Merge branch 'main' into dev

# Conflicts:
#	web/screens/Chat/ErrorMessage/index.tsx

core/src/browser/extensions/engines/OAIEngine.ts

@@ -102,12 +102,20 @@ export abstract class OAIEngine extends AIEngine {
         events.emit(MessageEvent.OnMessageUpdate, message)
       },
       error: async (err: any) => {
+        console.error(`Inference error: ${JSON.stringify(err, null, 2)}`)
         if (this.isCancelled || message.content.length) {
           message.status = MessageStatus.Stopped
           events.emit(MessageEvent.OnMessageUpdate, message)
           return
         }
         message.status = MessageStatus.Error
+        message.content[0] = {
+          type: ContentType.Text,
+          text: {
+            value: err.message,
+            annotations: [],
+          },
+        }
         message.error_code = err.code
         events.emit(MessageEvent.OnMessageUpdate, message)
       },

core/src/types/message/messageEntity.ts

@@ -83,6 +83,8 @@ export enum MessageStatus {
 export enum ErrorCode {
   InvalidApiKey = 'invalid_api_key',
 
+  InsufficientQuota = 'insufficient_quota',
+
   Unknown = 'unknown',
 }
 

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

@@ -18,7 +18,7 @@ export class RetrievalTool extends InferenceTool {
     tool?: AssistantTool
   ): Promise<MessageRequest> {
     if (!data.model || !data.messages) {
-      return Promise.resolve(this.normalize(data))
+      return Promise.resolve(data)
     }
 
     const latestMessage = data.messages[data.messages.length - 1]
@@ -38,6 +38,8 @@ export class RetrievalTool extends InferenceTool {
           docFile,
           data.model?.engine
         )
+      } else {
+        return Promise.resolve(data)
       }
     } else if (
       // Check whether we need to ingest document or not
@@ -48,7 +50,7 @@ export class RetrievalTool extends InferenceTool {
     ) {
       // No document ingested, reroute the result to inference engine
 
-      return Promise.resolve(this.normalize(data))
+      return Promise.resolve(data)
     }
     // 2. Load agent on thread changed
     if (this.retrievalThreadId !== data.threadId) {

extensions/model-extension/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@janhq/model-extension",
-  "version": "1.0.28",
+  "version": "1.0.30",
   "description": "Model Management Extension provides model exploration and seamless downloads",
   "main": "dist/index.js",
   "module": "dist/module.js",

models/mistral-ins-7b-q4/model.json

@@ -5,7 +5,7 @@
       "url": "https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-GGUF/resolve/main/mistral-7b-instruct-v0.2.Q4_K_M.gguf"
     }
   ],
-  "id": "mistral-7b",
+  "id": "mistral-ins-7b-q4",
   "object": "model",
   "name": "Mistral Instruct 7B Q4",
   "version": "1.0",
@@ -29,7 +29,7 @@
     "author": "MistralAI",
     "tags": ["Featured", "7B", "Foundational Model"],
     "size": 4370000000,
-    "cover": "https://raw.githubusercontent.com/janhq/jan/dev/models/mistral-7b/cover.png"
+    "cover": "https://raw.githubusercontent.com/janhq/jan/dev/models/mistral-ins-7b-q4/cover.png"
   },
   "engine": "nitro"
 }

models/stealth-v1.2-7b/model.json

@@ -5,7 +5,7 @@
       "url": "https://huggingface.co/janhq/stealth-v1.3-GGUF/resolve/main/stealth-v1.3.Q4_K_M.gguf"
     }
   ],
-  "id": "stealth-7b",
+  "id": "stealth-v1.2-7b",
   "object": "model",
   "name": "Stealth 7B Q4",
   "version": "1.0",

models/tinyllama-1.1b/model.json

@@ -0,0 +1,34 @@
+{
+  "sources": [
+    {
+      "filename": "tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf",
+      "url": "https://huggingface.co/TheBloke/TinyLlama-1.1B-Chat-v1.0-GGUF/resolve/main/tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf"
+    }
+  ],
+  "id": "tinyllama-1.1b",
+  "object": "model",
+  "name": "TinyLlama Chat 1.1B Q4",
+  "version": "1.0",
+  "description": "TinyLlama is a tiny model with only 1.1B. It's a good model for less powerful computers.",
+  "format": "gguf",
+  "settings": {
+    "ctx_len": 4096,
+    "prompt_template": "<|system|>\n{system_message}<|user|>\n{prompt}<|assistant|>",
+    "llama_model_path": "tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf"
+  },
+  "parameters": {
+    "temperature": 0.7,
+    "top_p": 0.95,
+    "stream": true,
+    "max_tokens": 2048,
+    "stop": [],
+    "frequency_penalty": 0,
+    "presence_penalty": 0
+  },
+  "metadata": {
+    "author": "TinyLlama",
+    "tags": ["Tiny", "Foundation Model"],
+    "size": 669000000
+  },
+  "engine": "nitro"
+}

