48,8 +51,8 @@ We offer a handful of precompiled models for Ampere and Ada cards that you can i ![alt text](image.png) -:::info -Due to our limited resources, we only prebuilt a few demo models. You can always build your desired models directly on your machine. [Read here](#build-your-own-tensorrt-models). +:::warning +If you are our nightly builds, you may have to reinstall the TensorRT-LLM extension each time you update the app. We're working on better extension lifecyles - stay tuned. ::: ## Configure Settings