Merge pull request #1948 from janhq/dev

Sync dev branch to docs branch

.devcontainer/devcontainer.json

@@ -0,0 +1,4 @@
+{
+    "name": "jan",
+    "image": "node:20"
+}

.github/scripts/auto-sign.sh

@@ -8,3 +8,5 @@ fi
 
 # If both variables are set, execute the following commands
 find "$APP_PATH" \( -type f -perm +111 -o -name "*.node" \) -exec codesign -s "$DEVELOPER_ID" --options=runtime {} \;
+
+find "$APP_PATH" -type f -name "*.o" -exec codesign -s "$DEVELOPER_ID" --options=runtime {} \;

.github/workflows/clean-cloudflare-page-preview-url-and-r2.yml

@@ -55,10 +55,10 @@ jobs:
     steps:
       - name: install-aws-cli-action
         uses: unfor19/install-aws-cli-action@v1
-      - name: Delete object older than 7 days
+      - name: Delete object older than 10 days
         run: |
           # Get the list of objects in the 'latest' folder
-          OBJECTS=$(aws s3api list-objects --bucket ${{ secrets.CLOUDFLARE_R2_BUCKET_NAME }} --query 'Contents[?LastModified<`'$(date -d "$current_date -30 days" -u +"%Y-%m-%dT%H:%M:%SZ")'`].{Key: Key}' --endpoint-url https://${{ secrets.CLOUDFLARE_ACCOUNT_ID }}.r2.cloudflarestorage.com | jq -c .)
+          OBJECTS=$(aws s3api list-objects --bucket ${{ secrets.CLOUDFLARE_R2_BUCKET_NAME }} --query 'Contents[?LastModified<`'$(date -d "$current_date -10 days" -u +"%Y-%m-%dT%H:%M:%SZ")'`].{Key: Key}' --endpoint-url https://${{ secrets.CLOUDFLARE_ACCOUNT_ID }}.r2.cloudflarestorage.com | jq -c .)
           
           # Create a JSON file for the delete operation
           echo "{\"Objects\": $OBJECTS, \"Quiet\": false}" > delete.json

.github/workflows/jan-electron-build-nightly.yml

@@ -1,8 +1,14 @@
 name: Jan Build Electron App Nightly or Manual
 
 on:
+  push:
+    branches:
+      - main
+    paths-ignore:
+      - 'README.md'
+      - 'docs/**'
   schedule:
-    - cron: '0 20 * * 2,3,4' # At 8 PM UTC on Tuesday, Wednesday, and Thursday, which is 3 AM UTC+7
+    - cron: '0 20 * * 1,2,3' # At 8 PM UTC on Monday, Tuesday, and Wednesday which is 3 AM UTC+7 Tuesday, Wednesday, and Thursday
   workflow_dispatch:
     inputs:
       public_provider:
@@ -23,19 +29,36 @@ jobs:
       - name: Set public provider
         id: set-public-provider
         run: |
-          if [ ${{ github.event == 'workflow_dispatch' }} ]; then
+          if [ "${{ github.event_name }}" == "workflow_dispatch" ]; then
             echo "::set-output name=public_provider::${{ github.event.inputs.public_provider }}"
             echo "::set-output name=ref::${{ github.ref }}"
           else
+            if [ "${{ github.event_name }}" == "schedule" ]; then
               echo "::set-output name=public_provider::cloudflare-r2"
               echo "::set-output name=ref::refs/heads/dev"
+            elif [ "${{ github.event_name }}" == "push" ]; then
+              echo "::set-output name=public_provider::cloudflare-r2"
+              echo "::set-output name=ref::${{ github.ref }}"
+            else
+              echo "::set-output name=public_provider::none"
+              echo "::set-output name=ref::${{ github.ref }}"
+            fi
           fi
   # Job create Update app version based on latest release tag with build number and save to output
   get-update-version:
     uses: ./.github/workflows/template-get-update-version.yml
 
-  build-macos:
+  build-macos-x64:
-    uses: ./.github/workflows/template-build-macos.yml
+    uses: ./.github/workflows/template-build-macos-x64.yml
+    needs: [get-update-version, set-public-provider]
+    secrets: inherit
+    with:
+      ref: ${{ needs.set-public-provider.outputs.ref }}
+      public_provider: ${{ needs.set-public-provider.outputs.public_provider }}
+      new_version: ${{ needs.get-update-version.outputs.new_version }}
+
+  build-macos-arm64:
+    uses: ./.github/workflows/template-build-macos-arm64.yml
     needs: [get-update-version, set-public-provider]
     secrets: inherit
     with:
@@ -62,8 +85,51 @@ jobs:
       public_provider: ${{ needs.set-public-provider.outputs.public_provider }}
       new_version: ${{ needs.get-update-version.outputs.new_version }}
 
+  combine-latest-mac-yml:
+    needs: [set-public-provider, build-macos-x64, build-macos-arm64]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Getting the repo
+        uses: actions/checkout@v3
+        with:
+          ref: ${{ needs.set-public-provider.outputs.ref }}
+      - name: Download mac-x64 artifacts
+        uses: actions/download-artifact@v3
+        with:
+          name: latest-mac-x64
+          path: ./latest-mac-x64
+      - name: Download mac-arm artifacts
+        uses: actions/download-artifact@v3
+        with:
+          name: latest-mac-arm64
+          path: ./latest-mac-arm64
+
+      - name: 'Merge latest-mac.yml'
+        # unfortunately electron-builder doesn't understand that we have two different releases for mac-x64 and mac-arm, so we need to manually merge the latest files
+        # see https://github.com/electron-userland/electron-builder/issues/5592
+        run: |
+          ls -la .
+          ls -la ./latest-mac-x64
+          ls -la ./latest-mac-arm64
+          ls -la ./electron
+          cp ./electron/merge-latest-ymls.js /tmp/merge-latest-ymls.js
+          npm install js-yaml --prefix /tmp
+          node /tmp/merge-latest-ymls.js ./latest-mac-x64/latest-mac.yml ./latest-mac-arm64/latest-mac.yml ./latest-mac.yml
+          cat ./latest-mac.yml
+      
+      - name: Upload latest-mac.yml
+        if: ${{ needs.set-public-provider.outputs.public_provider == 'cloudflare-r2' }}
+        run: |
+          aws s3api put-object --endpoint-url https://${{ secrets.CLOUDFLARE_ACCOUNT_ID }}.r2.cloudflarestorage.com --bucket ${{ secrets.CLOUDFLARE_R2_BUCKET_NAME }} --key "latest/latest-mac.yml" --body "./latest-mac.yml"
+        env:
+          AWS_ACCESS_KEY_ID: ${{ secrets.CLOUDFLARE_R2_ACCESS_KEY_ID }}
+          AWS_SECRET_ACCESS_KEY: ${{ secrets.CLOUDFLARE_R2_SECRET_ACCESS_KEY }}
+          AWS_DEFAULT_REGION: auto
+          AWS_EC2_METADATA_DISABLED: "true"
+
+    
   noti-discord-nightly-and-update-url-readme:
-    needs: [build-macos, build-windows-x64, build-linux-x64, get-update-version, set-public-provider]
+    needs: [build-macos-x64, build-macos-arm64, build-windows-x64, build-linux-x64, get-update-version, set-public-provider, combine-latest-mac-yml]
     secrets: inherit
     if: github.event_name == 'schedule'
     uses: ./.github/workflows/template-noti-discord-and-update-url-readme.yml
@@ -73,8 +139,19 @@ jobs:
       push_to_branch: dev
       new_version: ${{ needs.get-update-version.outputs.new_version }}
 
+  noti-discord-pre-release-and-update-url-readme:
+    needs: [build-macos-x64, build-macos-arm64, build-windows-x64, build-linux-x64, get-update-version, set-public-provider, combine-latest-mac-yml]
+    secrets: inherit
+    if: github.event_name == 'push'
+    uses: ./.github/workflows/template-noti-discord-and-update-url-readme.yml
+    with:
+      ref: refs/heads/dev
+      build_reason: Pre-release
+      push_to_branch: dev
+      new_version: ${{ needs.get-update-version.outputs.new_version }}
+
   noti-discord-manual-and-update-url-readme:
-    needs: [build-macos, build-windows-x64, build-linux-x64, get-update-version, set-public-provider]
+    needs: [build-macos-x64, build-macos-arm64, build-windows-x64, build-linux-x64, get-update-version, set-public-provider, combine-latest-mac-yml]
     secrets: inherit
     if: github.event_name == 'workflow_dispatch' && github.event.inputs.public_provider == 'cloudflare-r2'
     uses: ./.github/workflows/template-noti-discord-and-update-url-readme.yml

.github/workflows/jan-electron-build-pre-release.yml

@@ -1,52 +0,0 @@
-name: Jan Build Electron Pre Release
-
-on:
-  push:
-    branches:
-      - main
-    paths:
-      - "!README.md"
-
-jobs:
-
-  # Job create Update app version based on latest release tag with build number and save to output
-  get-update-version:
-    uses: ./.github/workflows/template-get-update-version.yml
-
-  build-macos:
-    uses: ./.github/workflows/template-build-macos.yml
-    secrets: inherit
-    needs: [get-update-version]
-    with:
-      ref: ${{ github.ref }}
-      public_provider: cloudflare-r2
-      new_version: ${{ needs.get-update-version.outputs.new_version }}
-
-  build-windows-x64:
-    uses: ./.github/workflows/template-build-windows-x64.yml
-    secrets: inherit
-    needs: [get-update-version]
-    with:
-      ref: ${{ github.ref }}
-      public_provider: cloudflare-r2
-      new_version: ${{ needs.get-update-version.outputs.new_version }}
-
-  build-linux-x64:
-    uses: ./.github/workflows/template-build-linux-x64.yml
-    secrets: inherit
-    needs: [get-update-version]
-    with:
-      ref: ${{ github.ref }}
-      public_provider: cloudflare-r2
-      new_version: ${{ needs.get-update-version.outputs.new_version }}
-
-  noti-discord-nightly-and-update-url-readme:
-    needs: [build-macos, build-windows-x64, build-linux-x64, get-update-version]
-    secrets: inherit
-    if: github.event_name == 'push' && github.ref == 'refs/heads/main'
-    uses: ./.github/workflows/template-noti-discord-and-update-url-readme.yml
-    with:
-      ref: refs/heads/dev
-      build_reason: Nightly
-      push_to_branch: dev
-      new_version: ${{ needs.get-update-version.outputs.new_version }}

.github/workflows/jan-electron-build.yml

@@ -9,8 +9,42 @@ jobs:
   get-update-version:
     uses: ./.github/workflows/template-get-update-version.yml
 
-  build-macos:
+  create-draft-release:
-    uses: ./.github/workflows/template-build-macos.yml
+    runs-on: ubuntu-latest
+    if: github.event_name == 'push' && startsWith(github.ref, 'refs/tags/')
+    outputs:
+      upload_url: ${{ steps.create_release.outputs.upload_url }}
+      version: ${{ steps.get_version.outputs.version }}
+    permissions:
+      contents: write
+    steps:
+      - name: Extract tag name without v prefix
+        id: get_version
+        run: echo "VERSION=${GITHUB_REF#refs/tags/v}" >> $GITHUB_ENV && echo "::set-output name=version::${GITHUB_REF#refs/tags/v}"
+        env:
+          GITHUB_REF: ${{ github.ref }}
+      - name: Create Draft Release
+        id: create_release
+        uses: actions/create-release@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          tag_name: ${{ github.ref_name }}
+          release_name: "${{ env.VERSION }}"
+          draft: true
+          prerelease: false
+
+  build-macos-x64:
+    uses: ./.github/workflows/template-build-macos-x64.yml
+    secrets: inherit
+    needs: [get-update-version]
+    with:
+      ref: ${{ github.ref }}
+      public_provider: github
+      new_version: ${{ needs.get-update-version.outputs.new_version }}
+
+  build-macos-arm64:
+    uses: ./.github/workflows/template-build-macos-arm64.yml
     secrets: inherit
     needs: [get-update-version]
     with:
@@ -36,8 +70,52 @@ jobs:
       public_provider: github
       new_version: ${{ needs.get-update-version.outputs.new_version }}
 
+  combine-latest-mac-yml:
+    needs: [build-macos-x64, build-macos-arm64, create-draft-release]
+    runs-on: ubuntu-latest
+    permissions:
+      contents: write
+    steps:
+      - name: Getting the repo
+        uses: actions/checkout@v3
+
+      - name: Download mac-x64 artifacts
+        uses: actions/download-artifact@v3
+        with:
+          name: latest-mac-x64
+          path: ./latest-mac-x64
+      - name: Download mac-arm artifacts
+        uses: actions/download-artifact@v3
+        with:
+          name: latest-mac-arm64
+          path: ./latest-mac-arm64
+
+      - name: 'Merge latest-mac.yml'
+        # unfortunately electron-builder doesn't understand that we have two different releases for mac-x64 and mac-arm, so we need to manually merge the latest files
+        # see https://github.com/electron-userland/electron-builder/issues/5592
+        run: |
+          ls -la .
+          ls -la ./latest-mac-x64
+          ls -la ./latest-mac-arm64
+          ls -la ./electron
+          cp ./electron/merge-latest-ymls.js /tmp/merge-latest-ymls.js
+          npm install js-yaml --prefix /tmp
+          node /tmp/merge-latest-ymls.js ./latest-mac-x64/latest-mac.yml ./latest-mac-arm64/latest-mac.yml ./latest-mac.yml
+          cat ./latest-mac.yml
+      
+      - name: Yet Another Upload Release Asset Action  
+        uses: shogo82148/actions-upload-release-asset@v1.7.2
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          upload_url: ${{ needs.create-draft-release.outputs.upload_url }}
+          asset_path: ./latest-mac.yml
+          asset_name: latest-mac.yml
+          asset_content_type: text/yaml
+          overwrite: true
+
   update_release_draft:
-    needs: [build-macos, build-windows-x64, build-linux-x64]
+    needs: [build-macos-x64, build-macos-arm64, build-windows-x64, build-linux-x64, combine-latest-mac-yml]
     permissions:
       # write permission is required to create a github release
       contents: write

.github/workflows/template-build-linux-x64.yml

@@ -98,8 +98,8 @@ jobs:
           make build-and-publish
         env:
           GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          ANALYTICS_ID: ${{ secrets.JAN_APP_POSTHOG_PROJECT_API_KEY }}
+          ANALYTICS_ID: ${{ secrets.JAN_APP_UMAMI_PROJECT_API_KEY }}
-          ANALYTICS_HOST: ${{ secrets.JAN_APP_POSTHOG_URL }}
+          ANALYTICS_HOST: ${{ secrets.JAN_APP_UMAMI_URL }}
 
       - name: Upload Artifact .deb file
         if: inputs.public_provider != 'github'

.github/workflows/template-build-macos-arm64.yml

@@ -0,0 +1,160 @@
+name: build-macos
+on:
+  workflow_call:
+    inputs:
+      ref:
+        required: true
+        type: string
+        default: 'refs/heads/main'
+      public_provider:
+        required: true
+        type: string
+        default: none
+        description: 'none: build only, github: build and publish to github, cloudflare: build and publish to cloudflare'
+      new_version:
+        required: true
+        type: string
+        default: ''
+      cloudflare_r2_path:
+        required: false
+        type: string
+        default: '/latest/'
+    secrets:
+      CLOUDFLARE_R2_BUCKET_NAME:
+        required: false
+      CLOUDFLARE_R2_ACCESS_KEY_ID:
+        required: false
+      CLOUDFLARE_R2_SECRET_ACCESS_KEY:
+        required: false
+      CLOUDFLARE_ACCOUNT_ID:
+        required: false
+      CODE_SIGN_P12_BASE64:
+        required: false
+      CODE_SIGN_P12_PASSWORD:
+        required: false
+      APPLE_ID:
+        required: false
+      APPLE_APP_SPECIFIC_PASSWORD:
+        required: false
+      DEVELOPER_ID:
+        required: false
+
+jobs:
+  build-macos:
+    runs-on: macos-silicon
+    environment: production
+    permissions:
+      contents: write
+    steps:
+      - name: Getting the repo
+        uses: actions/checkout@v3
+        with:
+          ref: ${{ inputs.ref }}
+
+      - name: Installing node
+        uses: actions/setup-node@v1
+        with:
+          node-version: 20
+      - name: Unblock keychain
+        run: |
+          security unlock-keychain -p ${{ secrets.KEYCHAIN_PASSWORD }} ~/Library/Keychains/login.keychain-db
+      # - uses: actions/setup-python@v5
+      #   with:
+      #     python-version: '3.11'
+
+      # - name: Install jq
+      #   uses: dcarbone/install-jq-action@v2.0.1
+
+      - name: Update app version based on latest release tag with build number
+        if: inputs.public_provider != 'github'
+        run: |
+          echo "Version: ${{ inputs.new_version }}"
+          # Update the version in electron/package.json
+          jq --arg version "${{ inputs.new_version }}" '.version = $version' electron/package.json > /tmp/package.json
+          mv /tmp/package.json electron/package.json
+
+          jq --arg version "${{ inputs.new_version }}" '.version = $version' web/package.json > /tmp/package.json
+          mv /tmp/package.json web/package.json
+
+          jq '.build.publish = [{"provider": "generic", "url": "${{ secrets.CLOUDFLARE_R2_PUBLIC_URL }}", "channel": "latest"}, {"provider": "s3", "bucket": "${{ secrets.CLOUDFLARE_R2_BUCKET_NAME }}", "region": "auto", "endpoint": "https://${{ secrets.CLOUDFLARE_ACCOUNT_ID }}.r2.cloudflarestorage.com", "path": "${{ inputs.cloudflare_r2_path }}", "channel": "latest"}]' electron/package.json > /tmp/package.json
+          mv /tmp/package.json electron/package.json
+          cat electron/package.json
+
+      - name: Update app version base on tag
+        if: github.event_name == 'push' && startsWith(github.ref, 'refs/tags/') && inputs.public_provider == 'github'
+        run: |
+          if [[ ! "${VERSION_TAG}" =~ ^v[0-9]+\.[0-9]+\.[0-9]+$ ]]; then
+              echo "Error: Tag is not valid!"
+              exit 1
+          fi
+          jq --arg version "${VERSION_TAG#v}" '.version = $version' electron/package.json > /tmp/package.json
+          mv /tmp/package.json electron/package.json
+          jq --arg version "${VERSION_TAG#v}" '.version = $version' web/package.json > /tmp/package.json
+          mv /tmp/package.json web/package.json
+        env:
+          VERSION_TAG: ${{ inputs.new_version }}
+
+      # - name: Get Cer for code signing
+      #   run: base64 -d <<< "$CODE_SIGN_P12_BASE64" > /tmp/codesign.p12
+      #   shell: bash
+      #   env:
+      #     CODE_SIGN_P12_BASE64: ${{ secrets.CODE_SIGN_P12_BASE64 }}
+
+      # - uses: apple-actions/import-codesign-certs@v2
+      #   continue-on-error: true
+      #   with:
+      #     p12-file-base64: ${{ secrets.CODE_SIGN_P12_BASE64 }}
+      #     p12-password: ${{ secrets.CODE_SIGN_P12_PASSWORD }}
+
+      - name: Build and publish app to cloudflare r2 or github artifactory
+        if: inputs.public_provider != 'github'
+        run: |
+          # check public_provider is true or not
+          echo "public_provider is ${{ inputs.public_provider }}"
+          if [ "${{ inputs.public_provider }}" == "none" ]; then
+            make build    
+          else
+            make build-and-publish
+          fi
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          # CSC_LINK: "/tmp/codesign.p12"
+          # CSC_KEY_PASSWORD: ${{ secrets.CODE_SIGN_P12_PASSWORD }}
+          # CSC_IDENTITY_AUTO_DISCOVERY: "true"
+          APPLE_ID: ${{ secrets.APPLE_ID }}
+          APPLE_APP_SPECIFIC_PASSWORD: ${{ secrets.APPLE_APP_SPECIFIC_PASSWORD }}
+          APP_PATH: "."
+          DEVELOPER_ID: ${{ secrets.DEVELOPER_ID }}
+          AWS_ACCESS_KEY_ID: ${{ secrets.CLOUDFLARE_R2_ACCESS_KEY_ID }}
+          AWS_SECRET_ACCESS_KEY: ${{ secrets.CLOUDFLARE_R2_SECRET_ACCESS_KEY }}
+          AWS_DEFAULT_REGION: auto
+          AWS_EC2_METADATA_DISABLED: "true"
+
+      - name: Build and publish app to github
+        if: github.event_name == 'push' && startsWith(github.ref, 'refs/tags/') && inputs.public_provider == 'github'
+        run: |
+          make build-and-publish
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          # CSC_LINK: "/tmp/codesign.p12"
+          # CSC_KEY_PASSWORD: ${{ secrets.CODE_SIGN_P12_PASSWORD }}
+          # CSC_IDENTITY_AUTO_DISCOVERY: "true"
+          APPLE_ID: ${{ secrets.APPLE_ID }}
+          APPLE_APP_SPECIFIC_PASSWORD: ${{ secrets.APPLE_APP_SPECIFIC_PASSWORD }}
+          APP_PATH: "."
+          DEVELOPER_ID: ${{ secrets.DEVELOPER_ID }}
+          ANALYTICS_ID: ${{ secrets.JAN_APP_UMAMI_PROJECT_API_KEY }}
+          ANALYTICS_HOST: ${{ secrets.JAN_APP_UMAMI_URL }}
+
+      - name: Upload Artifact
+        if: inputs.public_provider != 'github'
+        uses: actions/upload-artifact@v2
+        with:
+          name: jan-mac-arm64-${{ inputs.new_version }}
+          path: ./electron/dist/jan-mac-arm64-${{ inputs.new_version }}.dmg
+
+      - name: Upload Artifact
+        uses: actions/upload-artifact@v2
+        with:
+          name: latest-mac-arm64
+          path: ./electron/dist/latest-mac.yml

.github/workflows/template-build-macos-x64.yml

@@ -137,8 +137,8 @@ jobs:
           APPLE_APP_SPECIFIC_PASSWORD: ${{ secrets.APPLE_APP_SPECIFIC_PASSWORD }}
           APP_PATH: "."
           DEVELOPER_ID: ${{ secrets.DEVELOPER_ID }}
-          ANALYTICS_ID: ${{ secrets.JAN_APP_POSTHOG_PROJECT_API_KEY }}
+          ANALYTICS_ID: ${{ secrets.JAN_APP_UMAMI_PROJECT_API_KEY }}
-          ANALYTICS_HOST: ${{ secrets.JAN_APP_POSTHOG_URL }}
+          ANALYTICS_HOST: ${{ secrets.JAN_APP_UMAMI_URL }}
 
       - name: Upload Artifact
         if: inputs.public_provider != 'github'
@@ -148,9 +148,8 @@ jobs:
           path: ./electron/dist/jan-mac-x64-${{ inputs.new_version }}.dmg
 
       - name: Upload Artifact
-        if: inputs.public_provider != 'github'
         uses: actions/upload-artifact@v2
         with:
-          name: jan-mac-arm64-${{ inputs.new_version }}
+          name: latest-mac-x64
-          path: ./electron/dist/jan-mac-arm64-${{ inputs.new_version }}.dmg
+          path: ./electron/dist/latest-mac.yml
 