models/trinity-v1.2-7b/model.json

@@ -5,7 +5,7 @@
       "url": "https://huggingface.co/janhq/trinity-v1.2-GGUF/resolve/main/trinity-v1.2.Q4_K_M.gguf"
     }
   ],
-  "id": "trinity-7b",
+  "id": "trinity-v1.2-7b",
   "object": "model",
   "name": "Trinity-v1.2 7B Q4",
   "version": "1.0",
@@ -28,7 +28,7 @@
     "author": "Jan",
     "tags": ["7B", "Merged", "Featured"],
     "size": 4370000000,
-    "cover": "https://raw.githubusercontent.com/janhq/jan/dev/models/trinity-7b/cover.png"
+    "cover": "https://raw.githubusercontent.com/janhq/jan/dev/models/trinity-v1.2-7b/cover.png"
   },
   "engine": "nitro"
 }

web/containers/AutoLink/index.tsx

@@ -0,0 +1,34 @@
+import React from 'react'
+
+type Props = {
+  text: string
+}
+
+const AutoLink: React.FC<Props> = ({ text }) => {
+  const delimiter =
+    /((?:https?:\/\/)?(?:(?:[a-z0-9]?(?:[a-z0-9-]{1,61}[a-z0-9])?\.[^.|\s])+[a-z.]*[a-z]+|(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)(?:\.(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)){3})(?::\d{1,5})*[a-z0-9.,_/~#&=;%+?\-\\(\\)]*)/gi
+
+  return (
+    <>
+      {text.split(delimiter).map((word) => {
+        const match = word.match(delimiter)
+        if (match) {
+          const url = match[0]
+          return (
+            <a
+              key={url}
+              target="blank"
+              href={url.startsWith('http') ? url : `http://${url}`}
+              className="text-primary dark:text-blue-400"
+            >
+              {url}
+            </a>
+          )
+        }
+        return word
+      })}
+    </>
+  )
+}
+
+export default React.memo(AutoLink)

web/hooks/useSendChatMessage.ts

@@ -234,6 +234,10 @@ export default function useSendChatMessage() {
       selectedModelRef.current?.id ??
       activeThreadRef.current.assistants[0].model.id
 
+    if (base64Blob) {
+      setFileUpload([])
+    }
+
     if (modelRef.current?.id !== modelId) {
       setQueuedMessage(true)
       const error = await startModel(modelId).catch((error: Error) => error)
@@ -261,10 +265,6 @@ export default function useSendChatMessage() {
     // Reset states
     setReloadModel(false)
     setEngineParamsUpdate(false)
-
-    if (base64Blob) {
-      setFileUpload([])
-    }
   }
 
   return {

web/screens/Chat/ErrorMessage/index.tsx

@@ -3,6 +3,7 @@ import { Button } from '@janhq/uikit'
 import { useAtomValue, useSetAtom } from 'jotai'
 import { RefreshCcw } from 'lucide-react'
 
+import AutoLink from '@/containers/AutoLink'
 import ModalTroubleShooting, {
   modalTroubleShootingAtom,
 } from '@/containers/ModalTroubleShoot'
@@ -110,9 +111,9 @@ const ErrorMessage = ({ message }: { message: ThreadMessage }) => {
           ) : (
             <div
               key={message.id}
-              className="flex flex-col items-center text-center text-sm font-medium text-gray-500"
+              className="mx-6 flex flex-col items-center space-y-2 text-center text-sm font-medium text-gray-500"
             >
-              {getErrorTitle()}
+              <AutoLink text={getErrorTitle()} />
               <p>
                 Jan’s in beta. Access&nbsp;
                 <span

web/screens/ExploreModels/ExploreModelItemHeader/index.tsx

@@ -1,4 +1,4 @@
-import { useCallback } from 'react'
+import { useCallback, useState } from 'react'
 
 import { Model } from '@janhq/core'
 import {
@@ -70,6 +70,7 @@ const ExploreModelItemHeader: React.FC<Props> = ({ model, onClick, open }) => {
   const { requestCreateNewThread } = useCreateNewThread()
   const totalRam = useAtomValue(totalRamAtom)
   const { settings } = useSettings()
+  const [imageLoaded, setImageLoaded] = useState(true)
 
   const nvidiaTotalVram = useAtomValue(nvidiaTotalVramAtom)
   const setMainViewState = useSetAtom(mainViewStateAtom)
@@ -143,9 +144,10 @@ const ExploreModelItemHeader: React.FC<Props> = ({ model, onClick, open }) => {
       className="cursor-pointer rounded-t-md bg-background"
       onClick={onClick}
     >
-      {model.metadata.cover && (
-        <div className="relative h-full w-full ">
+      {model.metadata.cover && imageLoaded && (
+        <div className="relative h-full w-full">
           <img
+            onError={() => setImageLoaded(false)}
             src={model.metadata.cover}
             className="h-[250px] w-full object-cover"
             alt={`Cover - ${model.id}`}