.github/workflows/template-build-windows-x64.yml

@@ -127,8 +127,8 @@ jobs:
           make build-and-publish
         env:
           GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          ANALYTICS_ID: ${{ secrets.JAN_APP_POSTHOG_PROJECT_API_KEY }}
+          ANALYTICS_ID: ${{ secrets.JAN_APP_UMAMI_PROJECT_API_KEY }}
-          ANALYTICS_HOST: ${{ secrets.JAN_APP_POSTHOG_URL }}
+          ANALYTICS_HOST: ${{ secrets.JAN_APP_UMAMI_URL }}
           AZURE_KEY_VAULT_URI: ${{ secrets.AZURE_KEY_VAULT_URI }}
           AZURE_CLIENT_ID: ${{ secrets.AZURE_CLIENT_ID }}
           AZURE_TENANT_ID: ${{ secrets.AZURE_TENANT_ID }}

.github/workflows/update-release-url.yml

@@ -17,7 +17,7 @@ jobs:
         with:
           fetch-depth: "0"
           token: ${{ secrets.PAT_SERVICE_ACCOUNT }}
-          ref: main
+          ref: dev
 
       - name: Get Latest Release
         uses: pozetroninc/github-action-get-latest-release@v0.7.0
@@ -46,4 +46,4 @@ jobs:
           git config --global user.name "Service Account"
           git add README.md
           git commit -m "Update README.md with Stable Download URLs"
-          git -c http.extraheader="AUTHORIZATION: bearer ${{ secrets.PAT_SERVICE_ACCOUNT }}" push origin HEAD:main
+          git -c http.extraheader="AUTHORIZATION: bearer ${{ secrets.PAT_SERVICE_ACCOUNT }}" push origin HEAD:dev

.gitignore

@@ -12,6 +12,8 @@ build
 electron/renderer
 electron/models
 electron/docs
+electron/engines
+server/pre-install
 package-lock.json
 
 *.log
@@ -26,3 +28,4 @@ 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
+

README.md

@@ -43,31 +43,31 @@ Jan is an open-source ChatGPT alternative that runs 100% offline on your compute
   <tr style="text-align:center">
     <td style="text-align:center"><b>Stable (Recommended)</b></td>
     <td style="text-align:center">
-      <a href='https://github.com/janhq/jan/releases/download/v0.4.4/jan-win-x64-0.4.4.exe'>
+      <a href='https://github.com/janhq/jan/releases/download/v0.4.6/jan-win-x64-0.4.6.exe'>
         <img src='./docs/static/img/windows.png' style="height:14px; width: 14px" />
         <b>jan.exe</b>
       </a>
     </td>
     <td style="text-align:center">
-      <a href='https://github.com/janhq/jan/releases/download/v0.4.4/jan-mac-x64-0.4.4.dmg'>
+      <a href='https://github.com/janhq/jan/releases/download/v0.4.6/jan-mac-x64-0.4.6.dmg'>
         <img src='./docs/static/img/mac.png' style="height:15px; width: 15px" />
         <b>Intel</b>
       </a>
     </td>
     <td style="text-align:center">
-      <a href='https://github.com/janhq/jan/releases/download/v0.4.4/jan-mac-arm64-0.4.4.dmg'>
+      <a href='https://github.com/janhq/jan/releases/download/v0.4.6/jan-mac-arm64-0.4.6.dmg'>
         <img src='./docs/static/img/mac.png' style="height:15px; width: 15px" />
         <b>M1/M2</b>
       </a>
     </td>
     <td style="text-align:center">
-      <a href='https://github.com/janhq/jan/releases/download/v0.4.4/jan-linux-amd64-0.4.4.deb'>
+      <a href='https://github.com/janhq/jan/releases/download/v0.4.6/jan-linux-amd64-0.4.6.deb'>
         <img src='./docs/static/img/linux.png' style="height:14px; width: 14px" />
         <b>jan.deb</b>
       </a>
     </td>
     <td style="text-align:center">
-      <a href='https://github.com/janhq/jan/releases/download/v0.4.4/jan-linux-x86_64-0.4.4.AppImage'>
+      <a href='https://github.com/janhq/jan/releases/download/v0.4.6/jan-linux-x86_64-0.4.6.AppImage'>
         <img src='./docs/static/img/linux.png' style="height:14px; width: 14px" />
         <b>jan.AppImage</b>
       </a>
@@ -76,31 +76,31 @@ Jan is an open-source ChatGPT alternative that runs 100% offline on your compute
   <tr style="text-align:center">
     <td style="text-align:center"><b>Experimental (Nightly Build)</b></td>
     <td style="text-align:center">
-      <a href='https://delta.jan.ai/latest/jan-win-x64-0.4.4-180.exe'>
+      <a href='https://delta.jan.ai/latest/jan-win-x64-0.4.6-263.exe'>
         <img src='./docs/static/img/windows.png' style="height:14px; width: 14px" />
         <b>jan.exe</b>
       </a>
     </td>
     <td style="text-align:center">
-      <a href='https://delta.jan.ai/latest/jan-mac-x64-0.4.4-180.dmg'>
+      <a href='https://delta.jan.ai/latest/jan-mac-x64-0.4.6-263.dmg'>
         <img src='./docs/static/img/mac.png' style="height:15px; width: 15px" />
         <b>Intel</b>
       </a>
     </td>
     <td style="text-align:center">
-      <a href='https://delta.jan.ai/latest/jan-mac-arm64-0.4.4-180.dmg'>
+      <a href='https://delta.jan.ai/latest/jan-mac-arm64-0.4.6-263.dmg'>
         <img src='./docs/static/img/mac.png' style="height:15px; width: 15px" />
         <b>M1/M2</b>
       </a>
     </td>
     <td style="text-align:center">
-      <a href='https://delta.jan.ai/latest/jan-linux-amd64-0.4.4-180.deb'>
+      <a href='https://delta.jan.ai/latest/jan-linux-amd64-0.4.6-263.deb'>
         <img src='./docs/static/img/linux.png' style="height:14px; width: 14px" />
         <b>jan.deb</b>
       </a>
     </td>
     <td style="text-align:center">
-      <a href='https://delta.jan.ai/latest/jan-linux-x86_64-0.4.4-180.AppImage'>
+      <a href='https://delta.jan.ai/latest/jan-linux-x86_64-0.4.6-263.AppImage'>
         <img src='./docs/static/img/linux.png' style="height:14px; width: 14px" />
         <b>jan.AppImage</b>
       </a>

core/src/api/index.ts

@@ -3,7 +3,6 @@
  * @description Enum of all the routes exposed by the app
  */
 export enum AppRoute {
-  appDataPath = 'appDataPath',
   openExternalUrl = 'openExternalUrl',
   openAppDirectory = 'openAppDirectory',
   openFileExplore = 'openFileExplorer',
@@ -12,6 +11,7 @@ export enum AppRoute {
   updateAppConfiguration = 'updateAppConfiguration',
   relaunch = 'relaunch',
   joinPath = 'joinPath',
+  isSubdirectory = 'isSubdirectory',
   baseName = 'baseName',
   startServer = 'startServer',
   stopServer = 'stopServer',
@@ -61,7 +61,9 @@ export enum FileManagerRoute {
   syncFile = 'syncFile',
   getJanDataFolderPath = 'getJanDataFolderPath',
   getResourcePath = 'getResourcePath',
+  getUserHomePath = 'getUserHomePath',
   fileStat = 'fileStat',
+  writeBlob = 'writeBlob',
 }
 
 export type ApiFunction = (...args: any[]) => any

core/src/core.ts

@@ -22,7 +22,11 @@ const executeOnMain: (extension: string, method: string, ...args: any[]) => Prom
  * @param {object} network - Optional object to specify proxy/whether to ignore SSL certificates.
  * @returns {Promise<any>} A promise that resolves when the file is downloaded.
  */
-const downloadFile: (url: string, fileName: string, network?: { proxy?: string, ignoreSSL?: boolean }) => Promise<any> = (url, fileName, network) => {
+const downloadFile: (
+  url: string,
+  fileName: string,
+  network?: { proxy?: string; ignoreSSL?: boolean }
+) => Promise<any> = (url, fileName, network) => {
   return global.core?.api?.downloadFile(url, fileName, network)
 }
 
@@ -79,6 +83,12 @@ const openExternalUrl: (url: string) => Promise<any> = (url) =>
  */
 const getResourcePath: () => Promise<string> = () => global.core.api?.getResourcePath()
 
+/**
+ * Gets the user's home path.
+ * @returns return user's home path
+ */
+const getUserHomePath = (): Promise<string> => global.core.api?.getUserHomePath()
+
 /**
  * Log to file from browser processes.
  *
@@ -87,6 +97,17 @@ const getResourcePath: () => Promise<string> = () => global.core.api?.getResourc
 const log: (message: string, fileName?: string) => void = (message, fileName) =>
   global.core.api?.log(message, fileName)
 
+/**
+ * Check whether the path is a subdirectory of another path.
+ *
+ * @param from - The path to check.
+ * @param to - The path to check against.
+ *
+ * @returns {Promise<boolean>} - A promise that resolves with a boolean indicating whether the path is a subdirectory.
+ */
+const isSubdirectory: (from: string, to: string) => Promise<boolean> = (from: string, to: string) =>
+  global.core.api?.isSubdirectory(from, to)
+
 /**
  * Register extension point function type definition
  */
@@ -94,7 +115,7 @@ export type RegisterExtensionPoint = (
   extensionName: string,
   extensionId: string,
   method: Function,
-  priority?: number,
+  priority?: number
 ) => void
 
 /**
@@ -111,5 +132,7 @@ export {
   openExternalUrl,
   baseName,
   log,
+  isSubdirectory,
+  getUserHomePath,
   FileStat,
 }

core/src/fs.ts

@@ -1,4 +1,4 @@
-import { FileStat } from "./types"
+import { FileStat } from './types'
 
 /**
  * Writes data to a file at the specified path.
@@ -6,6 +6,15 @@ import { FileStat } from "./types"
  */
 const writeFileSync = (...args: any[]) => global.core.api?.writeFileSync(...args)
 
+/**
+ * Writes blob data to a file at the specified path.
+ * @param path - The path to file.
+ * @param data - The blob data.
+ * @returns
+ */
+const writeBlob: (path: string, data: string) => Promise<any> = (path, data) =>
+  global.core.api?.writeBlob(path, data)
+
 /**
  * Reads the contents of a file at the specified path.
  * @returns {Promise<any>} A Promise that resolves with the contents of the file.
@@ -60,7 +69,6 @@ const syncFile: (src: string, dest: string) => Promise<any> = (src, dest) =>
  */
 const copyFileSync = (...args: any[]) => global.core.api?.copyFileSync(...args)
 
-
 /**
  * Gets the file's stats.
  *
@@ -70,7 +78,6 @@ const copyFileSync = (...args: any[]) => global.core.api?.copyFileSync(...args)
 const fileStat: (path: string) => Promise<FileStat | undefined> = (path) =>
   global.core.api?.fileStat(path)
 
-
 // TODO: Export `dummy` fs functions automatically
 // Currently adding these manually
 export const fs = {
@@ -84,5 +91,6 @@ export const fs = {
   appendFileSync,
   copyFileSync,
   syncFile,
-  fileStat
+  fileStat,
+  writeBlob,
 }

core/src/node/api/common/builder.ts

@@ -2,7 +2,8 @@ import fs from 'fs'
 import { JanApiRouteConfiguration, RouteConfiguration } from './configuration'
 import { join } from 'path'
 import { ContentType, MessageStatus, Model, ThreadMessage } from './../../../index'
-import { getJanDataFolderPath } from '../../utils'
+import { getEngineConfiguration, getJanDataFolderPath } from '../../utils'
+import { DEFAULT_CHAT_COMPLETION_URL } from './consts'
 
 export const getBuilder = async (configuration: RouteConfiguration) => {
   const directoryPath = join(getJanDataFolderPath(), configuration.dirName)
@@ -309,7 +310,7 @@ export const chatCompletions = async (request: any, reply: any) => {
   const engineConfiguration = await getEngineConfiguration(requestedModel.engine)
 
   let apiKey: string | undefined = undefined
-  let apiUrl: string = 'http://127.0.0.1:3928/inferences/llamacpp/chat_completion' // default nitro url
+  let apiUrl: string = DEFAULT_CHAT_COMPLETION_URL
 
   if (engineConfiguration) {
     apiKey = engineConfiguration.api_key
@@ -320,7 +321,7 @@ export const chatCompletions = async (request: any, reply: any) => {
     'Content-Type': 'text/event-stream',
     'Cache-Control': 'no-cache',
     'Connection': 'keep-alive',
-    "Access-Control-Allow-Origin": "*"
+    'Access-Control-Allow-Origin': '*',
   })
 
   const headers: Record<string, any> = {
@@ -346,13 +347,3 @@ export const chatCompletions = async (request: any, reply: any) => {
     response.body.pipe(reply.raw)
   }
 }
-
-const getEngineConfiguration = async (engineId: string) => {
-  if (engineId !== 'openai') {
-    return undefined
-  }
-  const directoryPath = join(getJanDataFolderPath(), 'engines')
-  const filePath = join(directoryPath, `${engineId}.json`)
-  const data = await fs.readFileSync(filePath, 'utf-8')
-  return JSON.parse(data)
-}

core/src/node/api/common/consts.ts

@@ -0,0 +1,19 @@
+// The PORT to use for the Nitro subprocess
+export const NITRO_DEFAULT_PORT = 3928
+
+// The HOST address to use for the Nitro subprocess
+export const LOCAL_HOST = '127.0.0.1'
+
+export const SUPPORTED_MODEL_FORMAT = '.gguf'
+
+// The URL for the Nitro subprocess
+const NITRO_HTTP_SERVER_URL = `http://${LOCAL_HOST}:${NITRO_DEFAULT_PORT}`
+// The URL for the Nitro subprocess to load a model
+export const NITRO_HTTP_LOAD_MODEL_URL = `${NITRO_HTTP_SERVER_URL}/inferences/llamacpp/loadmodel`
+// The URL for the Nitro subprocess to validate a model
+export const NITRO_HTTP_VALIDATE_MODEL_URL = `${NITRO_HTTP_SERVER_URL}/inferences/llamacpp/modelstatus`
+
+// The URL for the Nitro subprocess to kill itself
+export const NITRO_HTTP_KILL_URL = `${NITRO_HTTP_SERVER_URL}/processmanager/destroy`
+
+export const DEFAULT_CHAT_COMPLETION_URL = `http://${LOCAL_HOST}:${NITRO_DEFAULT_PORT}/inferences/llamacpp/chat_completion` // default nitro url

core/src/node/api/common/startStopModel.ts

@@ -0,0 +1,351 @@
+import fs from 'fs'
+import { join } from 'path'
+import { getJanDataFolderPath, getJanExtensionsPath, getSystemResourceInfo } from '../../utils'
+import { logServer } from '../../log'
+import { ChildProcessWithoutNullStreams, spawn } from 'child_process'
+import { Model, ModelSettingParams, PromptTemplate } from '../../../types'
+import {
+  LOCAL_HOST,
+  NITRO_DEFAULT_PORT,
+  NITRO_HTTP_KILL_URL,
+  NITRO_HTTP_LOAD_MODEL_URL,
+  NITRO_HTTP_VALIDATE_MODEL_URL,
+  SUPPORTED_MODEL_FORMAT,
+} from './consts'
+
+// The subprocess instance for Nitro
+let subprocess: ChildProcessWithoutNullStreams | undefined = undefined
+
+// TODO: move this to core type
+interface NitroModelSettings extends ModelSettingParams {
+  llama_model_path: string
+  cpu_threads: number
+}
+
+export const startModel = async (modelId: string, settingParams?: ModelSettingParams) => {
+  try {
+    await runModel(modelId, settingParams)
+
+    return {
+      message: `Model ${modelId} started`,
+    }
+  } catch (e) {
+    return {
+      error: e,
+    }
+  }
+}
+
+const runModel = async (modelId: string, settingParams?: ModelSettingParams): Promise<void> => {
+  const janDataFolderPath = getJanDataFolderPath()
+  const modelFolderFullPath = join(janDataFolderPath, 'models', modelId)
+
+  if (!fs.existsSync(modelFolderFullPath)) {
+    throw `Model not found: ${modelId}`
+  }
+
+  const files: string[] = fs.readdirSync(modelFolderFullPath)
+
+  // Look for GGUF model file
+  const ggufBinFile = files.find((file) => file.toLowerCase().includes(SUPPORTED_MODEL_FORMAT))
+
+  const modelMetadataPath = join(modelFolderFullPath, 'model.json')
+  const modelMetadata: Model = JSON.parse(fs.readFileSync(modelMetadataPath, 'utf-8'))
+
+  if (!ggufBinFile) {
+    throw 'No GGUF model file found'
+  }
+  const modelBinaryPath = join(modelFolderFullPath, ggufBinFile)
+
+  const nitroResourceProbe = await getSystemResourceInfo()
+  const nitroModelSettings: NitroModelSettings = {
+    ...modelMetadata.settings,
+    ...settingParams,
+    llama_model_path: modelBinaryPath,
+    // This is critical and requires real CPU physical core count (or performance core)
+    cpu_threads: Math.max(1, nitroResourceProbe.numCpuPhysicalCore),
+    ...(modelMetadata.settings.mmproj && {
+      mmproj: join(modelFolderFullPath, modelMetadata.settings.mmproj),
+    }),
+  }
+
+  logServer(`[NITRO]::Debug: Nitro model settings: ${JSON.stringify(nitroModelSettings)}`)
+
+  // Convert settings.prompt_template to system_prompt, user_prompt, ai_prompt
+  if (modelMetadata.settings.prompt_template) {
+    const promptTemplate = modelMetadata.settings.prompt_template
+    const prompt = promptTemplateConverter(promptTemplate)
+    if (prompt?.error) {
+      return Promise.reject(prompt.error)
+    }
+    nitroModelSettings.system_prompt = prompt.system_prompt
+    nitroModelSettings.user_prompt = prompt.user_prompt
+    nitroModelSettings.ai_prompt = prompt.ai_prompt
+  }
+
+  await runNitroAndLoadModel(modelId, nitroModelSettings)
+}
+
+// TODO: move to util
+const promptTemplateConverter = (promptTemplate: string): PromptTemplate => {
+  // Split the string using the markers
+  const systemMarker = '{system_message}'
+  const promptMarker = '{prompt}'
+
+  if (promptTemplate.includes(systemMarker) && promptTemplate.includes(promptMarker)) {
+    // Find the indices of the markers
+    const systemIndex = promptTemplate.indexOf(systemMarker)
+    const promptIndex = promptTemplate.indexOf(promptMarker)
+
+    // Extract the parts of the string
+    const system_prompt = promptTemplate.substring(0, systemIndex)
+    const user_prompt = promptTemplate.substring(systemIndex + systemMarker.length, promptIndex)
+    const ai_prompt = promptTemplate.substring(promptIndex + promptMarker.length)
+
+    // Return the split parts
+    return { system_prompt, user_prompt, ai_prompt }
+  } else if (promptTemplate.includes(promptMarker)) {
+    // Extract the parts of the string for the case where only promptMarker is present
+    const promptIndex = promptTemplate.indexOf(promptMarker)
+    const user_prompt = promptTemplate.substring(0, promptIndex)
+    const ai_prompt = promptTemplate.substring(promptIndex + promptMarker.length)
+
+    // Return the split parts
+    return { user_prompt, ai_prompt }
+  }
+
+  // Return an error if none of the conditions are met
+  return { error: 'Cannot split prompt template' }
+}
+
+const runNitroAndLoadModel = async (modelId: string, modelSettings: NitroModelSettings) => {
+  // Gather system information for CPU physical cores and memory
+  const tcpPortUsed = require('tcp-port-used')
+
+  await stopModel(modelId)
+  await tcpPortUsed.waitUntilFree(NITRO_DEFAULT_PORT, 300, 5000)
+
+  /**
+   * There is a problem with Windows process manager
+   * Should wait for awhile to make sure the port is free and subprocess is killed
+   * The tested threshold is 500ms
+   **/
+  if (process.platform === 'win32') {
+    await new Promise((resolve) => setTimeout(resolve, 500))
+  }
+
+  await spawnNitroProcess()
+  await loadLLMModel(modelSettings)
+  await validateModelStatus()
+}
+
+const spawnNitroProcess = async (): Promise<void> => {
+  logServer(`[NITRO]::Debug: Spawning Nitro subprocess...`)
+
+  let binaryFolder = join(
+    getJanExtensionsPath(),
+    '@janhq',
+    'inference-nitro-extension',
+    'dist',
+    'bin'
+  )
+
+  let executableOptions = executableNitroFile()
+  const tcpPortUsed = require('tcp-port-used')
+
+  const args: string[] = ['1', LOCAL_HOST, NITRO_DEFAULT_PORT.toString()]
+  // Execute the binary
+  logServer(
+    `[NITRO]::Debug: Spawn nitro at path: ${executableOptions.executablePath}, and args: ${args}`
+  )
+  subprocess = spawn(
+    executableOptions.executablePath,
+    ['1', LOCAL_HOST, NITRO_DEFAULT_PORT.toString()],
+    {
+      cwd: binaryFolder,
+      env: {
+        ...process.env,
+        CUDA_VISIBLE_DEVICES: executableOptions.cudaVisibleDevices,
+      },
+    }
+  )
+
+  // Handle subprocess output
+  subprocess.stdout.on('data', (data: any) => {
+    logServer(`[NITRO]::Debug: ${data}`)
+  })
+
+  subprocess.stderr.on('data', (data: any) => {
+    logServer(`[NITRO]::Error: ${data}`)
+  })
+
+  subprocess.on('close', (code: any) => {
+    logServer(`[NITRO]::Debug: Nitro exited with code: ${code}`)
+    subprocess = undefined
+  })
+
+  tcpPortUsed.waitUntilUsed(NITRO_DEFAULT_PORT, 300, 30000).then(() => {
+    logServer(`[NITRO]::Debug: Nitro is ready`)
+  })
+}
+
+type NitroExecutableOptions = {
+  executablePath: string
+  cudaVisibleDevices: string
+}
+
+const executableNitroFile = (): NitroExecutableOptions => {
+  const nvidiaInfoFilePath = join(getJanDataFolderPath(), 'settings', 'settings.json')
+  let binaryFolder = join(
+    getJanExtensionsPath(),
+    '@janhq',
+    'inference-nitro-extension',
+    'dist',
+    'bin'
+  )
+
+  let cudaVisibleDevices = ''
+  let binaryName = 'nitro'
+  /**
+   * The binary folder is different for each platform.
+   */
+  if (process.platform === 'win32') {
+    /**
+     *  For Windows: win-cpu, win-cuda-11-7, win-cuda-12-0
+     */
+    let nvidiaInfo = JSON.parse(fs.readFileSync(nvidiaInfoFilePath, 'utf-8'))
+    if (nvidiaInfo['run_mode'] === 'cpu') {
+      binaryFolder = join(binaryFolder, 'win-cpu')
+    } else {
+      if (nvidiaInfo['cuda'].version === '12') {
+        binaryFolder = join(binaryFolder, 'win-cuda-12-0')
+      } else {
+        binaryFolder = join(binaryFolder, 'win-cuda-11-7')
+      }
+      cudaVisibleDevices = nvidiaInfo['gpu_highest_vram']
+    }
+    binaryName = 'nitro.exe'
+  } else if (process.platform === 'darwin') {
+    /**
+     *  For MacOS: mac-arm64 (Silicon), mac-x64 (InteL)
+     */
+    if (process.arch === 'arm64') {
+      binaryFolder = join(binaryFolder, 'mac-arm64')
+    } else {
+      binaryFolder = join(binaryFolder, 'mac-x64')
+    }
+  } else {
+    /**
+     *  For Linux: linux-cpu, linux-cuda-11-7, linux-cuda-12-0
+     */
+    let nvidiaInfo = JSON.parse(fs.readFileSync(nvidiaInfoFilePath, 'utf-8'))
+    if (nvidiaInfo['run_mode'] === 'cpu') {
+      binaryFolder = join(binaryFolder, 'linux-cpu')
+    } else {
+      if (nvidiaInfo['cuda'].version === '12') {
+        binaryFolder = join(binaryFolder, 'linux-cuda-12-0')
+      } else {
+        binaryFolder = join(binaryFolder, 'linux-cuda-11-7')
+      }
+      cudaVisibleDevices = nvidiaInfo['gpu_highest_vram']
+    }
+  }
+
+  return {
+    executablePath: join(binaryFolder, binaryName),
+    cudaVisibleDevices,
+  }
+}
+
+const validateModelStatus = async (): Promise<void> => {
+  // Send a GET request to the validation URL.
+  // Retry the request up to 3 times if it fails, with a delay of 500 milliseconds between retries.
+  const fetchRT = require('fetch-retry')
+  const fetchRetry = fetchRT(fetch)
+
+  return fetchRetry(NITRO_HTTP_VALIDATE_MODEL_URL, {
+    method: 'GET',
+    headers: {
+      'Content-Type': 'application/json',
+    },
+    retries: 5,
+    retryDelay: 500,
+  }).then(async (res: Response) => {
+    logServer(`[NITRO]::Debug: Validate model state success with response ${JSON.stringify(res)}`)
+    // If the response is OK, check model_loaded status.
+    if (res.ok) {
+      const body = await res.json()
+      // If the model is loaded, return an empty object.
+      // Otherwise, return an object with an error message.
+      if (body.model_loaded) {
+        return Promise.resolve()
+      }
+    }
+    return Promise.reject('Validate model status failed')
+  })
+}
+
+const loadLLMModel = async (settings: NitroModelSettings): Promise<Response> => {
+  logServer(`[NITRO]::Debug: Loading model with params ${JSON.stringify(settings)}`)
+  const fetchRT = require('fetch-retry')
+  const fetchRetry = fetchRT(fetch)
+
+  return fetchRetry(NITRO_HTTP_LOAD_MODEL_URL, {
+    method: 'POST',
+    headers: {
+      'Content-Type': 'application/json',
+    },
+    body: JSON.stringify(settings),
+    retries: 3,
+    retryDelay: 500,
+  })
+    .then((res: any) => {
+      logServer(`[NITRO]::Debug: Load model success with response ${JSON.stringify(res)}`)
+      return Promise.resolve(res)
+    })
+    .catch((err: any) => {
+      logServer(`[NITRO]::Error: Load model failed with error ${err}`)
+      return Promise.reject(err)
+    })
+}
+
+/**
+ * Stop model and kill nitro process.
+ */
+export const stopModel = async (_modelId: string) => {
+  if (!subprocess) {
+    return {
+      error: "Model isn't running",
+    }
+  }
+  return new Promise((resolve, reject) => {
+    const controller = new AbortController()
+    setTimeout(() => {
+      controller.abort()
+      reject({
+        error: 'Failed to stop model: Timedout',
+      })
+    }, 5000)
+    const tcpPortUsed = require('tcp-port-used')
+    logServer(`[NITRO]::Debug: Request to kill Nitro`)
+
+    fetch(NITRO_HTTP_KILL_URL, {
+      method: 'DELETE',
+      signal: controller.signal,
+    })
+      .then(() => {
+        subprocess?.kill()
+        subprocess = undefined
+      })
+      .catch(() => {
+        // don't need to do anything, we still kill the subprocess
+      })
+      .then(() => tcpPortUsed.waitUntilFree(NITRO_DEFAULT_PORT, 300, 5000))
+      .then(() => logServer(`[NITRO]::Debug: Nitro process is terminated`))
+      .then(() =>
+        resolve({
+          message: 'Model stopped',
+        })
+      )
+  })
+}

core/src/node/api/routes/common.ts

@@ -10,6 +10,10 @@ import {
 } from '../common/builder'
 
 import { JanApiRouteConfiguration } from '../common/configuration'
+import { startModel, stopModel } from '../common/startStopModel'
+import { ModelSettingParams } from '../../../types'
+import { getJanDataFolderPath } from '../../utils'
+import { normalizeFilePath } from '../../path'
 
 export const commonRouter = async (app: HttpServer) => {
   // Common Routes
@@ -17,26 +21,47 @@ export const commonRouter = async (app: HttpServer) => {
     app.get(`/${key}`, async (_request) => getBuilder(JanApiRouteConfiguration[key]))
 
     app.get(`/${key}/:id`, async (request: any) =>
-      retrieveBuilder(JanApiRouteConfiguration[key], request.params.id),
+      retrieveBuilder(JanApiRouteConfiguration[key], request.params.id)
     )
 
     app.delete(`/${key}/:id`, async (request: any) =>
-      deleteBuilder(JanApiRouteConfiguration[key], request.params.id),
+      deleteBuilder(JanApiRouteConfiguration[key], request.params.id)
     )
   })
 
   // Download Model Routes
   app.get(`/models/download/:modelId`, async (request: any) =>
-    downloadModel(request.params.modelId, { ignoreSSL: request.query.ignoreSSL === 'true', proxy: request.query.proxy }),
+    downloadModel(request.params.modelId, {
+      ignoreSSL: request.query.ignoreSSL === 'true',
+      proxy: request.query.proxy,
+    })
   )
 
+  app.put(`/models/:modelId/start`, async (request: any) => {
+    let settingParams: ModelSettingParams | undefined = undefined
+    if (Object.keys(request.body).length !== 0) {
+      settingParams = JSON.parse(request.body) as ModelSettingParams
+    }
+
+    return startModel(request.params.modelId, settingParams)
+  })
+
+  app.put(`/models/:modelId/stop`, async (request: any) => stopModel(request.params.modelId))
+
   // Chat Completion Routes
   app.post(`/chat/completions`, async (request: any, reply: any) => chatCompletions(request, reply))
 
   // App Routes
   app.post(`/app/${AppRoute.joinPath}`, async (request: any, reply: any) => {
     const args = JSON.parse(request.body) as any[]
-    reply.send(JSON.stringify(join(...args[0])))
+
+    const paths = args[0].map((arg: string) =>
+      typeof arg === 'string' && (arg.startsWith(`file:/`) || arg.startsWith(`file:\\`))
+        ? join(getJanDataFolderPath(), normalizeFilePath(arg))
+        : arg
+    )
+
+    reply.send(JSON.stringify(join(...paths)))
   })
 
   app.post(`/app/${AppRoute.baseName}`, async (request: any, reply: any) => {

core/src/node/api/routes/download.ts

@@ -4,55 +4,55 @@ import { DownloadManager } from '../../download'
 import { HttpServer } from '../HttpServer'
 import { createWriteStream } from 'fs'
 import { getJanDataFolderPath } from '../../utils'
-import { normalizeFilePath } from "../../path";
+import { normalizeFilePath } from '../../path'
 
 export const downloadRouter = async (app: HttpServer) => {
   app.post(`/${DownloadRoute.downloadFile}`, async (req, res) => {
-    const strictSSL = !(req.query.ignoreSSL === "true");
+    const strictSSL = !(req.query.ignoreSSL === 'true')
-    const proxy = req.query.proxy?.startsWith("http") ? req.query.proxy : undefined;
+    const proxy = req.query.proxy?.startsWith('http') ? req.query.proxy : undefined
-    const body = JSON.parse(req.body as any);
+    const body = JSON.parse(req.body as any)
     const normalizedArgs = body.map((arg: any) => {
-      if (typeof arg === "string") {
+      if (typeof arg === 'string' && arg.startsWith('file:')) {
-        return join(getJanDataFolderPath(), normalizeFilePath(arg));
+        return join(getJanDataFolderPath(), normalizeFilePath(arg))
       }
-      return arg;
+      return arg
-    });
+    })
 
-    const localPath = normalizedArgs[1];
+    const localPath = normalizedArgs[1]
-    const fileName = localPath.split("/").pop() ?? "";
+    const fileName = localPath.split('/').pop() ?? ''
 
-    const request = require("request");
+    const request = require('request')
-    const progress = require("request-progress");
+    const progress = require('request-progress')
 
-    const rq = request({ url: normalizedArgs[0], strictSSL, proxy });
+    const rq = request({ url: normalizedArgs[0], strictSSL, proxy })
     progress(rq, {})
-      .on("progress", function (state: any) {
+      .on('progress', function (state: any) {
-        console.log("download onProgress", state);
+        console.log('download onProgress', state)
       })
-      .on("error", function (err: Error) {
+      .on('error', function (err: Error) {
-        console.log("download onError", err);
+        console.log('download onError', err)
       })
-      .on("end", function () {
+      .on('end', function () {
-        console.log("download onEnd");
+        console.log('download onEnd')
       })
-      .pipe(createWriteStream(normalizedArgs[1]));
+      .pipe(createWriteStream(normalizedArgs[1]))
 
-    DownloadManager.instance.setRequest(fileName, rq);
+    DownloadManager.instance.setRequest(fileName, rq)
-  });
+  })
 
   app.post(`/${DownloadRoute.abortDownload}`, async (req, res) => {
-    const body = JSON.parse(req.body as any);
+    const body = JSON.parse(req.body as any)
     const normalizedArgs = body.map((arg: any) => {
-      if (typeof arg === "string") {
+      if (typeof arg === 'string' && arg.startsWith('file:')) {
-        return join(getJanDataFolderPath(), normalizeFilePath(arg));
+        return join(getJanDataFolderPath(), normalizeFilePath(arg))
       }
-      return arg;
+      return arg
-    });
+    })
 
-    const localPath = normalizedArgs[0];
+    const localPath = normalizedArgs[0]
-    const fileName = localPath.split("/").pop() ?? "";
+    const fileName = localPath.split('/').pop() ?? ''
-    const rq = DownloadManager.instance.networkRequests[fileName];
+    const rq = DownloadManager.instance.networkRequests[fileName]
-    DownloadManager.instance.networkRequests[fileName] = undefined;
+    DownloadManager.instance.networkRequests[fileName] = undefined
-    rq?.abort();
+    rq?.abort()
-  });
+  })
-};
+}

core/src/node/api/routes/fileManager.ts

@@ -8,5 +8,7 @@ export const fsRouter = async (app: HttpServer) => {
 
   app.post(`/app/${FileManagerRoute.getResourcePath}`, async (request: any, reply: any) => {})
 
+  app.post(`/app/${FileManagerRoute.getUserHomePath}`, async (request: any, reply: any) => {})
+
   app.post(`/app/${FileManagerRoute.fileStat}`, async (request: any, reply: any) => {})
 }

core/src/node/api/routes/fs.ts

@@ -2,6 +2,7 @@ import { FileSystemRoute } from '../../../api'
 import { join } from 'path'
 import { HttpServer } from '../HttpServer'
 import { getJanDataFolderPath } from '../../utils'
+import { normalizeFilePath } from '../../path'
 
 export const fsRouter = async (app: HttpServer) => {
   const moduleName = 'fs'
@@ -13,10 +14,10 @@ export const fsRouter = async (app: HttpServer) => {
         const result = await import(moduleName).then((mdl) => {
           return mdl[route](
             ...body.map((arg: any) =>
-              typeof arg === 'string' && arg.includes('file:/')
+              typeof arg === 'string' && (arg.startsWith(`file:/`) || arg.startsWith(`file:\\`))
-                ? join(getJanDataFolderPath(), arg.replace('file:/', ''))
+                ? join(getJanDataFolderPath(), normalizeFilePath(arg))
-                : arg,
+                : arg
-            ),
+            )
           )
         })
         res.status(200).send(result)

core/src/node/log.ts

@@ -2,38 +2,36 @@ import fs from 'fs'
 import util from 'util'
 import { getAppLogPath, getServerLogPath } from './utils'
 
-export const log = function (message: string) {
+export const log = (message: string) => {
-  const appLogPath = getAppLogPath()
+  const path = getAppLogPath()
   if (!message.startsWith('[')) {
     message = `[APP]::${message}`
   }
 
   message = `${new Date().toISOString()} ${message}`
 
-  if (fs.existsSync(appLogPath)) {
+  writeLog(message, path)
-    var log_file = fs.createWriteStream(appLogPath, {
-      flags: 'a',
-    })
-    log_file.write(util.format(message) + '\n')
-    log_file.close()
-    console.debug(message)
-  }
 }
 
-export const logServer = function (message: string) {
+export const logServer = (message: string) => {
-  const serverLogPath = getServerLogPath()
+  const path = getServerLogPath()
   if (!message.startsWith('[')) {
     message = `[SERVER]::${message}`
   }
 
   message = `${new Date().toISOString()} ${message}`
+  writeLog(message, path)
+}
 
-  if (fs.existsSync(serverLogPath)) {
+const writeLog = (message: string, logPath: string) => {
-    var log_file = fs.createWriteStream(serverLogPath, {
+  if (!fs.existsSync(logPath)) {
+    fs.writeFileSync(logPath, message)
+  } else {
+    const logFile = fs.createWriteStream(logPath, {
       flags: 'a',
     })
-    log_file.write(util.format(message) + '\n')
+    logFile.write(util.format(message) + '\n')
-    log_file.close()
+    logFile.close()
     console.debug(message)
   }
 }

core/src/node/utils/index.ts

@@ -1,16 +1,18 @@
-import { AppConfiguration } from "../../types";
+import { AppConfiguration, SystemResourceInfo } from '../../types'
-import { join } from "path";
+import { join } from 'path'
-import fs from "fs";
+import fs from 'fs'
-import os from "os";
+import os from 'os'
+import { log, logServer } from '../log'
+import childProcess from 'child_process'
 
 // TODO: move this to core
-const configurationFileName = "settings.json";
+const configurationFileName = 'settings.json'
 
 // TODO: do no specify app name in framework module
-const defaultJanDataFolder = join(os.homedir(), "jan");
+const defaultJanDataFolder = join(os.homedir(), 'jan')
 const defaultAppConfig: AppConfiguration = {
   data_folder: defaultJanDataFolder,
-};
+}
 
 /**
  * Getting App Configurations.
@@ -20,39 +22,39 @@ const defaultAppConfig: AppConfiguration = {
 export const getAppConfigurations = (): AppConfiguration => {
   // Retrieve Application Support folder path
   // Fallback to user home directory if not found
-  const configurationFile = getConfigurationFilePath();
+  const configurationFile = getConfigurationFilePath()
 
   if (!fs.existsSync(configurationFile)) {
     // create default app config if we don't have one
-    console.debug(`App config not found, creating default config at ${configurationFile}`);
+    console.debug(`App config not found, creating default config at ${configurationFile}`)
-    fs.writeFileSync(configurationFile, JSON.stringify(defaultAppConfig));
+    fs.writeFileSync(configurationFile, JSON.stringify(defaultAppConfig))
-    return defaultAppConfig;
+    return defaultAppConfig
   }
 
   try {
     const appConfigurations: AppConfiguration = JSON.parse(
-      fs.readFileSync(configurationFile, "utf-8"),
+      fs.readFileSync(configurationFile, 'utf-8')
-    );
+    )
-    return appConfigurations;
+    return appConfigurations
   } catch (err) {
-    console.error(`Failed to read app config, return default config instead! Err: ${err}`);
+    console.error(`Failed to read app config, return default config instead! Err: ${err}`)
-    return defaultAppConfig;
+    return defaultAppConfig
+  }
 }
-};
 
 const getConfigurationFilePath = () =>
   join(
-    global.core?.appPath() || process.env[process.platform == "win32" ? "USERPROFILE" : "HOME"],
+    global.core?.appPath() || process.env[process.platform == 'win32' ? 'USERPROFILE' : 'HOME'],
-    configurationFileName,
+    configurationFileName
-  );
+  )
 
 export const updateAppConfiguration = (configuration: AppConfiguration): Promise<void> => {
-  const configurationFile = getConfigurationFilePath();
+  const configurationFile = getConfigurationFilePath()
-  console.debug("updateAppConfiguration, configurationFile: ", configurationFile);
+  console.debug('updateAppConfiguration, configurationFile: ', configurationFile)
 
-  fs.writeFileSync(configurationFile, JSON.stringify(configuration));
+  fs.writeFileSync(configurationFile, JSON.stringify(configuration))
-  return Promise.resolve();
+  return Promise.resolve()
-};
+}
 
 /**
  * Utility function to get server log path
@@ -60,13 +62,13 @@ export const updateAppConfiguration = (configuration: AppConfiguration): Promise
  * @returns {string} The log path.
  */
 export const getServerLogPath = (): string => {
-  const appConfigurations = getAppConfigurations();
+  const appConfigurations = getAppConfigurations()
-  const logFolderPath = join(appConfigurations.data_folder, "logs");
+  const logFolderPath = join(appConfigurations.data_folder, 'logs')
   if (!fs.existsSync(logFolderPath)) {
-    fs.mkdirSync(logFolderPath, { recursive: true });
+    fs.mkdirSync(logFolderPath, { recursive: true })
+  }
+  return join(logFolderPath, 'server.log')
 }
-  return join(logFolderPath, "server.log");
-};
 
 /**
  * Utility function to get app log path
@@ -74,13 +76,13 @@ export const getServerLogPath = (): string => {
  * @returns {string} The log path.
  */
 export const getAppLogPath = (): string => {
-  const appConfigurations = getAppConfigurations();
+  const appConfigurations = getAppConfigurations()
-  const logFolderPath = join(appConfigurations.data_folder, "logs");
+  const logFolderPath = join(appConfigurations.data_folder, 'logs')
   if (!fs.existsSync(logFolderPath)) {
-    fs.mkdirSync(logFolderPath, { recursive: true });
+    fs.mkdirSync(logFolderPath, { recursive: true })
+  }
+  return join(logFolderPath, 'app.log')
 }
-  return join(logFolderPath, "app.log");
-};
 
 /**
  * Utility function to get data folder path
@@ -88,9 +90,9 @@ export const getAppLogPath = (): string => {
  * @returns {string} The data folder path.
  */
 export const getJanDataFolderPath = (): string => {
-  const appConfigurations = getAppConfigurations();
+  const appConfigurations = getAppConfigurations()
-  return appConfigurations.data_folder;
+  return appConfigurations.data_folder
-};
+}
 
 /**
  * Utility function to get extension path
@@ -98,6 +100,70 @@ export const getJanDataFolderPath = (): string => {
  * @returns {string} The extensions path.
  */
 export const getJanExtensionsPath = (): string => {
-  const appConfigurations = getAppConfigurations();
+  const appConfigurations = getAppConfigurations()
-  return join(appConfigurations.data_folder, "extensions");
+  return join(appConfigurations.data_folder, 'extensions')
-};
+}
+
+/**
+ * Utility function to physical cpu count
+ *
+ * @returns {number} The physical cpu count.
+ */
+export const physicalCpuCount = async (): Promise<number> => {
+  const platform = os.platform()
+  if (platform === 'linux') {
+    const output = await exec('lscpu -p | egrep -v "^#" | sort -u -t, -k 2,4 | wc -l')
+    return parseInt(output.trim(), 10)
+  } else if (platform === 'darwin') {
+    const output = await exec('sysctl -n hw.physicalcpu_max')
+    return parseInt(output.trim(), 10)
+  } else if (platform === 'win32') {
+    const output = await exec('WMIC CPU Get NumberOfCores')
+    return output
+      .split(os.EOL)
+      .map((line: string) => parseInt(line))
+      .filter((value: number) => !isNaN(value))
+      .reduce((sum: number, number: number) => sum + number, 1)
+  } else {
+    const cores = os.cpus().filter((cpu: any, index: number) => {
+      const hasHyperthreading = cpu.model.includes('Intel')
+      const isOdd = index % 2 === 1
+      return !hasHyperthreading || isOdd
+    })
+    return cores.length
+  }
+}
+
+const exec = async (command: string): Promise<string> => {
+  return new Promise((resolve, reject) => {
+    childProcess.exec(command, { encoding: 'utf8' }, (error, stdout) => {
+      if (error) {
+        reject(error)
+      } else {
+        resolve(stdout)
+      }
+    })
+  })
+}
+
+export const getSystemResourceInfo = async (): Promise<SystemResourceInfo> => {
+  const cpu = await physicalCpuCount()
+  const message = `[NITRO]::CPU informations - ${cpu}`
+  log(message)
+  logServer(message)
+
+  return {
+    numCpuPhysicalCore: cpu,
+    memAvailable: 0, // TODO: this should not be 0
+  }
+}
+
+export const getEngineConfiguration = async (engineId: string) => {
+  if (engineId !== 'openai') {
+    return undefined
+  }
+  const directoryPath = join(getJanDataFolderPath(), 'engines')
+  const filePath = join(directoryPath, `${engineId}.json`)
+  const data = fs.readFileSync(filePath, 'utf-8')
+  return JSON.parse(data)
+}

core/src/types/assistant/assistantEntity.ts

@@ -2,6 +2,13 @@
  * Assistant type defines the shape of an assistant object.
  * @stored
  */
+
+export type AssistantTool = {
+  type: string
+  enabled: boolean
+  settings: any
+}
+
 export type Assistant = {
   /** Represents the avatar of the user. */
   avatar: string
@@ -22,7 +29,7 @@ export type Assistant = {
   /** Represents the instructions for the object. */
   instructions?: string
   /** Represents the tools associated with the object. */
-  tools?: any
+  tools?: AssistantTool[]
   /** Represents the file identifiers associated with the object. */
   file_ids: string[]
   /** Represents the metadata of the object. */

core/src/types/config/appConfigEvent.ts

@@ -0,0 +1,6 @@
+/**
+ * App configuration event name
+ */
+export enum AppConfigurationEventName {
+  OnConfigurationUpdate = 'OnConfigurationUpdate',
+}

core/src/types/config/index.ts

@@ -1 +1,2 @@
 export * from './appConfigEntity'
+export * from './appConfigEvent'

core/src/types/index.ts

@@ -6,3 +6,4 @@ export * from './inference'
 export * from './monitoring'
 export * from './file'
 export * from './config'
+export * from './miscellaneous'

core/src/types/inference/inferenceEntity.ts

@@ -1,3 +1,5 @@
+import { ContentType, ContentValue } from '../message'
+
 /**
  * The role of the author of this message.
  */
@@ -13,7 +15,32 @@ export enum ChatCompletionRole {
  */
 export type ChatCompletionMessage = {
   /** The contents of the message. **/
-  content?: string
+  content?: ChatCompletionMessageContent
   /** The role of the author of this message. **/
   role: ChatCompletionRole
 }
+
+export type ChatCompletionMessageContent =
+  | string
+  | (ChatCompletionMessageContentText &
+      ChatCompletionMessageContentImage &
+      ChatCompletionMessageContentDoc)[]
+
+export enum ChatCompletionMessageContentType {
+  Text = 'text',
+  Image = 'image_url',
+  Doc = 'doc_url',
+}
+
+export type ChatCompletionMessageContentText = {
+  type: ChatCompletionMessageContentType
+  text: string
+}
+export type ChatCompletionMessageContentImage = {
+  type: ChatCompletionMessageContentType
+  image_url: { url: string }
+}
+export type ChatCompletionMessageContentDoc = {
+  type: ChatCompletionMessageContentType
+  doc_url: { url: string }
+}

core/src/types/message/messageEntity.ts

@@ -1,5 +1,6 @@
 import { ChatCompletionMessage, ChatCompletionRole } from '../inference'
 import { ModelInfo } from '../model'
+import { Thread } from '../thread'
 
 /**
  * The `ThreadMessage` type defines the shape of a thread's message object.
@@ -35,7 +36,10 @@ export type ThreadMessage = {
 export type MessageRequest = {
   id?: string
 
-  /** The thread id of the message request. **/
+  /**
+   * @deprecated Use thread object instead
+   * The thread id of the message request.
+   */
   threadId: string
 
   /**
@@ -48,6 +52,10 @@ export type MessageRequest = {
 
   /** Settings for constructing a chat completion request **/
   model?: ModelInfo
+
+  /** The thread of this message is belong to. **/
+  // TODO: deprecate threadId field
+  thread?: Thread
 }
 
 /**
@@ -62,7 +70,7 @@ export enum MessageStatus {
   /** Message loaded with error. **/
   Error = 'error',
   /** Message is cancelled streaming */
-  Stopped = "stopped"
+  Stopped = 'stopped',
 }
 
 /**
@@ -71,6 +79,7 @@ export enum MessageStatus {
 export enum ContentType {
   Text = 'text',
   Image = 'image',
+  Pdf = 'pdf',
 }
 
 /**
@@ -80,6 +89,8 @@ export enum ContentType {
 export type ContentValue = {
   value: string
   annotations: string[]
+  name?: string
+  size?: number
 }
 
 /**

core/src/types/miscellaneous/index.ts

@@ -0,0 +1,2 @@
+export * from './systemResourceInfo'
+export * from './promptTemplate'

core/src/types/miscellaneous/promptTemplate.ts

@@ -0,0 +1,6 @@
+export type PromptTemplate = {
+  system_prompt?: string
+  ai_prompt?: string
+  user_prompt?: string
+  error?: string
+}

core/src/types/miscellaneous/systemResourceInfo.ts

@@ -0,0 +1,4 @@
+export type SystemResourceInfo = {
+  numCpuPhysicalCore: number
+  memAvailable: number
+}

core/src/types/model/modelEntity.ts

@@ -7,6 +7,7 @@ export type ModelInfo = {
   settings: ModelSettingParams
   parameters: ModelRuntimeParams
   engine?: InferenceEngine
+  proxyEngine?: InferenceEngine
 }
 
 /**
@@ -18,7 +19,8 @@ export enum InferenceEngine {
   nitro = 'nitro',
   openai = 'openai',
   triton_trtllm = 'triton_trtllm',
-  hf_endpoint = 'hf_endpoint',
+
+  tool_retrieval_enabled = 'tool_retrieval_enabled',
 }
 
 export type ModelArtifact = {
@@ -90,6 +92,13 @@ export type Model = {
    * The model engine.
    */
   engine: InferenceEngine
+
+  proxyEngine?: InferenceEngine
+
+  /**
+   * Is multimodal or not.
+   */
+  visionModel?: boolean
 }
 
 export type ModelMetadata = {
@@ -114,6 +123,7 @@ export type ModelSettingParams = {
   user_prompt?: string
   llama_model_path?: string
   mmproj?: string
+  cont_batching?: boolean
 }
 
 /**
@@ -129,4 +139,5 @@ export type ModelRuntimeParams = {
   stop?: string[]
   frequency_penalty?: number
   presence_penalty?: number
+  engine?: string
 }

core/src/types/thread/index.ts

@@ -1,2 +1,3 @@
 export * from './threadEntity'
 export * from './threadInterface'
+export * from './threadEvent'

core/src/types/thread/threadEntity.ts

@@ -1,3 +1,4 @@
+import { AssistantTool } from '../assistant'
 import { ModelInfo } from '../model'
 
 /**
@@ -30,6 +31,7 @@ export type ThreadAssistantInfo = {
   assistant_name: string
   model: ModelInfo
   instructions?: string
+  tools?: AssistantTool[]
 }
 
 /**
@@ -41,5 +43,4 @@ export type ThreadState = {
   waitingForResponse: boolean
   error?: Error
   lastMessage?: string
-  isFinishInit?: boolean
 }

core/src/types/thread/threadEvent.ts

@@ -0,0 +1,4 @@
+export enum ThreadEvent {
+  /** The `OnThreadStarted` event is emitted when a thread is started. */
+  OnThreadStarted = 'OnThreadStarted',
+}

docs/.env.example

@@ -1,5 +1,5 @@
 GTM_ID=xxxx
-POSTHOG_PROJECT_API_KEY=xxxx
+UMAMI_PROJECT_API_KEY=xxxx
-POSTHOG_APP_URL=xxxx
+UMAMI_APP_URL=xxxx
 ALGOLIA_API_KEY=xxxx
 ALGOLIA_APP_ID=xxxx

docs/docs/template/QA_script.md

@@ -1,6 +1,6 @@
 # [Release Version] QA Script 
 
-**Release Version:**
+**Release Version:** v0.4.6
 
 **Operating System:**
 
@@ -25,10 +25,10 @@
 
 ### 3. Users uninstall app
 
-- [ ] :key: Check that the uninstallation process removes all components of the app from the system.
+- [ ] :key::warning: Check that the uninstallation process removes the app successfully from the system.
 - [ ] Clean the Jan root directory and open the app to check if it creates all the necessary folders, especially models and extensions.
 - [ ] When updating the app, check if the `/models` directory has any JSON files that change according to the update.
-- [ ] Verify if updating the app also updates extensions correctly (test functionality changes; support notifications for necessary tests with each version related to extensions update).
+- [ ] Verify if updating the app also updates extensions correctly (test functionality changes, support notifications for necessary tests with each version related to extensions update).
 
 ### 4. Users close app
 
@@ -60,49 +60,45 @@
 - [ ] :key: Ensure that the conversation thread is maintained without any loss of data upon sending multiple messages.
 - [ ] Test for the ability to send different types of messages (e.g., text, emojis, code blocks).
 - [ ] :key: Validate the scroll functionality in the chat window for lengthy conversations.
-- [ ] Check if the user can renew responses multiple times.
 - [ ] Check if the user can copy the response.
 - [ ] Check if the user can delete responses.
-- [ ] :warning: Test if the user deletes the message midway, then the assistant stops that response.
 - [ ] :key: Check the `clear message` button works.
 - [ ] :key: Check the `delete entire chat` works.
-- [ ] :warning: Check if deleting all the chat retains the system prompt.
+- [ ] Check if deleting all the chat retains the system prompt.
 - [ ] Check the output format of the AI (code blocks, JSON, markdown, ...).
 - [ ] :key: Validate that there is appropriate error handling and messaging if the assistant fails to respond.
 - [ ] Test assistant's ability to maintain context over multiple exchanges.
 - [ ] :key: Check the `create new chat` button works correctly
 - [ ] Confirm that by changing `models` mid-thread the app can still handle it.
-- [ ] Check that by changing `instructions` mid-thread the app can still handle it.
+- [ ] Check the `regenerate` button renews the response (single / multiple times).
-- [ ] Check the `regenerate` button renews the response.
+- [ ] Check the `Instructions` update correctly after the user updates it midway (mid-thread).
-- [ ] Check the `Instructions` update correctly after the user updates it midway.
 
 ### 2. Users can customize chat settings like model parameters via both the GUI & thread.json
 
-- [ ] :key: Confirm that the chat settings options are accessible via the GUI.
+- [ ] :key: Confirm that the Threads settings options are accessible.
 - [ ] Test the functionality to adjust model parameters (e.g., Temperature, Top K, Top P) from the GUI and verify they are reflected in the chat behavior.
 - [ ] :key: Ensure that changes can be saved and persisted between sessions.
 - [ ] Validate that users can access and modify the thread.json file.
 - [ ] :key: Check that changes made in thread.json are correctly applied to the chat session upon reload or restart.
-- [ ] Verify if there is a revert option to go back to previous settings after changes are made.
-- [ ] Test for user feedback or confirmation after saving changes to settings.
 - [ ] Check the maximum and minimum limits of the adjustable parameters and how they affect the assistant's responses.
 - [ ] :key: Validate user permissions for those who can change settings and persist them.
 - [ ] :key: Ensure that users switch between threads with different models, the app can handle it.
 
-### 3. Users can click on a history thread
+### 3. Model dropdown
+- [ ] :key: Model list should highlight recommended based on user RAM
+- [ ] Model size should display (for both installed and imported models)
 
+### 4. Users can click on a history thread
 - [ ] Test the ability to click on any thread in the history panel.
 - [ ] :key: Verify that clicking a thread brings up the past conversation in the main chat window.
 - [ ] :key: Ensure that the selected thread is highlighted or otherwise indicated in the history panel.
 - [ ] Confirm that the chat window displays the entire conversation from the selected history thread without any missing messages.
 - [ ] :key: Check the performance and accuracy of the history feature when dealing with a large number of threads.
 - [ ] Validate that historical threads reflect the exact state of the chat at that time, including settings.
-- [ ] :key: :warning: Test the search functionality within the history panel for quick navigation.
 - [ ] :key: Verify the ability to delete or clean old threads.
 - [ ] :key: Confirm that changing the title of the thread updates correctly.
 
-### 4. Users can config instructions for the assistant.
+### 5. Users can config instructions for the assistant.
-
 - [ ] Ensure there is a clear interface to input or change instructions for the assistant.
 - [ ] Test if the instructions set by the user are being followed by the assistant in subsequent conversations.
 - [ ] :key: Validate that changes to instructions are updated in real time and do not require a restart of the application or session.
@@ -112,6 +108,8 @@
 - [ ] Validate that instructions can be saved with descriptive names for easy retrieval.
 - [ ] :key: Check if the assistant can handle conflicting instructions and how it resolves them.
 - [ ] Ensure that instruction configurations are documented for user reference.
+- [ ] :key: RAG - Users can import documents and the system should process queries about the uploaded file, providing accurate and appropriate responses in the conversation thread.
+
 
 ## D. Hub
 
@@ -125,8 +123,7 @@
 
 - [ ] Display the best model for their RAM at the top.
 - [ ] :key: Ensure that models are labeled with RAM requirements and compatibility.
-- [ ] :key: Validate that the download function is disabled for models that exceed the user's system capabilities.
+- [ ] :warning: Test that the platform provides alternative recommendations for models not suitable due to RAM limitations.
-- [ ] Test that the platform provides alternative recommendations for models not suitable due to RAM limitations.
 - [ ] :key: Check the download model functionality and validate if the cancel download feature works correctly.
 
 ### 3. Users can download models via a HuggingFace URL (coming soon)
@@ -139,7 +136,7 @@
 
 - [ ] :key: Have clear instructions so users can do their own.
 - [ ] :key: Ensure the new model updates after restarting the app.
-- [ ] Ensure it raises clear errors for users to fix the problem while adding a new model.
+- [ ] :warning:Ensure it raises clear errors for users to fix the problem while adding a new model.
 
 ### 5. Users can use the model as they want
 
@@ -149,9 +146,13 @@
 - [ ] Check if starting another model stops the other model entirely.
 - [ ] Check the `Explore models` navigate correctly to the model panel.
 - [ ] :key: Check when deleting a model it will delete all the files on the user's computer.
-- [ ] The recommended tags should present right for the user's hardware.
+- [ ] :warning:The recommended tags should present right for the user's hardware.
 - [ ] Assess that the descriptions of models are accurate and informative.
 
+### 6. Users can Integrate With a Remote Server
+- [ ] :key: Import openAI GPT model https://jan.ai/guides/using-models/integrate-with-remote-server/ and the model displayed in Hub / Thread dropdown
+- [ ] Users can use the remote model properly
+
 ## E. System Monitor
 
 ### 1. Users can see disk and RAM utilization
@@ -181,7 +182,7 @@
 - [ ] Confirm that the application saves the theme preference and persists it across sessions.
 - [ ] Validate that all elements of the UI are compatible with the theme changes and maintain legibility and contrast.
 
-### 2. Users change the extensions
+### 2. Users change the extensions [TBU]
 
 - [ ] Confirm that the `Extensions` tab lists all available plugins.
 - [ ] :key: Test the toggle switch for each plugin to ensure it enables or disables the plugin correctly.
@@ -208,3 +209,19 @@
 - [ ] :key: Test that the application prevents the installation of incompatible or corrupt plugin files.
 - [ ] :key: Check that the user can uninstall or disable custom plugins as easily as pre-installed ones.
 - [ ] Verify that the application's performance remains stable after the installation of custom plugins.
+
+### 5. Advanced Settings
+- [ ] Attemp to test downloading model from hub using **HTTP Proxy** [guideline](https://github.com/janhq/jan/pull/1562)
+- [ ] Users can move **Jan data folder**
+- [ ] Users can click on Reset button to **factory reset** app settings to its original state & delete all usage data.
+
+## G. Local API server
+
+### 1. Local Server Usage with Server Options
+- [ ] :key: Explore API Reference: Swagger API for sending/receiving requests
+    - [ ] Use default server option
+    - [ ] Configure and use custom server options
+- [ ] Test starting/stopping the local API server with different Model/Model settings
+- [ ] Server logs captured with correct Server Options provided
+- [ ] Verify functionality of Open logs/Clear feature
+- [ ] Ensure that threads and other functions impacting the model are disabled while the local server is running

electron/handlers/app.ts

@@ -1,5 +1,5 @@
 import { app, ipcMain, dialog, shell } from 'electron'
-import { join, basename } from 'path'
+import { join, basename, relative as getRelative, isAbsolute } from 'path'
 import { WindowManager } from './../managers/window'
 import { getResourcePath } from './../utils/path'
 import { AppRoute, AppConfiguration } from '@janhq/core'
@@ -50,6 +50,27 @@ export function handleAppIPCs() {
     join(...paths)
   )
 
+  /**
+   * Checks if the given path is a subdirectory of the given directory.
+   *
+   * @param _event - The IPC event object.
+   * @param from - The path to check.
+   * @param to - The directory to check against.
+   *
+   * @returns {Promise<boolean>} - A promise that resolves with the result.
+   */
+  ipcMain.handle(
+    AppRoute.isSubdirectory,
+    async (_event, from: string, to: string) => {
+      const relative = getRelative(from, to)
+      const isSubdir =
+        relative && !relative.startsWith('..') && !isAbsolute(relative)
+
+      if (isSubdir === '') return false
+      else return isSubdir
+    }
+  )
+
   /**
    * Retrieve basename from given path, respect to the current OS.
    */

electron/handlers/fileManager.ts

@@ -1,4 +1,4 @@
-import { ipcMain } from 'electron'
+import { ipcMain, app } from 'electron'
 // @ts-ignore
 import reflect from '@alumna/reflect'
 
@@ -38,6 +38,10 @@ export function handleFileMangerIPCs() {
     getResourcePath()
   )
 
+  ipcMain.handle(FileManagerRoute.getUserHomePath, async (_event) =>
+    app.getPath('home')
+  )
+
   // handle fs is directory here
   ipcMain.handle(
     FileManagerRoute.fileStat,
@@ -59,4 +63,20 @@ export function handleFileMangerIPCs() {
       return fileStat
     }
   )
+
+  ipcMain.handle(
+    FileManagerRoute.writeBlob,
+    async (_event, path: string, data: string): Promise<void> => {
+      try {
+        const normalizedPath = normalizeFilePath(path)
+        const dataBuffer = Buffer.from(data, 'base64')
+        fs.writeFileSync(
+          join(getJanDataFolderPath(), normalizedPath),
+          dataBuffer
+        )
+      } catch (err) {
+        console.error(`writeFile ${path} result: ${err}`)
+      }
+    }
+  )
 }

electron/handlers/fs.ts

@@ -1,9 +1,9 @@
 import { ipcMain } from 'electron'
 
-import { FileSystemRoute } from '@janhq/core'
-import { join } from 'path'
 import { getJanDataFolderPath, normalizeFilePath } from '@janhq/core/node'
-
+import fs from 'fs'
+import { FileManagerRoute, FileSystemRoute } from '@janhq/core'
+import { join } from 'path'
 /**
  * Handles file system operations.
  */
@@ -15,7 +15,7 @@ export function handleFsIPCs() {
         mdl[route](
           ...args.map((arg) =>
             typeof arg === 'string' &&
-            (arg.includes(`file:/`) || arg.includes(`file:\\`))
+            (arg.startsWith(`file:/`) || arg.startsWith(`file:\\`))
               ? join(getJanDataFolderPath(), normalizeFilePath(arg))
               : arg
           )

electron/main.ts

@@ -28,6 +28,22 @@ import { setupCore } from './utils/setup'
 
 app
   .whenReady()
+  .then(async () => {
+    if (!app.isPackaged) {
+      // Which means you're running from source code
+      const { default: installExtension, REACT_DEVELOPER_TOOLS } = await import(
+        'electron-devtools-installer'
+      ) // Don't use import on top level, since the installer package is dev-only
+      try {
+        const name = installExtension(REACT_DEVELOPER_TOOLS)
+        console.log(`Added Extension: ${name}`)
+      } catch (err) {
+        console.log('An error occurred while installing devtools:')
+        console.error(err)
+        // Only log the error and don't throw it because it's not critical
+      }
+    }
+  })
   .then(setupCore)
   .then(createUserSpace)
   .then(migrateExtensions)

electron/merge-latest-ymls.js

@@ -0,0 +1,27 @@
+const yaml = require('js-yaml')
+const fs = require('fs')
+
+// get two file paths from arguments:
+const [, , ...args] = process.argv
+const file1 = args[0]
+const file2 = args[1]
+const file3 = args[2]
+
+// check that all arguments are present and throw error instead
+if (!file1 || !file2 || !file3) {
+  throw new Error('Please provide 3 file paths as arguments: path to file1, to file2 and destination path')
+}
+
+const doc1 = yaml.load(fs.readFileSync(file1, 'utf8'))
+console.log('doc1: ', doc1)
+
+const doc2 = yaml.load(fs.readFileSync(file2, 'utf8'))
+console.log('doc2: ', doc2)
+
+const merged = { ...doc1, ...doc2 }
+merged.files.push(...doc1.files)
+
+console.log('merged', merged)
+
+const mergedYml = yaml.dump(merged)
+fs.writeFileSync(file3, mergedYml, 'utf8')

electron/package.json

@@ -63,11 +63,11 @@
     "build:test:darwin": "tsc -p . && electron-builder -p never -m --dir",
     "build:test:win32": "tsc -p . && electron-builder -p never -w --dir",
     "build:test:linux": "tsc -p . && electron-builder -p never -l --dir",
-    "build:darwin": "tsc -p . && electron-builder -p never -m --x64 --arm64",
+    "build:darwin": "tsc -p . && electron-builder -p never -m",
     "build:win32": "tsc -p . && electron-builder -p never -w",
     "build:linux": "tsc -p . && electron-builder -p never -l deb -l AppImage",
     "build:publish": "run-script-os",
-    "build:publish:darwin": "tsc -p . && electron-builder -p always -m --x64 --arm64",
+    "build:publish:darwin": "tsc -p . && electron-builder -p always -m",
     "build:publish:win32": "tsc -p . && electron-builder -p always -w",
     "build:publish:linux": "tsc -p . && electron-builder -p always -l deb -l AppImage"
   },
@@ -86,7 +86,7 @@
     "request": "^2.88.2",
     "request-progress": "^3.0.0",
     "rimraf": "^5.0.5",
-    "typescript": "^5.3.3",
+    "typescript": "^5.2.2",
     "ulid": "^2.3.0",
     "use-debounce": "^9.0.4"
   },
@@ -99,6 +99,7 @@
     "@typescript-eslint/parser": "^6.7.3",
     "electron": "28.0.0",
     "electron-builder": "^24.9.1",
+    "electron-devtools-installer": "^3.2.0",
     "electron-playwright-helpers": "^1.6.0",
     "eslint-plugin-react": "^7.33.2",
     "run-script-os": "^1.1.6"

electron/playwright.config.ts

@@ -1,9 +1,16 @@
-import { PlaywrightTestConfig } from "@playwright/test";
+import { PlaywrightTestConfig } from '@playwright/test'
 
 const config: PlaywrightTestConfig = {
-  testDir: "./tests",
+  testDir: './tests/e2e',
   retries: 0,
-  timeout: 120000,
+  globalTimeout: 300000,
-};
+  use: {
+    screenshot: 'only-on-failure',
+    video: 'retain-on-failure',
+    trace: 'retain-on-failure',
+  },
 
-export default config;
+  reporter: [['html', { outputFolder: './playwright-report' }]],
+}
+
+export default config

electron/tests/e2e/hub.e2e.spec.ts

@@ -0,0 +1,34 @@
+import {
+  page,
+  test,
+  setupElectron,
+  teardownElectron,
+  TIMEOUT,
+} from '../pages/basePage'
+import { expect } from '@playwright/test'
+
+test.beforeAll(async () => {
+  const appInfo = await setupElectron()
+  expect(appInfo.asar).toBe(true)
+  expect(appInfo.executable).toBeTruthy()
+  expect(appInfo.main).toBeTruthy()
+  expect(appInfo.name).toBe('jan')
+  expect(appInfo.packageJson).toBeTruthy()
+  expect(appInfo.packageJson.name).toBe('jan')
+  expect(appInfo.platform).toBeTruthy()
+  expect(appInfo.platform).toBe(process.platform)
+  expect(appInfo.resourcesDir).toBeTruthy()
+})
+
+test.afterAll(async () => {
+  await teardownElectron()
+})
+
+test('explores hub', async () => {
+  await page.getByTestId('Hub').first().click({
+    timeout: TIMEOUT,
+  })
+  await page.getByTestId('hub-container-test-id').isVisible({
+    timeout: TIMEOUT,
+  })
+})

electron/tests/e2e/navigation.e2e.spec.ts

@@ -0,0 +1,38 @@
+import { expect } from '@playwright/test'
+import {
+  page,
+  setupElectron,
+  TIMEOUT,
+  test,
+  teardownElectron,
+} from '../pages/basePage'
+
+test.beforeAll(async () => {
+  await setupElectron()
+})
+
+test.afterAll(async () => {
+  await teardownElectron()
+})
+
+test('renders left navigation panel', async () => {
+  const systemMonitorBtn = await page
+    .getByTestId('System Monitor')
+    .first()
+    .isEnabled({
+      timeout: TIMEOUT,
+    })
+  const settingsBtn = await page
+    .getByTestId('Thread')
+    .first()
+    .isEnabled({ timeout: TIMEOUT })
+  expect([systemMonitorBtn, settingsBtn].filter((e) => !e).length).toBe(0)
+  // Chat section should be there
+  await page.getByTestId('Local API Server').first().click({
+    timeout: TIMEOUT,
+  })
+  const localServer = page.getByTestId('local-server-testid').first()
+  await expect(localServer).toBeVisible({
+    timeout: TIMEOUT,
+  })
+})

electron/tests/e2e/settings.e2e.spec.ts

@@ -0,0 +1,23 @@
+import { expect } from '@playwright/test'
+
+import {
+  setupElectron,
+  teardownElectron,
+  test,
+  page,
+  TIMEOUT,
+} from '../pages/basePage'
+
+test.beforeAll(async () => {
+  await setupElectron()
+})
+
+test.afterAll(async () => {
+  await teardownElectron()
+})
+
+test('shows settings', async () => {
+  await page.getByTestId('Settings').first().click({ timeout: TIMEOUT })
+  const settingDescription = page.getByTestId('testid-setting-description')
+  await expect(settingDescription).toBeVisible({ timeout: TIMEOUT })
+})

electron/tests/explore.e2e.spec.ts

@@ -1,41 +0,0 @@
-import { _electron as electron } from 'playwright'
-import { ElectronApplication, Page, expect, test } from '@playwright/test'
-
-import {
-  findLatestBuild,
-  parseElectronApp,
-  stubDialog,
-} from 'electron-playwright-helpers'
-
-let electronApp: ElectronApplication
-let page: Page
-
-test.beforeAll(async () => {
-  process.env.CI = 'e2e'
-
-  const latestBuild = findLatestBuild('dist')
-  expect(latestBuild).toBeTruthy()
-
-  // parse the packaged Electron app and find paths and other info
-  const appInfo = parseElectronApp(latestBuild)
-  expect(appInfo).toBeTruthy()
-
-  electronApp = await electron.launch({
-    args: [appInfo.main], // main file from package.json
-    executablePath: appInfo.executable, // path to the Electron executable
-  })
-  await stubDialog(electronApp, 'showMessageBox', { response: 1 })
-
-  page = await electronApp.firstWindow()
-})
-
-test.afterAll(async () => {
-  await electronApp.close()
-  await page.close()
-})
-
-test('explores models', async () => {
-  await page.getByTestId('Hub').first().click()
-  await page.getByTestId('testid-explore-models').isVisible()
-  //   More test cases here...
-})

electron/tests/main.e2e.spec.ts

@@ -1,55 +0,0 @@
-import { _electron as electron } from 'playwright'
-import { ElectronApplication, Page, expect, test } from '@playwright/test'
-
-import {
-  findLatestBuild,
-  parseElectronApp,
-  stubDialog,
-} from 'electron-playwright-helpers'
-
-let electronApp: ElectronApplication
-let page: Page
-
-test.beforeAll(async () => {
-  process.env.CI = 'e2e'
-
-  const latestBuild = findLatestBuild('dist')
-  expect(latestBuild).toBeTruthy()
-
-  // parse the packaged Electron app and find paths and other info
-  const appInfo = parseElectronApp(latestBuild)
-  expect(appInfo).toBeTruthy()
-  expect(appInfo.asar).toBe(true)
-  expect(appInfo.executable).toBeTruthy()
-  expect(appInfo.main).toBeTruthy()
-  expect(appInfo.name).toBe('jan')
-  expect(appInfo.packageJson).toBeTruthy()
-  expect(appInfo.packageJson.name).toBe('jan')
-  expect(appInfo.platform).toBeTruthy()
-  expect(appInfo.platform).toBe(process.platform)
-  expect(appInfo.resourcesDir).toBeTruthy()
-
-  electronApp = await electron.launch({
-    args: [appInfo.main], // main file from package.json
-    executablePath: appInfo.executable, // path to the Electron executable
-  })
-  await stubDialog(electronApp, 'showMessageBox', { response: 1 })
-
-  page = await electronApp.firstWindow()
-})
-
-test.afterAll(async () => {
-  await electronApp.close()
-  await page.close()
-})
-
-test('renders the home page', async () => {
-  expect(page).toBeDefined()
-
-  // Welcome text is available
-  const welcomeText = await page
-    .getByTestId('testid-welcome-title')
-    .first()
-    .isVisible()
-  expect(welcomeText).toBe(false)
-})

electron/tests/navigation.e2e.spec.ts

@@ -1,54 +0,0 @@
-import { _electron as electron } from 'playwright'
-import { ElectronApplication, Page, expect, test } from '@playwright/test'
-
-import {
-  findLatestBuild,
-  parseElectronApp,
-  stubDialog,
-} from 'electron-playwright-helpers'
-
-let electronApp: ElectronApplication
-let page: Page
-
-test.beforeAll(async () => {
-  process.env.CI = 'e2e'
-
-  const latestBuild = findLatestBuild('dist')
-  expect(latestBuild).toBeTruthy()
-
-  // parse the packaged Electron app and find paths and other info
-  const appInfo = parseElectronApp(latestBuild)
-  expect(appInfo).toBeTruthy()
-
-  electronApp = await electron.launch({
-    args: [appInfo.main], // main file from package.json
-    executablePath: appInfo.executable, // path to the Electron executable
-  })
-  await stubDialog(electronApp, 'showMessageBox', { response: 1 })
-
-  page = await electronApp.firstWindow()
-})
-
-test.afterAll(async () => {
-  await electronApp.close()
-  await page.close()
-})
-
-test('renders left navigation panel', async () => {
-  // Chat section should be there
-  const chatSection = await page.getByTestId('Chat').first().isVisible()
-  expect(chatSection).toBe(false)
-
-  // Home actions
-  /* Disable unstable feature tests
-   ** const botBtn = await page.getByTestId("Bot").first().isEnabled();
-   ** Enable back when it is whitelisted
-   */
-
-  const systemMonitorBtn = await page
-    .getByTestId('System Monitor')
-    .first()
-    .isEnabled()
-  const settingsBtn = await page.getByTestId('Settings').first().isEnabled()
-  expect([systemMonitorBtn, settingsBtn].filter((e) => !e).length).toBe(0)
-})

electron/tests/pages/basePage.ts

@@ -0,0 +1,67 @@
+import {
+  expect,
+  test as base,
+  _electron as electron,
+  ElectronApplication,
+  Page,
+} from '@playwright/test'
+import {
+  findLatestBuild,
+  parseElectronApp,
+  stubDialog,
+} from 'electron-playwright-helpers'
+
+export const TIMEOUT: number = parseInt(process.env.TEST_TIMEOUT || '300000')
+
+export let electronApp: ElectronApplication
+export let page: Page
+
+export async function setupElectron() {
+  process.env.CI = 'e2e'
+
+  const latestBuild = findLatestBuild('dist')
+  expect(latestBuild).toBeTruthy()
+
+  // parse the packaged Electron app and find paths and other info
+  const appInfo = parseElectronApp(latestBuild)
+  expect(appInfo).toBeTruthy()
+
+  electronApp = await electron.launch({
+    args: [appInfo.main], // main file from package.json
+    executablePath: appInfo.executable, // path to the Electron executable
+  })
+  await stubDialog(electronApp, 'showMessageBox', { response: 1 })
+
+  page = await electronApp.firstWindow({
+    timeout: TIMEOUT,
+  })
+  // Return appInfo for future use
+  return appInfo
+}
+
+export async function teardownElectron() {
+  await page.close()
+  await electronApp.close()
+}
+
+export const test = base.extend<{
+  attachScreenshotsToReport: void
+}>({
+  attachScreenshotsToReport: [
+    async ({ request }, use, testInfo) => {
+      await use()
+
+      // After the test, we can check whether the test passed or failed.
+      if (testInfo.status !== testInfo.expectedStatus) {
+        const screenshot = await page.screenshot()
+        await testInfo.attach('screenshot', {
+          body: screenshot,
+          contentType: 'image/png',
+        })
+      }
+    },
+    { auto: true },
+  ],
+})
+
+test.setTimeout(TIMEOUT)

electron/tests/settings.e2e.spec.ts

@@ -1,40 +0,0 @@
-import { _electron as electron } from 'playwright'
-import { ElectronApplication, Page, expect, test } from '@playwright/test'
-
-import {
-  findLatestBuild,
-  parseElectronApp,
-  stubDialog,
-} from 'electron-playwright-helpers'
-
-let electronApp: ElectronApplication
-let page: Page
-
-test.beforeAll(async () => {
-  process.env.CI = 'e2e'
-
-  const latestBuild = findLatestBuild('dist')
-  expect(latestBuild).toBeTruthy()
-
-  // parse the packaged Electron app and find paths and other info
-  const appInfo = parseElectronApp(latestBuild)
-  expect(appInfo).toBeTruthy()
-
-  electronApp = await electron.launch({
-    args: [appInfo.main], // main file from package.json
-    executablePath: appInfo.executable, // path to the Electron executable
-  })
-  await stubDialog(electronApp, 'showMessageBox', { response: 1 })
-
-  page = await electronApp.firstWindow()
-})
-
-test.afterAll(async () => {
-  await electronApp.close()
-  await page.close()
-})
-
-test('shows settings', async () => {
-  await page.getByTestId('Settings').first().click()
-  await page.getByTestId('testid-setting-description').isVisible()
-})

electron/tests/system-monitor.e2e.spec.ts

@@ -1,41 +0,0 @@
-import { _electron as electron } from 'playwright'
-import { ElectronApplication, Page, expect, test } from '@playwright/test'
-
-import {
-  findLatestBuild,
-  parseElectronApp,
-  stubDialog,
-} from 'electron-playwright-helpers'
-
-let electronApp: ElectronApplication
-let page: Page
-
-test.beforeAll(async () => {
-  process.env.CI = 'e2e'
-
-  const latestBuild = findLatestBuild('dist')
-  expect(latestBuild).toBeTruthy()
-
-  // parse the packaged Electron app and find paths and other info
-  const appInfo = parseElectronApp(latestBuild)
-  expect(appInfo).toBeTruthy()
-
-  electronApp = await electron.launch({
-    args: [appInfo.main], // main file from package.json
-    executablePath: appInfo.executable, // path to the Electron executable
-  })
-  await stubDialog(electronApp, 'showMessageBox', { response: 1 })
-
-  page = await electronApp.firstWindow()
-})
-
-test.afterAll(async () => {
-  await electronApp.close()
-  await page.close()
-})
-
-test('shows system monitor', async () => {
-  await page.getByTestId('System Monitor').first().click()
-  await page.getByTestId('testid-system-monitor').isVisible()
-  //   More test cases here...
-})

extensions/assistant-extension/package.json

@@ -3,26 +3,50 @@
   "version": "1.0.0",
   "description": "This extension enables assistants, including Jan, a default assistant that can call all downloaded models",
   "main": "dist/index.js",
-  "module": "dist/module.js",
+  "node": "dist/node/index.js",
   "author": "Jan <service@jan.ai>",
   "license": "AGPL-3.0",
   "scripts": {
-    "build": "tsc -b . && webpack --config webpack.config.js",
+    "build": "tsc --module commonjs && rollup -c rollup.config.ts",
-    "build:publish": "rimraf *.tgz --glob && npm run build && npm pack && cpx *.tgz ../../electron/pre-install"
+    "build:publish:linux": "rimraf *.tgz --glob && npm run build && npm pack && cpx *.tgz ../../electron/pre-install",
+    "build:publish:darwin": "rimraf *.tgz --glob && npm run build && ../../.github/scripts/auto-sign.sh && npm pack && cpx *.tgz ../../electron/pre-install",
+    "build:publish:win32": "rimraf *.tgz --glob && npm run build && npm pack && cpx *.tgz ../../electron/pre-install",
+    "build:publish": "run-script-os"
   },
   "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/pdf-parse": "^1.1.4",
+    "cpx": "^1.5.0",
     "rimraf": "^3.0.2",
-    "webpack": "^5.88.2",
+    "rollup": "^2.38.5",
-    "webpack-cli": "^5.1.4"
+    "rollup-plugin-define": "^1.0.1",
+    "rollup-plugin-sourcemaps": "^0.6.3",
+    "rollup-plugin-typescript2": "^0.36.0",
+    "typescript": "^5.3.3",
+    "run-script-os": "^1.1.6"
   },
   "dependencies": {
     "@janhq/core": "file:../../core",
+    "@langchain/community": "0.0.13",
+    "hnswlib-node": "^1.4.2",
+    "langchain": "^0.0.214",
     "path-browserify": "^1.0.1",
+    "pdf-parse": "^1.1.1",
     "ts-loader": "^9.5.0"
   },
   "files": [
     "dist/*",
     "package.json",
     "README.md"
+  ],
+  "bundleDependencies": [
+    "@janhq/core",
+    "@langchain/community",
+    "hnswlib-node",
+    "langchain",
+    "pdf-parse"
   ]
 }

extensions/assistant-extension/rollup.config.ts

@@ -0,0 +1,81 @@
+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");
+
+const pkg = require("./package.json");
+
+export default [
+  {
+    input: `src/index.ts`,
+    output: [{ file: pkg.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}`),
+        EXTENSION_NAME: JSON.stringify(packageJson.name),
+        VERSION: JSON.stringify(packageJson.version),
+      }),
+      // 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: [{ dir: "dist/node", format: "cjs", sourcemap: false }],
+    // Indicate here external modules you don't wanna include in your bundle (i.e.: 'lodash')
+    external: [
+      "@janhq/core/node",
+      "@langchain/community",
+      "langchain",
+      "langsmith",
+      "path",
+      "hnswlib-node",
+    ],
+    watch: {
+      include: "src/node/**",
+    },
+    // inlineDynamicImports: true,
+    plugins: [
+      // Allow json resolution
+      json(),
+      // Compile TypeScript files
+      typescript({ useTsconfigDeclarationDir: true }),
+      // Allow bundling cjs modules (unlike webpack, rollup doesn't understand cjs)
+      commonjs({
+        ignoreDynamicRequires: true,
+      }),
+      // 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(),
+    ],
+  },
+];

extensions/assistant-extension/src/@types/global.d.ts

@@ -1 +1,3 @@
-declare const MODULE: string;
+declare const NODE: string;
+declare const EXTENSION_NAME: string;
+declare const VERSION: string;

extensions/assistant-extension/src/index.ts

@@ -1,33 +1,174 @@
-import { fs, Assistant } from "@janhq/core";
+import {
-import { AssistantExtension } from "@janhq/core";
+  fs,
-import { join } from "path";
+  Assistant,
+  MessageRequest,
+  events,
+  InferenceEngine,
+  MessageEvent,
+  InferenceEvent,
+  joinPath,
+  executeOnMain,
+  AssistantExtension,
+} from "@janhq/core";
 
 export default class JanAssistantExtension extends AssistantExtension {
   private static readonly _homeDir = "file://assistants";
 
+  controller = new AbortController();
+  isCancelled = false;
+  retrievalThreadId: string | undefined = undefined;
+
   async onLoad() {
     // making the assistant directory
-    if (!(await fs.existsSync(JanAssistantExtension._homeDir)))
+    const assistantDirExist = await fs.existsSync(
-      fs.mkdirSync(JanAssistantExtension._homeDir).then(() => {
+      JanAssistantExtension._homeDir,
+    );
+    if (
+      localStorage.getItem(`${EXTENSION_NAME}-version`) !== VERSION ||
+      !assistantDirExist
+    ) {
+      if (!assistantDirExist)
+        await fs.mkdirSync(JanAssistantExtension._homeDir);
+
+      // Write assistant metadata
       this.createJanAssistant();
+      // Finished migration
+      localStorage.setItem(`${EXTENSION_NAME}-version`, VERSION);
+    }
+
+    // Events subscription
+    events.on(MessageEvent.OnMessageSent, (data: MessageRequest) =>
+      JanAssistantExtension.handleMessageRequest(data, this),
+    );
+
+    events.on(InferenceEvent.OnInferenceStopped, () => {
+      JanAssistantExtension.handleInferenceStopped(this);
     });
   }
 
+  private static async handleInferenceStopped(instance: JanAssistantExtension) {
+    instance.isCancelled = true;
+    instance.controller?.abort();
+  }
+
+  private static async handleMessageRequest(
+    data: MessageRequest,
+    instance: JanAssistantExtension,
+  ) {
+    instance.isCancelled = false;
+    instance.controller = new AbortController();
+
+    if (
+      data.model?.engine !== InferenceEngine.tool_retrieval_enabled ||
+      !data.messages ||
+      !data.thread?.assistants[0]?.tools
+    ) {
+      return;
+    }
+
+    const latestMessage = data.messages[data.messages.length - 1];
+
+    // Ingest the document if needed
+    if (
+      latestMessage &&
+      latestMessage.content &&
+      typeof latestMessage.content !== "string"
+    ) {
+      const docFile = latestMessage.content[1]?.doc_url?.url;
+      if (docFile) {
+        await executeOnMain(
+          NODE,
+          "toolRetrievalIngestNewDocument",
+          docFile,
+          data.model?.proxyEngine,
+        );
+      }
+    }
+
+    // Load agent on thread changed
+    if (instance.retrievalThreadId !== data.threadId) {
+      await executeOnMain(NODE, "toolRetrievalLoadThreadMemory", data.threadId);
+
+      instance.retrievalThreadId = data.threadId;
+
+      // Update the text splitter
+      await executeOnMain(
+        NODE,
+        "toolRetrievalUpdateTextSplitter",
+        data.thread.assistants[0].tools[0]?.settings?.chunk_size ?? 4000,
+        data.thread.assistants[0].tools[0]?.settings?.chunk_overlap ?? 200,
+      );
+    }
+
+    if (latestMessage.content) {
+      const prompt =
+        typeof latestMessage.content === "string"
+          ? latestMessage.content
+          : latestMessage.content[0].text;
+      // Retrieve the result
+      console.debug("toolRetrievalQuery", latestMessage.content);
+      const retrievalResult = await executeOnMain(
+        NODE,
+        "toolRetrievalQueryResult",
+        prompt,
+      );
+
+      // Update the message content
+      // Using the retrieval template with the result and query
+      if (data.thread?.assistants[0].tools)
+        data.messages[data.messages.length - 1].content =
+          data.thread.assistants[0].tools[0].settings?.retrieval_template
+            ?.replace("{CONTEXT}", retrievalResult)
+            .replace("{QUESTION}", prompt);
+    }
+
+    // Filter out all the messages that are not text
+    data.messages = data.messages.map((message) => {
+      if (
+        message.content &&
+        typeof message.content !== "string" &&
+        (message.content.length ?? 0) > 0
+      ) {
+        return {
+          ...message,
+          content: [message.content[0]],
+        };
+      }
+      return message;
+    });
+
+    // Reroute the result to inference engine
+    const output = {
+      ...data,
+      model: {
+        ...data.model,
+        engine: data.model.proxyEngine,
+      },
+    };
+    events.emit(MessageEvent.OnMessageSent, output);
+  }
+
   /**
    * Called when the extension is unloaded.
    */
   onUnload(): void {}
 
   async createAssistant(assistant: Assistant): Promise<void> {
-    const assistantDir = join(JanAssistantExtension._homeDir, assistant.id);
+    const assistantDir = await joinPath([
+      JanAssistantExtension._homeDir,
+      assistant.id,
+    ]);
     if (!(await fs.existsSync(assistantDir))) await fs.mkdirSync(assistantDir);
 
     // store the assistant metadata json
-    const assistantMetadataPath = join(assistantDir, "assistant.json");
+    const assistantMetadataPath = await joinPath([
+      assistantDir,
+      "assistant.json",
+    ]);
     try {
       await fs.writeFileSync(
         assistantMetadataPath,
-        JSON.stringify(assistant, null, 2)
+        JSON.stringify(assistant, null, 2),
       );
     } catch (err) {
       console.error(err);
@@ -39,14 +180,17 @@ export default class JanAssistantExtension extends AssistantExtension {
     // get all the assistant metadata json
     const results: Assistant[] = [];
     const allFileName: string[] = await fs.readdirSync(
-      JanAssistantExtension._homeDir
+      JanAssistantExtension._homeDir,
     );
     for (const fileName of allFileName) {
-      const filePath = join(JanAssistantExtension._homeDir, fileName);
+      const filePath = await joinPath([
+        JanAssistantExtension._homeDir,
+        fileName,
+      ]);
 
       if (filePath.includes(".DS_Store")) continue;
       const jsonFiles: string[] = (await fs.readdirSync(filePath)).filter(
-        (file: string) => file === "assistant.json"
+        (file: string) => file === "assistant.json",
       );
 
       if (jsonFiles.length !== 1) {
@@ -55,8 +199,8 @@ export default class JanAssistantExtension extends AssistantExtension {
       }
 
       const content = await fs.readFileSync(
-        join(filePath, jsonFiles[0]),
+        await joinPath([filePath, jsonFiles[0]]),
-        "utf-8"
+        "utf-8",
       );
       const assistant: Assistant =
         typeof content === "object" ? content : JSON.parse(content);
@@ -73,7 +217,10 @@ export default class JanAssistantExtension extends AssistantExtension {
     }
 
     // remove the directory
-    const assistantDir = join(JanAssistantExtension._homeDir, assistant.id);
+    const assistantDir = await joinPath([
+      JanAssistantExtension._homeDir,
+      assistant.id,
+    ]);
     await fs.rmdirSync(assistantDir);
     return Promise.resolve();
   }
@@ -89,7 +236,24 @@ export default class JanAssistantExtension extends AssistantExtension {
       description: "A default assistant that can use all downloaded models",
       model: "*",
       instructions: "",
-      tools: undefined,
+      tools: [
+        {
+          type: "retrieval",
+          enabled: false,
+          settings: {
+            top_k: 2,
+            chunk_size: 1024,
+            chunk_overlap: 64,
+            retrieval_template: `Use the following pieces of context to answer the question at the end. If you don't know the answer, just say that you don't know, don't try to make up an answer.
+            ----------------
+            CONTEXT: {CONTEXT}
+            ----------------
+            QUESTION: {QUESTION}
+            ----------------
+            Helpful Answer:`,
+          },
+        },
+      ],
       file_ids: [],
       metadata: undefined,
     };

extensions/assistant-extension/src/node/engine.ts

@@ -0,0 +1,13 @@
+import fs from "fs";
+import path from "path";
+import { getJanDataFolderPath } from "@janhq/core/node";
+
+// Sec: Do not send engine settings over requests
+// Read it manually instead
+export const readEmbeddingEngine = (engineName: string) => {
+  const engineSettings = fs.readFileSync(
+    path.join(getJanDataFolderPath(), "engines", `${engineName}.json`),
+    "utf-8",
+  );
+  return JSON.parse(engineSettings);
+};

extensions/assistant-extension/src/node/index.ts

@@ -0,0 +1,39 @@
+import { getJanDataFolderPath, normalizeFilePath } from "@janhq/core/node";
+import { Retrieval } from "./tools/retrieval";
+import path from "path";
+
+const retrieval = new Retrieval();
+
+export async function toolRetrievalUpdateTextSplitter(
+  chunkSize: number,
+  chunkOverlap: number,
+) {
+  retrieval.updateTextSplitter(chunkSize, chunkOverlap);
+  return Promise.resolve();
+}
+export async function toolRetrievalIngestNewDocument(
+  file: string,
+  engine: string,
+) {
+  const filePath = path.join(getJanDataFolderPath(), normalizeFilePath(file));
+  const threadPath = path.dirname(filePath.replace("files", ""));
+  retrieval.updateEmbeddingEngine(engine);
+  await retrieval.ingestAgentKnowledge(filePath, `${threadPath}/memory`);
+  return Promise.resolve();
+}
+
+export async function toolRetrievalLoadThreadMemory(threadId: string) {
+  try {
+    await retrieval.loadRetrievalAgent(
+      path.join(getJanDataFolderPath(), "threads", threadId, "memory"),
+    );
+    return Promise.resolve();
+  } catch (err) {
+    console.debug(err);
+  }
+}
+
+export async function toolRetrievalQueryResult(query: string) {
+  const res = await retrieval.generateResult(query);
+  return Promise.resolve(res);
+}

extensions/assistant-extension/src/node/tools/retrieval/index.ts

@@ -0,0 +1,77 @@
+import { RecursiveCharacterTextSplitter } from "langchain/text_splitter";
+import { formatDocumentsAsString } from "langchain/util/document";
+import { PDFLoader } from "langchain/document_loaders/fs/pdf";
+
+import { HNSWLib } from "langchain/vectorstores/hnswlib";
+
+import { OpenAIEmbeddings } from "langchain/embeddings/openai";
+import { readEmbeddingEngine } from "../../engine";
+
+export class Retrieval {
+  public chunkSize: number = 100;
+  public chunkOverlap?: number = 0;
+  private retriever: any;
+
+  private embeddingModel?: OpenAIEmbeddings = undefined;
+  private textSplitter?: RecursiveCharacterTextSplitter;
+
+  constructor(chunkSize: number = 4000, chunkOverlap: number = 200) {
+    this.updateTextSplitter(chunkSize, chunkOverlap);
+  }
+
+  public updateTextSplitter(chunkSize: number, chunkOverlap: number): void {
+    this.chunkSize = chunkSize;
+    this.chunkOverlap = chunkOverlap;
+    this.textSplitter = new RecursiveCharacterTextSplitter({
+      chunkSize: chunkSize,
+      chunkOverlap: chunkOverlap,
+    });
+  }
+
+  public updateEmbeddingEngine(engine: string): void {
+    // Engine settings are not compatible with the current embedding model params
+    // Switch case manually for now
+    const settings = readEmbeddingEngine(engine);
+    if (engine === "nitro") {
+      this.embeddingModel = new OpenAIEmbeddings(
+        { openAIApiKey: "nitro-embedding" },
+        { basePath: "http://127.0.0.1:3928/v1" },
+      );
+    } else {
+      // Fallback to OpenAI Settings
+      this.embeddingModel = new OpenAIEmbeddings({
+        openAIApiKey: settings.api_key,
+      });
+    }
+  }
+
+  public ingestAgentKnowledge = async (
+    filePath: string,
+    memoryPath: string,
+  ): Promise<any> => {
+    const loader = new PDFLoader(filePath, {
+      splitPages: true,
+    });
+    if (!this.embeddingModel) return Promise.reject();
+    const doc = await loader.load();
+    const docs = await this.textSplitter!.splitDocuments(doc);
+    const vectorStore = await HNSWLib.fromDocuments(docs, this.embeddingModel);
+    return vectorStore.save(memoryPath);
+  };
+
+  public loadRetrievalAgent = async (memoryPath: string): Promise<void> => {
+    if (!this.embeddingModel) return Promise.reject();
+    const vectorStore = await HNSWLib.load(memoryPath, this.embeddingModel);
+    this.retriever = vectorStore.asRetriever(2);
+    return Promise.resolve();
+  };
+
+  public generateResult = async (query: string): Promise<string> => {
+    if (!this.retriever) {
+      return Promise.resolve(" ");
+    }
+    const relevantDocs = await this.retriever.getRelevantDocuments(query);
+    const serializedDoc = formatDocumentsAsString(relevantDocs);
+    return Promise.resolve(serializedDoc);
+  };
+}

extensions/assistant-extension/tsconfig.json

@@ -1,14 +1,20 @@
 {
   "compilerOptions": {
-    "target": "es2016",
-    "module": "ES6",
     "moduleResolution": "node",
-    "outDir": "./dist",
+    "target": "es5",
-    "esModuleInterop": true,
+    "module": "ES2020",
-    "forceConsistentCasingInFileNames": true,
+    "lib": ["es2015", "es2016", "es2017", "dom"],
-    "strict": false,
+    "strict": true,
+    "sourceMap": true,
+    "declaration": true,
+    "allowSyntheticDefaultImports": true,
+    "experimentalDecorators": true,
+    "emitDecoratorMetadata": true,
+    "declarationDir": "dist/types",
+    "outDir": "dist",
+    "importHelpers": true,
+    "typeRoots": ["node_modules/@types"],
     "skipLibCheck": true,
-    "rootDir": "./src"
   },
-  "include": ["./src"]
+  "include": ["src"],
 }

extensions/assistant-extension/webpack.config.js

@@ -1,38 +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"],
-    fallback: {
-      path: require.resolve("path-browserify"),
-    },
-  },
-  optimization: {
-    minimize: false,
-  },
-  // Add loaders and other configuration as needed for your project
-};

extensions/conversational-extension/src/index.ts

@@ -4,16 +4,15 @@ import {
   ConversationalExtension,
   Thread,
   ThreadMessage,
+  events,
 } from '@janhq/core'
 
 /**
  * JSONConversationalExtension is a ConversationalExtension implementation that provides
  * functionality for managing threads.
  */
-export default class JSONConversationalExtension
+export default class JSONConversationalExtension extends ConversationalExtension {
-  extends ConversationalExtension
+  private static readonly _threadFolder = 'file://threads'
-{
-  private static readonly _homeDir = 'file://threads'
   private static readonly _threadInfoFileName = 'thread.json'
   private static readonly _threadMessagesFileName = 'messages.jsonl'
 
@@ -21,8 +20,8 @@ export default class JSONConversationalExtension
    * Called when the extension is loaded.
    */
   async onLoad() {
-    if (!(await fs.existsSync(JSONConversationalExtension._homeDir)))
+    if (!(await fs.existsSync(JSONConversationalExtension._threadFolder)))
-      await fs.mkdirSync(JSONConversationalExtension._homeDir)
+      await fs.mkdirSync(JSONConversationalExtension._threadFolder)
     console.debug('JSONConversationalExtension loaded')
   }
 
@@ -69,7 +68,7 @@ export default class JSONConversationalExtension
   async saveThread(thread: Thread): Promise<void> {
     try {
       const threadDirPath = await joinPath([
-        JSONConversationalExtension._homeDir,
+        JSONConversationalExtension._threadFolder,
         thread.id,
       ])
       const threadJsonPath = await joinPath([
@@ -93,7 +92,7 @@ export default class JSONConversationalExtension
    */
   async deleteThread(threadId: string): Promise<void> {
     const path = await joinPath([
-      JSONConversationalExtension._homeDir,
+      JSONConversationalExtension._threadFolder,
       `${threadId}`,
     ])
     try {
@@ -110,7 +109,7 @@ export default class JSONConversationalExtension
   async addNewMessage(message: ThreadMessage): Promise<void> {
     try {
       const threadDirPath = await joinPath([
-        JSONConversationalExtension._homeDir,
+        JSONConversationalExtension._threadFolder,
         message.thread_id,
       ])
       const threadMessagePath = await joinPath([
@@ -119,6 +118,33 @@ export default class JSONConversationalExtension
       ])
       if (!(await fs.existsSync(threadDirPath)))
         await fs.mkdirSync(threadDirPath)
+
+      if (message.content[0]?.type === 'image') {
+        const filesPath = await joinPath([threadDirPath, 'files'])
+        if (!(await fs.existsSync(filesPath))) await fs.mkdirSync(filesPath)
+
+        const imagePath = await joinPath([filesPath, `${message.id}.png`])
+        const base64 = message.content[0].text.annotations[0]
+        await this.storeImage(base64, imagePath)
+        if ((await fs.existsSync(imagePath)) && message.content?.length) {
+          // Use file path instead of blob
+          message.content[0].text.annotations[0] = `threads/${message.thread_id}/files/${message.id}.png`
+        }
+      }
+
+      if (message.content[0]?.type === 'pdf') {
+        const filesPath = await joinPath([threadDirPath, 'files'])
+        if (!(await fs.existsSync(filesPath))) await fs.mkdirSync(filesPath)
+
+        const filePath = await joinPath([filesPath, `${message.id}.pdf`])
+        const blob = message.content[0].text.annotations[0]
+        await this.storeFile(blob, filePath)
+
+        if ((await fs.existsSync(filePath)) && message.content?.length) {
+          // Use file path instead of blob
+          message.content[0].text.annotations[0] = `threads/${message.thread_id}/files/${message.id}.pdf`
+        }
+      }
       await fs.appendFileSync(threadMessagePath, JSON.stringify(message) + '\n')
       Promise.resolve()
     } catch (err) {
@@ -126,13 +152,32 @@ export default class JSONConversationalExtension
     }
   }
 
+  async storeImage(base64: string, filePath: string): Promise<void> {
+    const base64Data = base64.replace(/^data:image\/\w+;base64,/, '')
+
+    try {
+      await fs.writeBlob(filePath, base64Data)
+    } catch (err) {
+      console.error(err)
+    }
+  }
+
+  async storeFile(base64: string, filePath: string): Promise<void> {
+    const base64Data = base64.replace(/^data:application\/pdf;base64,/, '')
+    try {
+      await fs.writeBlob(filePath, base64Data)
+    } catch (err) {
+      console.error(err)
+    }
+  }
+
   async writeMessages(
     threadId: string,
     messages: ThreadMessage[]
   ): Promise<void> {
     try {
       const threadDirPath = await joinPath([
-        JSONConversationalExtension._homeDir,
+        JSONConversationalExtension._threadFolder,
         threadId,
       ])
       const threadMessagePath = await joinPath([
@@ -160,7 +205,7 @@ export default class JSONConversationalExtension
   private async readThread(threadDirName: string): Promise<any> {
     return fs.readFileSync(
       await joinPath([
-        JSONConversationalExtension._homeDir,
+        JSONConversationalExtension._threadFolder,
         threadDirName,
         JSONConversationalExtension._threadInfoFileName,
       ]),
@@ -174,14 +219,14 @@ export default class JSONConversationalExtension
    */
   private async getValidThreadDirs(): Promise<string[]> {
     const fileInsideThread: string[] = await fs.readdirSync(
-      JSONConversationalExtension._homeDir
+      JSONConversationalExtension._threadFolder
     )
 
     const threadDirs: string[] = []
     for (let i = 0; i < fileInsideThread.length; i++) {
       if (fileInsideThread[i].includes('.DS_Store')) continue
       const path = await joinPath([
-        JSONConversationalExtension._homeDir,
+        JSONConversationalExtension._threadFolder,
         fileInsideThread[i],
       ])
 
@@ -201,7 +246,7 @@ export default class JSONConversationalExtension
   async getAllMessages(threadId: string): Promise<ThreadMessage[]> {
     try {
       const threadDirPath = await joinPath([
-        JSONConversationalExtension._homeDir,
+        JSONConversationalExtension._threadFolder,
         threadId,
       ])
 
@@ -218,18 +263,17 @@ export default class JSONConversationalExtension
         JSONConversationalExtension._threadMessagesFileName,
       ])
 
-      const result = await fs
+      let readResult = await fs.readFileSync(messageFilePath, 'utf-8')
-        .readFileSync(messageFilePath, 'utf-8')
+
-        .then((content) =>
+      if (typeof readResult === 'object') {
-          content
+        readResult = JSON.stringify(readResult)
-            .toString()
+      }
-            .split('\n')
+
-            .filter((line) => line !== '')
+      const result = readResult.split('\n').filter((line) => line !== '')
-        )
 
       const messages: ThreadMessage[] = []
       result.forEach((line: string) => {
-        messages.push(JSON.parse(line) as ThreadMessage)
+        messages.push(JSON.parse(line))
       })
       return messages
     } catch (err) {

extensions/inference-nitro-extension/bin/version.txt

@@ -1 +1 @@
-0.2.12
+0.3.5

extensions/inference-nitro-extension/package.json

@@ -35,11 +35,12 @@
     "rollup-plugin-sourcemaps": "^0.6.3",
     "rollup-plugin-typescript2": "^0.36.0",
     "run-script-os": "^1.1.6",
-    "typescript": "^5.3.3"
+    "typescript": "^5.2.2"
   },
   "dependencies": {
     "@janhq/core": "file:../../core",
     "@rollup/plugin-replace": "^5.0.5",
+    "@types/os-utils": "^0.0.4",
     "fetch-retry": "^5.0.6",
     "path-browserify": "^1.0.1",
     "rxjs": "^7.8.1",

extensions/inference-nitro-extension/rollup.config.ts

@@ -27,6 +27,9 @@ export default [
         TROUBLESHOOTING_URL: JSON.stringify(
           "https://jan.ai/guides/troubleshooting"
         ),
+        JAN_SERVER_INFERENCE_URL: JSON.stringify(
+          "http://localhost:1337/v1/chat/completions"
+        ),
       }),
       // Allow json resolution
       json(),

extensions/inference-nitro-extension/src/@types/global.d.ts

@@ -1,22 +1,7 @@
 declare const NODE: string;
 declare const INFERENCE_URL: string;
 declare const TROUBLESHOOTING_URL: string;
-
+declare const JAN_SERVER_INFERENCE_URL: string;
-/**
- * The parameters for the initModel function.
- * @property settings - The settings for the machine learning model.
- * @property settings.ctx_len - The context length.
- * @property settings.ngl - The number of generated tokens.
- * @property settings.cont_batching - Whether to use continuous batching.
- * @property settings.embedding - Whether to use embedding.
- */
-interface EngineSettings {
-  ctx_len: number;
-  ngl: number;
-  cpu_threads: number;
-  cont_batching: boolean;
-  embedding: boolean;
-}
 
 /**
  * The response from the initModel function.
@@ -26,8 +11,3 @@ interface ModelOperationResponse {
   error?: any;
   modelFile?: string;
 }
-
-interface ResourcesInfo {
-  numCpuPhysicalCore: number;
-  memAvailable: number;
-}

extensions/inference-nitro-extension/src/helpers/sse.ts

@@ -6,6 +6,7 @@ import { Observable } from "rxjs";
  * @returns An Observable that emits the generated response as a string.
  */
 export function requestInference(
+  inferenceUrl: string,
   recentMessages: any[],
   model: Model,
   controller?: AbortController
@@ -17,7 +18,7 @@ export function requestInference(
       stream: true,
       ...model.parameters,
     });
-    fetch(INFERENCE_URL, {
+    fetch(inferenceUrl, {
       method: "POST",
       headers: {
         "Content-Type": "application/json",

extensions/inference-nitro-extension/src/index.ts

@@ -24,6 +24,7 @@ import {
   MessageEvent,
   ModelEvent,
   InferenceEvent,
+  ModelSettingParams,
 } from "@janhq/core";
 import { requestInference } from "./helpers/sse";
 import { ulid } from "ulid";
@@ -45,12 +46,12 @@ export default class JanInferenceNitroExtension extends InferenceExtension {
 
   private _currentModel: Model | undefined;
 
-  private _engineSettings: EngineSettings = {
+  private _engineSettings: ModelSettingParams = {
     ctx_len: 2048,
     ngl: 100,
     cpu_threads: 1,
     cont_batching: false,
-    embedding: false,
+    embedding: true,
   };
 
   controller = new AbortController();
@@ -67,15 +68,28 @@ export default class JanInferenceNitroExtension extends InferenceExtension {
    */
   private nitroProcessInfo: any = undefined;
 
+  private inferenceUrl = "";
+
   /**
    * Subscribes to events emitted by the @janhq/core package.
    */
   async onLoad() {
     if (!(await fs.existsSync(JanInferenceNitroExtension._homeDir))) {
-      await fs
+      try {
-        .mkdirSync(JanInferenceNitroExtension._homeDir)
+        await fs.mkdirSync(JanInferenceNitroExtension._homeDir);
-        .catch((err: Error) => console.debug(err));
+      } catch (e) {
+        console.debug(e);
       }
+    }
+
+    // init inference url
+    // @ts-ignore
+    const electronApi = window?.electronAPI;
+    this.inferenceUrl = INFERENCE_URL;
+    if (!electronApi) {
+      this.inferenceUrl = JAN_SERVER_INFERENCE_URL;
+    }
+    console.debug("Inference url: ", this.inferenceUrl);
 
     if (!(await fs.existsSync(JanInferenceNitroExtension._settingsDir)))
       await fs.mkdirSync(JanInferenceNitroExtension._settingsDir);
@@ -133,6 +147,7 @@ export default class JanInferenceNitroExtension extends InferenceExtension {
 
     const modelFullPath = await joinPath(["models", model.id]);
 
+    this._currentModel = model;
     const nitroInitResult = await executeOnMain(NODE, "runModel", {
       modelFullPath,
       model,
@@ -143,7 +158,6 @@ export default class JanInferenceNitroExtension extends InferenceExtension {
       return;
     }
 
-    this._currentModel = model;
     events.emit(ModelEvent.OnModelReady, model);
 
     this.getNitroProcesHealthIntervalId = setInterval(
@@ -205,7 +219,11 @@ export default class JanInferenceNitroExtension extends InferenceExtension {
     return new Promise(async (resolve, reject) => {
       if (!this._currentModel) return Promise.reject("No model loaded");
 
-      requestInference(data.messages ?? [], this._currentModel).subscribe({
+      requestInference(
+        this.inferenceUrl,
+        data.messages ?? [],
+        this._currentModel
+      ).subscribe({
         next: (_content: any) => {},
         complete: async () => {
           resolve(message);
@@ -250,7 +268,12 @@ export default class JanInferenceNitroExtension extends InferenceExtension {
       ...(this._currentModel || {}),
       ...(data.model || {}),
     };
-    requestInference(data.messages ?? [], model, this.controller).subscribe({
+    requestInference(
+      this.inferenceUrl,
+      data.messages ?? [],
+      model,
+      this.controller
+    ).subscribe({
       next: (content: any) => {
         const messageContent: ThreadContent = {
           type: ContentType.Text,

extensions/inference-nitro-extension/src/node/execute.ts

@@ -25,12 +25,12 @@ export const executableNitroFile = (): NitroExecutableOptions => {
     if (nvidiaInfo["run_mode"] === "cpu") {
       binaryFolder = path.join(binaryFolder, "win-cpu");
     } else {
-      if (nvidiaInfo["cuda"].version === "12") {
+      if (nvidiaInfo["cuda"].version === "11") {
-        binaryFolder = path.join(binaryFolder, "win-cuda-12-0");
-      } else {
         binaryFolder = path.join(binaryFolder, "win-cuda-11-7");
+      } else {
+        binaryFolder = path.join(binaryFolder, "win-cuda-12-0");
       }
-      cudaVisibleDevices = nvidiaInfo["gpu_highest_vram"];
+      cudaVisibleDevices = nvidiaInfo["gpus_in_use"].join(",");
     }
     binaryName = "nitro.exe";
   } else if (process.platform === "darwin") {
@@ -50,12 +50,12 @@ export const executableNitroFile = (): NitroExecutableOptions => {
     if (nvidiaInfo["run_mode"] === "cpu") {
       binaryFolder = path.join(binaryFolder, "linux-cpu");
     } else {
-      if (nvidiaInfo["cuda"].version === "12") {
+      if (nvidiaInfo["cuda"].version === "11") {
-        binaryFolder = path.join(binaryFolder, "linux-cuda-12-0");
-      } else {
         binaryFolder = path.join(binaryFolder, "linux-cuda-11-7");
+      } else {
+        binaryFolder = path.join(binaryFolder, "linux-cuda-12-0");
       }
-      cudaVisibleDevices = nvidiaInfo["gpu_highest_vram"];
+      cudaVisibleDevices = nvidiaInfo["gpus_in_use"].join(",");
     }
   }
   return {

extensions/inference-nitro-extension/src/node/index.ts

@@ -3,11 +3,19 @@ import path from "path";
 import { ChildProcessWithoutNullStreams, spawn } from "child_process";
 import tcpPortUsed from "tcp-port-used";
 import fetchRT from "fetch-retry";
-import { log, getJanDataFolderPath } from "@janhq/core/node";
+import {
+  log,
+  getJanDataFolderPath,
+  getSystemResourceInfo,
+} from "@janhq/core/node";
 import { getNitroProcessInfo, updateNvidiaInfo } from "./nvidia";
-import { Model, InferenceEngine, ModelSettingParams } from "@janhq/core";
+import {
+  Model,
+  InferenceEngine,
+  ModelSettingParams,
+  PromptTemplate,
+} from "@janhq/core";
 import { executableNitroFile } from "./execute";
-import { physicalCpuCount } from "./utils";
 
 // Polyfill fetch with retry
 const fetchRetry = fetchRT(fetch);
@@ -19,25 +27,6 @@ interface ModelInitOptions {
   modelFullPath: string;
   model: Model;
 }
-
-/**
- * The response object of Prompt Template parsing.
- */
-interface PromptTemplate {
-  system_prompt?: string;
-  ai_prompt?: string;
-  user_prompt?: string;
-  error?: string;
-}
-
-/**
- * Model setting args for Nitro model load.
- */
-interface ModelSettingArgs extends ModelSettingParams {
-  llama_model_path: string;
-  cpu_threads: number;
-}
-
 // The PORT to use for the Nitro subprocess
 const PORT = 3928;
 // The HOST address to use for the Nitro subprocess
@@ -60,7 +49,7 @@ let subprocess: ChildProcessWithoutNullStreams | undefined = undefined;
 // The current model file url
 let currentModelFile: string = "";
 // The current model settings
-let currentSettings: ModelSettingArgs | undefined = undefined;
+let currentSettings: ModelSettingParams | undefined = undefined;
 
 /**
  * Stops a Nitro subprocess.
@@ -78,7 +67,7 @@ function stopModel(): Promise<void> {
  * TODO: Should pass absolute of the model file instead of just the name - So we can modurize the module.ts to npm package
  */
 async function runModel(
-  wrapper: ModelInitOptions
+  wrapper: ModelInitOptions,
 ): Promise<ModelOperationResponse | void> {
   if (wrapper.model.engine !== InferenceEngine.nitro) {
     // Not a nitro model
@@ -96,7 +85,7 @@ async function runModel(
   const ggufBinFile = files.find(
     (file) =>
       file === path.basename(currentModelFile) ||
-      file.toLowerCase().includes(SUPPORTED_MODEL_FORMAT)
+      file.toLowerCase().includes(SUPPORTED_MODEL_FORMAT),
   );
 
   if (!ggufBinFile) return Promise.reject("No GGUF model file found");
@@ -106,7 +95,7 @@ async function runModel(
   if (wrapper.model.engine !== InferenceEngine.nitro) {
     return Promise.reject("Not a nitro model");
   } else {
-    const nitroResourceProbe = await getResourcesInfo();
+    const nitroResourceProbe = await getSystemResourceInfo();
     // Convert settings.prompt_template to system_prompt, user_prompt, ai_prompt
     if (wrapper.model.settings.prompt_template) {
       const promptTemplate = wrapper.model.settings.prompt_template;
@@ -133,7 +122,6 @@ async function runModel(
         mmproj: path.join(modelFolderPath, wrapper.model.settings.mmproj),
       }),
     };
-    console.log(currentSettings);
     return runNitroAndLoadModel();
   }
 }
@@ -192,10 +180,10 @@ function promptTemplateConverter(promptTemplate: string): PromptTemplate {
     const system_prompt = promptTemplate.substring(0, systemIndex);
     const user_prompt = promptTemplate.substring(
       systemIndex + systemMarker.length,
-      promptIndex
+      promptIndex,
     );
     const ai_prompt = promptTemplate.substring(
-      promptIndex + promptMarker.length
+      promptIndex + promptMarker.length,
     );
 
     // Return the split parts
@@ -205,7 +193,7 @@ function promptTemplateConverter(promptTemplate: string): PromptTemplate {
     const promptIndex = promptTemplate.indexOf(promptMarker);
     const user_prompt = promptTemplate.substring(0, promptIndex);
     const ai_prompt = promptTemplate.substring(
-      promptIndex + promptMarker.length
+      promptIndex + promptMarker.length,
     );
 
     // Return the split parts
@@ -221,6 +209,9 @@ function promptTemplateConverter(promptTemplate: string): PromptTemplate {
  * @returns A Promise that resolves when the model is loaded successfully, or rejects with an error message if the model is not found or fails to load.
  */
 function loadLLMModel(settings: any): Promise<Response> {
+  if (!settings?.ngl) {
+    settings.ngl = 100;
+  }
   log(`[NITRO]::Debug: Loading model with params ${JSON.stringify(settings)}`);
   return fetchRetry(NITRO_HTTP_LOAD_MODEL_URL, {
     method: "POST",
@@ -234,14 +225,14 @@ function loadLLMModel(settings: any): Promise<Response> {
     .then((res) => {
       log(
         `[NITRO]::Debug: Load model success with response ${JSON.stringify(
-          res
+          res,
-        )}`
+        )}`,
       );
       return Promise.resolve(res);
     })
     .catch((err) => {
       log(`[NITRO]::Error: Load model failed with error ${err}`);
-      return Promise.reject();
+      return Promise.reject(err);
     });
 }
 
@@ -263,8 +254,8 @@ async function validateModelStatus(): Promise<void> {
     retryDelay: 500,
   }).then(async (res: Response) => {
     log(
-      `[NITRO]::Debug: Validate model state success with response ${JSON.stringify(
+      `[NITRO]::Debug: Validate model state with response ${JSON.stringify(
-        res
+        res.status
       )}`
     );
     // If the response is OK, check model_loaded status.
@@ -273,9 +264,19 @@ async function validateModelStatus(): Promise<void> {
       // If the model is loaded, return an empty object.
       // Otherwise, return an object with an error message.
       if (body.model_loaded) {
+        log(
+          `[NITRO]::Debug: Validate model state success with response ${JSON.stringify(
+            body
+          )}`
+        );
         return Promise.resolve();
       }
     }
+    log(
+      `[NITRO]::Debug: Validate model state failed with response ${JSON.stringify(
+        res.statusText
+      )}`
+    );
     return Promise.reject("Validate model status failed");
   });
 }
@@ -316,7 +317,7 @@ function spawnNitroProcess(): Promise<any> {
     const args: string[] = ["1", LOCAL_HOST, PORT.toString()];
     // Execute the binary
     log(
-      `[NITRO]::Debug: Spawn nitro at path: ${executableOptions.executablePath}, and args: ${args}`
+      `[NITRO]::Debug: Spawn nitro at path: ${executableOptions.executablePath}, and args: ${args}`,
     );
     subprocess = spawn(
       executableOptions.executablePath,
@@ -327,7 +328,7 @@ function spawnNitroProcess(): Promise<any> {
           ...process.env,
           CUDA_VISIBLE_DEVICES: executableOptions.cudaVisibleDevices,
         },
-      }
+      },
     );
 
     // Handle subprocess output
@@ -352,22 +353,6 @@ function spawnNitroProcess(): Promise<any> {
   });
 }
 
-/**
- * Get the system resources information
- * TODO: Move to Core so that it can be reused
- */
-function getResourcesInfo(): Promise<ResourcesInfo> {
-  return new Promise(async (resolve) => {
-    const cpu = await physicalCpuCount();
-    log(`[NITRO]::CPU informations - ${cpu}`);
-    const response: ResourcesInfo = {
-      numCpuPhysicalCore: cpu,
-      memAvailable: 0,
-    };
-    resolve(response);
-  });
-}
-
 /**
  * Every module should have a dispose function
  * This will be called when the extension is unloaded and should clean up any resources

extensions/inference-nitro-extension/src/node/nvidia.ts

@@ -19,6 +19,8 @@ const DEFALT_SETTINGS = {
   },
   gpus: [],
   gpu_highest_vram: "",
+  gpus_in_use: [],
+  is_initial: true,
 };
 
 /**
@@ -48,11 +50,15 @@ export interface NitroProcessInfo {
  */
 export async function updateNvidiaInfo() {
   if (process.platform !== "darwin") {
-    await Promise.all([
+    let data;
-      updateNvidiaDriverInfo(),
+    try {
-      updateCudaExistence(),
+      data = JSON.parse(readFileSync(NVIDIA_INFO_FILE, "utf-8"));
-      updateGpuInfo(),
+    } catch (error) {
-    ]);
+      data = DEFALT_SETTINGS;
+      writeFileSync(NVIDIA_INFO_FILE, JSON.stringify(data, null, 2));
+    }
+    updateNvidiaDriverInfo();
+    updateGpuInfo();
   }
 }
 
@@ -73,12 +79,7 @@ export async function updateNvidiaDriverInfo(): Promise<void> {
   exec(
     "nvidia-smi --query-gpu=driver_version --format=csv,noheader",
     (error, stdout) => {
-      let data;
+      let data = JSON.parse(readFileSync(NVIDIA_INFO_FILE, "utf-8"));
-      try {
-        data = JSON.parse(readFileSync(NVIDIA_INFO_FILE, "utf-8"));
-      } catch (error) {
-        data = DEFALT_SETTINGS;
-      }
 
       if (!error) {
         const firstLine = stdout.split("\n")[0].trim();
@@ -107,7 +108,7 @@ export function checkFileExistenceInPaths(
 /**
  * Validate cuda for linux and windows
  */
-export function updateCudaExistence() {
+export function updateCudaExistence(data: Record<string, any> = DEFALT_SETTINGS): Record<string, any> {
   let filesCuda12: string[];
   let filesCuda11: string[];
   let paths: string[];
@@ -141,19 +142,14 @@ export function updateCudaExistence() {
     cudaVersion = "12";
   }
 
-  let data;
-  try {
-    data = JSON.parse(readFileSync(NVIDIA_INFO_FILE, "utf-8"));
-  } catch (error) {
-    data = DEFALT_SETTINGS;
-  }
-
   data["cuda"].exist = cudaExists;
   data["cuda"].version = cudaVersion;
-  if (cudaExists) {
+  console.log(data["is_initial"], data["gpus_in_use"]);
+  if (cudaExists && data["is_initial"] && data["gpus_in_use"].length > 0) {
     data.run_mode = "gpu";
   }
-  writeFileSync(NVIDIA_INFO_FILE, JSON.stringify(data, null, 2));
+  data.is_initial = false;
+  return data;
 }
 
 /**
@@ -161,14 +157,9 @@ export function updateCudaExistence() {
  */
 export async function updateGpuInfo(): Promise<void> {
   exec(
-    "nvidia-smi --query-gpu=index,memory.total --format=csv,noheader,nounits",
+    "nvidia-smi --query-gpu=index,memory.total,name --format=csv,noheader,nounits",
     (error, stdout) => {
-      let data;
+      let data = JSON.parse(readFileSync(NVIDIA_INFO_FILE, "utf-8"));
-      try {
-        data = JSON.parse(readFileSync(NVIDIA_INFO_FILE, "utf-8"));
-      } catch (error) {
-        data = DEFALT_SETTINGS;
-      }
 
       if (!error) {
         // Get GPU info and gpu has higher memory first
@@ -178,21 +169,27 @@ export async function updateGpuInfo(): Promise<void> {
           .trim()
           .split("\n")
           .map((line) => {
-            let [id, vram] = line.split(", ");
+            let [id, vram, name] = line.split(", ");
             vram = vram.replace(/\r/g, "");
             if (parseFloat(vram) > highestVram) {
               highestVram = parseFloat(vram);
               highestVramId = id;
             }
-            return { id, vram };
+            return { id, vram, name };
           });
 
-        data["gpus"] = gpus;
+        data.gpus = gpus;
-        data["gpu_highest_vram"] = highestVramId;
+        data.gpu_highest_vram = highestVramId;
       } else {
-        data["gpus"] = [];
+        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(NVIDIA_INFO_FILE, JSON.stringify(data, null, 2));
       Promise.resolve();
     }

extensions/inference-nitro-extension/src/node/utils.ts

@@ -1,56 +0,0 @@
-import os from "os";
-import childProcess from "child_process";
-
-function exec(command: string): Promise<string> {
-  return new Promise((resolve, reject) => {
-    childProcess.exec(command, { encoding: "utf8" }, (error, stdout) => {
-      if (error) {
-        reject(error);
-      } else {
-        resolve(stdout);
-      }
-    });
-  });
-}
-
-let amount: number;
-const platform = os.platform();
-
-export async function physicalCpuCount(): Promise<number> {
-  return new Promise((resolve, reject) => {
-    if (platform === "linux") {
-      exec('lscpu -p | egrep -v "^#" | sort -u -t, -k 2,4 | wc -l')
-        .then((output) => {
-          amount = parseInt(output.trim(), 10);
-          resolve(amount);
-        })
-        .catch(reject);
-    } else if (platform === "darwin") {
-      exec("sysctl -n hw.physicalcpu_max")
-        .then((output) => {
-          amount = parseInt(output.trim(), 10);
-          resolve(amount);
-        })
-        .catch(reject);
-    } else if (platform === "win32") {
-      exec("WMIC CPU Get NumberOfCores")
-        .then((output) => {
-          amount = output
-            .split(os.EOL)
-            .map((line: string) => parseInt(line))
-            .filter((value: number) => !isNaN(value))
-            .reduce((sum: number, number: number) => sum + number, 1);
-          resolve(amount);
-        })
-        .catch(reject);
-    } else {
-      const cores = os.cpus().filter((cpu: any, index: number) => {
-        const hasHyperthreading = cpu.model.includes("Intel");
-        const isOdd = index % 2 === 1;
-        return !hasHyperthreading || isOdd;
-      });
-      amount = cores.length;
-      resolve(amount);
-    }
-  });
-}

extensions/inference-openai-extension/src/index.ts

@@ -15,10 +15,13 @@ import {
   ThreadMessage,
   events,
   fs,
+  InferenceEngine,
   BaseExtension,
   MessageEvent,
   ModelEvent,
   InferenceEvent,
+  AppConfigurationEventName,
+  joinPath,
 } from "@janhq/core";
 import { requestInference } from "./helpers/sse";
 import { ulid } from "ulid";
@@ -30,7 +33,7 @@ import { join } from "path";
  * It also subscribes to events emitted by the @janhq/core package and handles new message requests.
  */
 export default class JanInferenceOpenAIExtension extends BaseExtension {
-  private static readonly _homeDir = "file://engines";
+  private static readonly _engineDir = "file://engines";
   private static readonly _engineMetadataFileName = "openai.json";
 
   private static _currentModel: OpenAIModel;
@@ -47,9 +50,9 @@ export default class JanInferenceOpenAIExtension extends BaseExtension {
    * Subscribes to events emitted by the @janhq/core package.
    */
   async onLoad() {
-    if (!(await fs.existsSync(JanInferenceOpenAIExtension._homeDir))) {
+    if (!(await fs.existsSync(JanInferenceOpenAIExtension._engineDir))) {
       await fs
-        .mkdirSync(JanInferenceOpenAIExtension._homeDir)
+        .mkdirSync(JanInferenceOpenAIExtension._engineDir)
         .catch((err) => console.debug(err));
     }
 
@@ -57,7 +60,7 @@ export default class JanInferenceOpenAIExtension extends BaseExtension {
 
     // Events subscription
     events.on(MessageEvent.OnMessageSent, (data) =>
-      JanInferenceOpenAIExtension.handleMessageRequest(data, this)
+      JanInferenceOpenAIExtension.handleMessageRequest(data, this),
     );
 
     events.on(ModelEvent.OnModelInit, (model: OpenAIModel) => {
@@ -70,6 +73,20 @@ export default class JanInferenceOpenAIExtension extends BaseExtension {
     events.on(InferenceEvent.OnInferenceStopped, () => {
       JanInferenceOpenAIExtension.handleInferenceStopped(this);
     });
+
+    const settingsFilePath = await joinPath([
+      JanInferenceOpenAIExtension._engineDir,
+      JanInferenceOpenAIExtension._engineMetadataFileName,
+    ]);
+
+    events.on(
+      AppConfigurationEventName.OnConfigurationUpdate,
+      (settingsKey: string) => {
+        // Update settings on changes
+        if (settingsKey === settingsFilePath)
+          JanInferenceOpenAIExtension.writeDefaultEngineSettings();
+      },
+    );
   }
 
   /**
@@ -80,8 +97,8 @@ export default class JanInferenceOpenAIExtension extends BaseExtension {
   static async writeDefaultEngineSettings() {
     try {
       const engineFile = join(
-        JanInferenceOpenAIExtension._homeDir,
+        JanInferenceOpenAIExtension._engineDir,
-        JanInferenceOpenAIExtension._engineMetadataFileName
+        JanInferenceOpenAIExtension._engineMetadataFileName,
       );
       if (await fs.existsSync(engineFile)) {
         const engine = await fs.readFileSync(engineFile, "utf-8");
@@ -90,7 +107,7 @@ export default class JanInferenceOpenAIExtension extends BaseExtension {
       } else {
         await fs.writeFileSync(
           engineFile,
-          JSON.stringify(JanInferenceOpenAIExtension._engineSettings, null, 2)
+          JSON.stringify(JanInferenceOpenAIExtension._engineSettings, null, 2),
         );
       }
     } catch (err) {
@@ -98,7 +115,7 @@ export default class JanInferenceOpenAIExtension extends BaseExtension {
     }
   }
   private static async handleModelInit(model: OpenAIModel) {
-    if (model.engine !== "openai") {
+    if (model.engine !== InferenceEngine.openai) {
       return;
     } else {
       JanInferenceOpenAIExtension._currentModel = model;
@@ -116,7 +133,7 @@ export default class JanInferenceOpenAIExtension extends BaseExtension {
   }
 
   private static async handleInferenceStopped(
-    instance: JanInferenceOpenAIExtension
+    instance: JanInferenceOpenAIExtension,
   ) {
     instance.isCancelled = true;
     instance.controller?.abort();
@@ -130,7 +147,7 @@ export default class JanInferenceOpenAIExtension extends BaseExtension {
    */
   private static async handleMessageRequest(
     data: MessageRequest,
-    instance: JanInferenceOpenAIExtension
+    instance: JanInferenceOpenAIExtension,
   ) {
     if (data.model.engine !== "openai") {
       return;
@@ -160,7 +177,7 @@ export default class JanInferenceOpenAIExtension extends BaseExtension {
         ...JanInferenceOpenAIExtension._currentModel,
         parameters: data.model.parameters,
       },
-      instance.controller
+      instance.controller,
     ).subscribe({
       next: (content) => {
         const messageContent: ThreadContent = {
@@ -181,7 +198,7 @@ export default class JanInferenceOpenAIExtension extends BaseExtension {
       },
       error: async (err) => {
         if (instance.isCancelled || message.content.length > 0) {
-          message.status = MessageStatus.Error;
+          message.status = MessageStatus.Stopped;
           events.emit(MessageEvent.OnMessageUpdate, message);
           return;
         }
@@ -193,7 +210,7 @@ export default class JanInferenceOpenAIExtension extends BaseExtension {
           },
         };
         message.content = [messageContent];
-        message.status = MessageStatus.Ready;
+        message.status = MessageStatus.Error;
         events.emit(MessageEvent.OnMessageUpdate, message);
       },
     });

extensions/inference-openai-extension/tsconfig.json

@@ -3,13 +3,12 @@
     "target": "es2016",
     "module": "ES6",
     "moduleResolution": "node",
-
     "outDir": "./dist",
     "esModuleInterop": true,
     "forceConsistentCasingInFileNames": true,
     "strict": false,
     "skipLibCheck": true,
-    "rootDir": "./src"
+    "rootDir": "./src",
   },
-  "include": ["./src"]
+  "include": ["./src"],
 }

extensions/inference-triton-trtllm-extension/tsconfig.json

@@ -3,13 +3,12 @@
     "target": "es2016",
     "module": "ES6",
     "moduleResolution": "node",
-
     "outDir": "./dist",
     "esModuleInterop": true,
     "forceConsistentCasingInFileNames": true,
     "strict": false,
     "skipLibCheck": true,
-    "rootDir": "./src"
+    "rootDir": "./src",
   },
-  "include": ["./src"]
+  "include": ["./src"],
 }

extensions/model-extension/src/index.ts

@@ -286,6 +286,7 @@ export default class JanModelExtension extends ModelExtension {
    * model.json file associated with it.
    *
    * This function will create a model.json file for the model.
+   * It works only with single binary file model.
    *
    * @param dirName the director which reside in ~/jan/models but does not have model.json file.
    */
@@ -302,9 +303,7 @@ export default class JanModelExtension extends ModelExtension {
     let binaryFileSize: number | undefined = undefined
 
     for (const file of files) {
-      if (file.endsWith(JanModelExtension._incompletedModelFileName)) continue
+      if (file.endsWith(JanModelExtension._supportedModelFormat)) {
-      if (file.endsWith('.json')) continue
-
         const path = await joinPath([JanModelExtension._homeDir, dirName, file])
         const fileStats = await fs.fileStat(path)
         if (fileStats.isDirectory) continue
@@ -312,13 +311,14 @@ export default class JanModelExtension extends ModelExtension {
         binaryFileName = file
         break
       }
+    }
 
     if (!binaryFileName) {
       console.warn(`Unable to find binary file for model ${dirName}`)
       return
     }
 
-    const defaultModel = await this.getDefaultModel()
+    const defaultModel = await this.getDefaultModel() as Model
     if (!defaultModel) {
       console.error('Unable to find default model')
       return
@@ -326,8 +326,19 @@ export default class JanModelExtension extends ModelExtension {
 
     const model: Model = {
       ...defaultModel,
+      // Overwrite default N/A fields
       id: dirName,
       name: dirName,
+      sources: [
+        {
+          url: binaryFileName,
+          filename: binaryFileName,
+        },
+      ],
+      settings: {
+        ...defaultModel.settings,
+        llama_model_path: binaryFileName,
+      },
       created: Date.now(),
       description: `${dirName} - user self import model`,
       metadata: {

extensions/monitoring-extension/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@janhq/monitoring-extension",
-  "version": "1.0.9",
+  "version": "1.0.10",
   "description": "This extension provides system health and OS level data",
   "main": "dist/index.js",
   "module": "dist/module.js",
@@ -26,6 +26,7 @@
     "README.md"
   ],
   "bundleDependencies": [
-    "node-os-utils"
+    "node-os-utils",
+    "@janhq/core"
   ]
 }

extensions/monitoring-extension/src/index.ts

@@ -1,5 +1,4 @@
-import { MonitoringExtension } from "@janhq/core";
+import { MonitoringExtension, executeOnMain } from "@janhq/core";
-import { executeOnMain } from "@janhq/core";
 
 /**
  * JanMonitoringExtension is a extension that provides system monitoring functionality.

extensions/monitoring-extension/src/module.ts

@@ -1,4 +1,14 @@
 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) => {
@@ -16,14 +26,44 @@ const getResourcesInfo = () =>
   });
 
 const getCurrentLoad = () =>
-  new Promise((resolve) => {
+  new Promise((resolve, reject) => {
     nodeOsUtils.cpu.usage().then((cpuPercentage) => {
-      const response = {
+      let data = {
-        cpu: {
+        run_mode: "cpu",
-          usage: cpuPercentage,
+        gpus_in_use: [],
-        },
       };
-      resolve(response);
+      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 !== "") {
+          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, stderr) => {
+              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: [] });
+      }
     });
   });
 

server/package.json

@@ -26,6 +26,8 @@
     "dotenv": "^16.3.1",
     "fastify": "^4.24.3",
     "request": "^2.88.2",
+    "fetch-retry": "^5.0.6",
+    "tcp-port-used": "^1.0.2",
     "request-progress": "^3.0.0"
   },
   "devDependencies": {
@@ -35,6 +37,7 @@
     "@typescript-eslint/parser": "^6.7.3",
     "eslint-plugin-react": "^7.33.2",
     "run-script-os": "^1.1.6",
+    "@types/tcp-port-used": "^1.0.4",
     "typescript": "^5.2.2"
   }
 }

uikit/package.json

@@ -18,6 +18,7 @@
   },
   "dependencies": {
     "@radix-ui/react-avatar": "^1.0.4",
+    "@radix-ui/react-checkbox": "^1.0.4",
     "@radix-ui/react-context": "^1.0.1",
     "@radix-ui/react-dialog": "^1.0.5",
     "@radix-ui/react-icons": "^1.3.0",

uikit/src/button/styles.scss

@@ -9,7 +9,7 @@
   }
 
   &-secondary-blue {
-    @apply bg-blue-200 text-blue-600 hover:bg-blue-500/50;
+    @apply bg-blue-200 text-blue-600 hover:bg-blue-300/50 dark:hover:bg-blue-200/80;
   }
 
   &-danger {
@@ -17,7 +17,7 @@
   }
 
   &-secondary-danger {
-    @apply bg-red-200 text-red-600 hover:bg-red-500/50;
+    @apply bg-red-200 text-red-600 hover:bg-red-300/50 dark:hover:bg-red-200/80;
   }
 
   &-outline {
@@ -67,14 +67,18 @@
 [type='submit'] {
   &.btn-primary {
     @apply bg-primary hover:bg-primary/90;
+    @apply disabled:pointer-events-none disabled:bg-zinc-100 disabled:text-zinc-400;
   }
   &.btn-secondary {
     @apply bg-secondary hover:bg-secondary/80;
+    @apply disabled:pointer-events-none disabled:bg-zinc-100 disabled:text-zinc-400;
   }
   &.btn-secondary-blue {
     @apply bg-blue-200 text-blue-900 hover:bg-blue-200/80;
+    @apply disabled:pointer-events-none disabled:bg-zinc-100 disabled:text-zinc-400;
   }
   &.btn-danger {
     @apply bg-danger hover:bg-danger/90;
+    @apply disabled:pointer-events-none disabled:bg-zinc-100 disabled:text-zinc-400;
   }
 }

uikit/src/checkbox/index.tsx

@@ -0,0 +1,29 @@
+'use client'
+
+import * as React from 'react'
+import * as CheckboxPrimitive from '@radix-ui/react-checkbox'
+import { CheckIcon } from '@radix-ui/react-icons'
+
+import { twMerge } from 'tailwind-merge'
+
+const Checkbox = React.forwardRef<
+  React.ElementRef<typeof CheckboxPrimitive.Root>,
+  React.ComponentPropsWithoutRef<typeof CheckboxPrimitive.Root>
+>(({ className, ...props }, ref) => (
+  <CheckboxPrimitive.Root
+    ref={ref}
+    className={twMerge('checkbox', className)}
+    {...props}
+  >
+    <CheckboxPrimitive.Indicator
+      className={twMerge(
+        'flex flex-shrink-0 items-center justify-center text-current'
+      )}
+    >
+      <CheckIcon className="checkbox--icon" />
+    </CheckboxPrimitive.Indicator>
+  </CheckboxPrimitive.Root>
+))
+Checkbox.displayName = CheckboxPrimitive.Root.displayName
+
+export { Checkbox }

uikit/src/checkbox/styles.scss

@@ -0,0 +1,7 @@
+.checkbox {
+  @apply border-border data-[state=checked]:bg-primary h-5 w-5 flex-shrink-0 rounded-md border data-[state=checked]:text-white;
+
+  &--icon {
+    @apply h-4 w-4;
+  }
+}

uikit/src/index.ts

@@ -12,3 +12,4 @@ export * from './command'
 export * from './textarea'
 export * from './select'
 export * from './slider'
+export * from './checkbox'

uikit/src/input/styles.scss

@@ -1,6 +1,6 @@
 .input {
   @apply border-border placeholder:text-muted-foreground flex h-9 w-full rounded-lg border bg-transparent px-3 py-1 transition-colors;
-  @apply disabled:cursor-not-allowed disabled:bg-zinc-100;
+  @apply disabled:cursor-not-allowed disabled:bg-zinc-100 disabled:dark:bg-zinc-800 disabled:dark:text-zinc-600;
   @apply focus-within:outline-none focus-visible:outline-0 focus-visible:ring-2 focus-visible:ring-blue-500 focus-visible:ring-offset-1;
   @apply file:border-0 file:bg-transparent file:font-medium;
 }

uikit/src/main.scss

@@ -16,6 +16,7 @@
 @import './textarea/styles.scss';
 @import './select/styles.scss';
 @import './slider/styles.scss';
+@import './checkbox/styles.scss';
 
 .animate-spin {
   animation: spin 1s linear infinite;

uikit/src/select/styles.scss

@@ -1,5 +1,6 @@
 .select {
-  @apply placeholder:text-muted-foreground border-border flex h-9 w-full items-center justify-between whitespace-nowrap rounded-md border bg-transparent px-3 py-2 text-sm shadow-sm disabled:cursor-not-allowed disabled:opacity-50 [&>span]:line-clamp-1;
+  @apply placeholder:text-muted-foreground border-border flex h-9 w-full items-center justify-between whitespace-nowrap rounded-md border bg-transparent px-3 py-2 text-sm shadow-sm disabled:cursor-not-allowed [&>span]:line-clamp-1;
+  @apply disabled:cursor-not-allowed disabled:bg-zinc-100 disabled:dark:bg-zinc-800 disabled:dark:text-zinc-600;
   @apply focus-within:outline-none focus-visible:outline-0 focus-visible:ring-2 focus-visible:ring-blue-500 focus-visible:ring-offset-1;
 
   &-caret {

web/containers/CardSidebar/index.tsx

@@ -13,10 +13,13 @@ import { useClickOutside } from '@/hooks/useClickOutside'
 
 import { usePath } from '@/hooks/usePath'
 
+import { openFileTitle } from '@/utils/titleUtils'
+
 import { activeThreadAtom } from '@/helpers/atoms/Thread.atom'
 
 interface Props {
   children: ReactNode
+  rightAction?: ReactNode
   title: string
   asChild?: boolean
   hideMoreVerticalAction?: boolean
@@ -25,6 +28,7 @@ export default function CardSidebar({
   children,
   title,
   asChild,
+  rightAction,
   hideMoreVerticalAction,
 }: Props) {
   const [show, setShow] = useState(true)
@@ -36,13 +40,6 @@ export default function CardSidebar({
 
   useClickOutside(() => setMore(false), null, [menu, toggle])
 
-  let openFolderTitle: string = 'Open Containing Folder'
-  if (isMac) {
-    openFolderTitle = 'Show in Finder'
-  } else if (isWindows) {
-    openFolderTitle = 'Show in File Explorer'
-  }
-
   return (
     <div
       className={twMerge(
@@ -53,27 +50,16 @@ export default function CardSidebar({
       <div
         className={twMerge(
           'relative flex items-center justify-between pl-4',
-          show && 'border-b border-border'
+          show && children && 'border-b border-border'
         )}
       >
-        <span className="font-bold">{title}</span>
+        <div className="flex items-center ">
-        <div className="flex">
-          {!asChild && (
-            <>
-              {!hideMoreVerticalAction && (
-                <div
-                  ref={setToggle}
-                  className="cursor-pointer rounded-lg bg-zinc-100 p-2 pr-0 dark:bg-zinc-900"
-                  onClick={() => setMore(!more)}
-                >
-                  <MoreVerticalIcon className="h-5 w-5" />
-                </div>
-              )}
-            </>
-          )}
           <button
-            onClick={() => setShow(!show)}
+            onClick={() => {
-            className="flex w-full flex-1 items-center space-x-2 rounded-lg bg-zinc-100 px-3 py-2 dark:bg-zinc-900"
+              if (!children) return
+              setShow(!show)
+            }}
+            className="flex w-full flex-1 items-center space-x-2 rounded-lg bg-zinc-100 py-2 pr-2 dark:bg-zinc-900"
           >
             <ChevronDownIcon
               className={twMerge(
@@ -82,6 +68,23 @@ export default function CardSidebar({
               )}
             />
           </button>
+          <span className="font-bold">{title}</span>
+        </div>
+        <div className="flex">
+          {rightAction && rightAction}
+          {!asChild && (
+            <>
+              {!hideMoreVerticalAction && (
+                <div
+                  ref={setToggle}
+                  className="cursor-pointer rounded-lg bg-zinc-100 p-2 px-3 dark:bg-zinc-900"
+                  onClick={() => setMore(!more)}
+                >
+                  <MoreVerticalIcon className="h-5 w-5" />
+                </div>
+              )}
+            </>
+          )}
         </div>
 
         {more && (
@@ -110,7 +113,7 @@ export default function CardSidebar({
                 {title === 'Model' ? (
                   <div className="flex flex-col">
                     <span className="font-medium text-black dark:text-muted-foreground">
-                      {openFolderTitle}
+                      {openFileTitle()}
                     </span>
                     <span className="mt-1 text-muted-foreground">
                       Opens thread.json. Changes affect this thread only.
@@ -118,7 +121,7 @@ export default function CardSidebar({
                   </div>
                 ) : (
                   <span className="text-bold text-black dark:text-muted-foreground">
-                    Show in Finder
+                    {openFileTitle()}
                   </span>
                 )}
               </>

web/containers/Checkbox/index.tsx

@@ -9,54 +9,26 @@ import {
   TooltipTrigger,
 } from '@janhq/uikit'
 
-import { useAtomValue, useSetAtom } from 'jotai'
 import { InfoIcon } from 'lucide-react'
 
-import { useActiveModel } from '@/hooks/useActiveModel'
-import useUpdateModelParameters from '@/hooks/useUpdateModelParameters'
-
-import { getConfigurationsData } from '@/utils/componentSettings'
-import { toSettingParams } from '@/utils/modelParam'
-
-import { serverEnabledAtom } from '@/helpers/atoms/LocalServer.atom'
-import {
-  engineParamsUpdateAtom,
-  getActiveThreadIdAtom,
-  getActiveThreadModelParamsAtom,
-} from '@/helpers/atoms/Thread.atom'
-
 type Props = {
   name: string
   title: string
+  enabled?: boolean
   description: string
   checked: boolean
+  onValueChanged?: (e: string | number | boolean) => void
 }
 
-const Checkbox: React.FC<Props> = ({ name, title, checked, description }) => {
+const Checkbox: React.FC<Props> = ({
-  const { updateModelParameter } = useUpdateModelParameters()
+  title,
-  const threadId = useAtomValue(getActiveThreadIdAtom)
+  checked,
-
+  enabled = true,
-  const activeModelParams = useAtomValue(getActiveThreadModelParamsAtom)
+  description,
-
+  onValueChanged,
-  const modelSettingParams = toSettingParams(activeModelParams)
+}) => {
-
-  const engineParams = getConfigurationsData(modelSettingParams)
-
-  const setEngineParamsUpdate = useSetAtom(engineParamsUpdateAtom)
-
-  const serverEnabled = useAtomValue(serverEnabledAtom)
-
-  const { stopModel } = useActiveModel()
-
   const onCheckedChange = (checked: boolean) => {
-    if (!threadId) return
+    onValueChanged?.(checked)
-    if (engineParams.some((x) => x.name.includes(name))) {
-      setEngineParamsUpdate(true)
-      stopModel()
-    } else {
-      setEngineParamsUpdate(false)
-    }
-    updateModelParameter(threadId, name, checked)
   }
 
   return (
@@ -80,7 +52,7 @@ const Checkbox: React.FC<Props> = ({ name, title, checked, description }) => {
       <Switch
         checked={checked}
         onCheckedChange={onCheckedChange}
-        disabled={serverEnabled}
+        disabled={!enabled}
       />
     </div>
   )

web/containers/DropdownListSidebar/index.tsx

@@ -26,6 +26,8 @@ import { useMainViewState } from '@/hooks/useMainViewState'
 
 import useRecommendedModel from '@/hooks/useRecommendedModel'
 
+import useUpdateModelParameters from '@/hooks/useUpdateModelParameters'
+
 import { toGibibytes } from '@/utils/converter'
 
 import ModelLabel from '../ModelLabel'
@@ -34,68 +36,40 @@ import OpenAiKeyInput from '../OpenAiKeyInput'
 import { serverEnabledAtom } from '@/helpers/atoms/LocalServer.atom'
 
 import {
-  ModelParams,
   activeThreadAtom,
-  getActiveThreadIdAtom,
   setThreadModelParamsAtom,
-  threadStatesAtom,
 } from '@/helpers/atoms/Thread.atom'
 
 export const selectedModelAtom = atom<Model | undefined>(undefined)
 
-export default function DropdownListSidebar() {
+// TODO: Move all of the unscoped logics outside of the component
-  const activeThreadId = useAtomValue(getActiveThreadIdAtom)
+const DropdownListSidebar = ({
+  strictedThread = true,
+}: {
+  strictedThread?: boolean
+}) => {
   const activeThread = useAtomValue(activeThreadAtom)
-  const threadStates = useAtomValue(threadStatesAtom)
   const [selectedModel, setSelectedModel] = useAtom(selectedModelAtom)
   const setThreadModelParams = useSetAtom(setThreadModelParamsAtom)
-  const { activeModel, stateModel } = useActiveModel()
+
+  const { stateModel } = useActiveModel()
   const [serverEnabled, setServerEnabled] = useAtom(serverEnabledAtom)
   const { setMainViewState } = useMainViewState()
-
+  const [loader, setLoader] = useState(0)
   const { recommendedModel, downloadedModels } = useRecommendedModel()
-
+  const { updateModelParameter } = useUpdateModelParameters()
-  /**
-   * Default value for max_tokens and ctx_len
-   * Its to avoid OOM issue since a model can set a big number for these settings
-   */
-  const defaultValue = (value?: number) => {
-    if (value && value < 4096) return value
-    return 4096
-  }
 
   useEffect(() => {
-    setSelectedModel(selectedModel || activeModel || recommendedModel)
+    if (!activeThread) return
 
-    if (activeThread) {
+    let model = downloadedModels.find(
-      const finishInit = threadStates[activeThread.id].isFinishInit ?? true
+      (model) => model.id === activeThread.assistants[0].model.id
-      if (finishInit) return
+    )
-      const modelParams: ModelParams = {
+    if (!model) {
-        ...recommendedModel?.parameters,
+      model = recommendedModel
-        ...recommendedModel?.settings,
-        /**
-         * This is to set default value for these settings instead of maximum value
-         * Should only apply when model.json has these settings
-         */
-        ...(recommendedModel?.parameters.max_tokens && {
-          max_tokens: defaultValue(recommendedModel?.parameters.max_tokens),
-        }),
-        ...(recommendedModel?.settings.ctx_len && {
-          ctx_len: defaultValue(recommendedModel?.settings.ctx_len),
-        }),
     }
-      setThreadModelParams(activeThread.id, modelParams)
+    setSelectedModel(model)
-    }
+  }, [recommendedModel, activeThread, downloadedModels, setSelectedModel])
-    // eslint-disable-next-line react-hooks/exhaustive-deps
-  }, [
-    recommendedModel,
-    activeThread,
-    setSelectedModel,
-    setThreadModelParams,
-    threadStates,
-  ])
-
-  const [loader, setLoader] = useState(0)
 
   // This is fake loader please fix this when we have realtime percentage when load model
   useEffect(() => {
@@ -132,25 +106,35 @@ export default function DropdownListSidebar() {
         setServerEnabled(false)
       }
 
-      if (activeThreadId) {
+      if (activeThread) {
         const modelParams = {
           ...model?.parameters,
           ...model?.settings,
         }
-        setThreadModelParams(activeThreadId, modelParams)
+        // Update model paramter to the thread state
+        setThreadModelParams(activeThread.id, modelParams)
+
+        // Update model parameter to the thread file
+        if (model)
+          updateModelParameter(activeThread.id, {
+            params: modelParams,
+            modelId: model.id,
+            engine: model.engine,
+          })
       }
     },
-    // eslint-disable-next-line react-hooks/exhaustive-deps
     [
       downloadedModels,
       serverEnabled,
-      activeThreadId,
+      activeThread,
-      activeModel,
+      setSelectedModel,
+      setServerEnabled,
       setThreadModelParams,
+      updateModelParameter,
     ]
   )
 
-  if (!activeThread) {
+  if (strictedThread && !activeThread) {
     return null
   }
 
@@ -236,10 +220,9 @@ export default function DropdownListSidebar() {
         </Select>
       </div>
 
-      <OpenAiKeyInput
+      <OpenAiKeyInput />
-        selectedModel={selectedModel}
-        serverEnabled={serverEnabled}
-      />
     </>
   )
 }
+
+export default DropdownListSidebar

web/containers/Layout/BottomBar/index.tsx

@@ -26,11 +26,12 @@ import { MainViewState } from '@/constants/screens'
 import { useActiveModel } from '@/hooks/useActiveModel'
 
 import { useDownloadState } from '@/hooks/useDownloadState'
-import { useGetDownloadedModels } from '@/hooks/useGetDownloadedModels'
+
 import useGetSystemResources from '@/hooks/useGetSystemResources'
 import { useMainViewState } from '@/hooks/useMainViewState'
 
 import { serverEnabledAtom } from '@/helpers/atoms/LocalServer.atom'
+import { downloadedModelsAtom } from '@/helpers/atoms/Model.atom'
 
 const menuLinks = [
   {
@@ -47,14 +48,22 @@ const menuLinks = [
 
 const BottomBar = () => {
   const { activeModel, stateModel } = useActiveModel()
-  const { ram, cpu } = useGetSystemResources()
+  const { ram, cpu, gpus } = useGetSystemResources()
   const progress = useAtomValue(appDownloadProgress)
-  const { downloadedModels } = useGetDownloadedModels()
+  const downloadedModels = useAtomValue(downloadedModelsAtom)
+
   const { setMainViewState } = useMainViewState()
   const { downloadStates } = useDownloadState()
   const setShowSelectModelModal = useSetAtom(showSelectModelModalAtom)
   const [serverEnabled] = useAtom(serverEnabledAtom)
 
+  const calculateGpuMemoryUsage = (gpu: Record<string, never>) => {
+    const total = parseInt(gpu.memoryTotal)
+    const free = parseInt(gpu.memoryFree)
+    if (!total || !free) return 0
+    return Math.round(((total - free) / total) * 100)
+  }
+
   return (
     <div className="fixed bottom-0 left-16 z-20 flex h-12 w-[calc(100%-64px)] items-center justify-between border-t border-border bg-background/80 px-3">
       <div className="flex flex-shrink-0 items-center gap-x-2">
@@ -117,6 +126,17 @@ const BottomBar = () => {
           <SystemItem name="CPU:" value={`${cpu}%`} />
           <SystemItem name="Mem:" value={`${ram}%`} />
         </div>
+        {gpus.length > 0 && (
+          <div className="flex items-center gap-x-2">
+            {gpus.map((gpu, index) => (
+              <SystemItem
+                key={index}
+                name={`GPU ${gpu.id}:`}
+                value={`${gpu.utilization}% Util, ${calculateGpuMemoryUsage(gpu)}% Mem`}
+              />
+            ))}
+          </div>
+        )}
         {/* VERSION is defined by webpack, please see next.config.js */}
         <span className="text-xs text-muted-foreground">
           Jan v{VERSION ?? ''}

web/containers/Layout/TopBar/CommandListDownloadedModel/index.tsx

@@ -11,7 +11,7 @@ import {
   Badge,
 } from '@janhq/uikit'
 
-import { useAtom } from 'jotai'
+import { useAtom, useAtomValue } from 'jotai'
 import { DatabaseIcon, CpuIcon } from 'lucide-react'
 
 import { showSelectModelModalAtom } from '@/containers/Providers/KeyListener'
@@ -19,14 +19,14 @@ import { showSelectModelModalAtom } from '@/containers/Providers/KeyListener'
 import { MainViewState } from '@/constants/screens'
 
 import { useActiveModel } from '@/hooks/useActiveModel'
-import { useGetDownloadedModels } from '@/hooks/useGetDownloadedModels'
 import { useMainViewState } from '@/hooks/useMainViewState'
 
 import { serverEnabledAtom } from '@/helpers/atoms/LocalServer.atom'
+import { downloadedModelsAtom } from '@/helpers/atoms/Model.atom'
 
 export default function CommandListDownloadedModel() {
   const { setMainViewState } = useMainViewState()
-  const { downloadedModels } = useGetDownloadedModels()
+  const downloadedModels = useAtomValue(downloadedModelsAtom)
   const { activeModel, startModel, stopModel } = useActiveModel()
   const [serverEnabled] = useAtom(serverEnabledAtom)
   const [showSelectModelModal, setShowSelectModelModal] = useAtom(

web/containers/Layout/TopBar/index.tsx

@@ -20,20 +20,22 @@ import { MainViewState } from '@/constants/screens'
 
 import { useClickOutside } from '@/hooks/useClickOutside'
 import { useCreateNewThread } from '@/hooks/useCreateNewThread'
-import useGetAssistants, { getAssistants } from '@/hooks/useGetAssistants'
 import { useMainViewState } from '@/hooks/useMainViewState'
 
 import { usePath } from '@/hooks/usePath'
 
 import { showRightSideBarAtom } from '@/screens/Chat/Sidebar'
 
+import { openFileTitle } from '@/utils/titleUtils'
+
+import { assistantsAtom } from '@/helpers/atoms/Assistant.atom'
 import { activeThreadAtom } from '@/helpers/atoms/Thread.atom'
 
 const TopBar = () => {
   const activeThread = useAtomValue(activeThreadAtom)
   const { mainViewState } = useMainViewState()
   const { requestCreateNewThread } = useCreateNewThread()
-  const { assistants } = useGetAssistants()
+  const assistants = useAtomValue(assistantsAtom)
   const [showRightSideBar, setShowRightSideBar] = useAtom(showRightSideBarAtom)
   const [showLeftSideBar, setShowLeftSideBar] = useAtom(showLeftSideBarAtom)
   const showing = useAtomValue(showRightSideBarAtom)
@@ -59,12 +61,7 @@ const TopBar = () => {
 
   const onCreateConversationClick = async () => {
     if (assistants.length === 0) {
-      const res = await getAssistants()
-      if (res.length === 0) {
       alert('No assistant available')
-        return
-      }
-      requestCreateNewThread(res[0])
     } else {
       requestCreateNewThread(assistants[0])
     }
@@ -126,7 +123,8 @@ const TopBar = () => {
               showing && 'border-l border-border'
             )}
           >
-            {activeThread && (
+            {((activeThread && mainViewState === MainViewState.Thread) ||
+              mainViewState === MainViewState.LocalServer) && (
               <div className="flex h-full w-52 items-center justify-between px-4">
                 {showing && (
                   <div className="relative flex h-full items-center">
@@ -161,7 +159,7 @@ const TopBar = () => {
                             className="text-muted-foreground"
                           />
                           <span className="font-medium text-black dark:text-muted-foreground">
-                            Show in Finder
+                            {openFileTitle()}
                           </span>
                         </div>
                         <div
@@ -206,7 +204,7 @@ const TopBar = () => {
                           />
                           <div className="flex flex-col">
                             <span className="font-medium text-black dark:text-muted-foreground">
-                              Show in Finder
+                              {openFileTitle()}
                             </span>
                           </div>
                         </div>

web/containers/Layout/index.tsx

@@ -12,7 +12,8 @@ import TopBar from '@/containers/Layout/TopBar'
 import { MainViewState } from '@/constants/screens'
 
 import { useMainViewState } from '@/hooks/useMainViewState'
-import { SUCCESS_SET_NEW_DESTINATION } from '@/hooks/useVaultDirectory'
+
+import { SUCCESS_SET_NEW_DESTINATION } from '@/screens/Settings/Advanced/DataFolder'
 
 const BaseLayout = (props: PropsWithChildren) => {
   const { children } = props
@@ -28,7 +29,7 @@ const BaseLayout = (props: PropsWithChildren) => {
     if (localStorage.getItem(SUCCESS_SET_NEW_DESTINATION) === 'true') {
       setMainViewState(MainViewState.Settings)
     }
-  }, [])
+  }, [setMainViewState])
 
   return (
     <div className="flex h-screen w-screen flex-1 overflow-hidden">

web/containers/Loader/GenerateResponse.tsx

@@ -0,0 +1,39 @@
+import React, { useEffect, useState } from 'react'
+
+export default function GenerateResponse() {
+  const [loader, setLoader] = useState(0)
+
+  // This is fake loader please fix this when we have realtime percentage when load model
+  useEffect(() => {
+    if (loader === 24) {
+      setTimeout(() => {
+        setLoader(loader + 1)
+      }, 250)
+    } else if (loader === 50) {
+      setTimeout(() => {
+        setLoader(loader + 1)
+      }, 250)
+    } else if (loader === 78) {
+      setTimeout(() => {
+        setLoader(loader + 1)
+      }, 250)
+    } else if (loader === 85) {
+      setLoader(85)
+    } else {
+      setLoader(loader + 1)
+    }
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [loader])
+
+  return (
+    <div className=" mb-1 mt-2 py-2 text-center">
+      <div className="relative inline-block overflow-hidden rounded-lg border border-neutral-50 bg-gray-50 px-4 py-2 font-semibold text-gray-600 shadow-lg">
+        <div
+          className="absolute left-0 top-0 h-full bg-gray-200"
+          style={{ width: `${loader}%` }}
+        />
+        <span className="relative z-10">Generating response...</span>
+      </div>
+    </div>
+  )
+}