chore: load, unload model and inference synchronously

core/src/extensions/ai-engines/RemoteOAIEngine.ts

@@ -1,46 +0,0 @@
-import { events } from '../../events'
-import { Model, ModelEvent } from '../../types'
-import { OAIEngine } from './OAIEngine'
-
-/**
- * Base OAI Remote Inference Provider
- * Added the implementation of loading and unloading model (applicable to local inference providers)
- */
-export abstract class RemoteOAIEngine extends OAIEngine {
-  // The inference engine
-  abstract apiKey: string
-  /**
-   * On extension load, subscribe to events.
-   */
-  onLoad() {
-    super.onLoad()
-    // These events are applicable to local inference providers
-    events.on(ModelEvent.OnModelInit, (model: Model) => this.loadModel(model))
-    events.on(ModelEvent.OnModelStop, (model: Model) => this.unloadModel(model))
-  }
-
-  /**
-   * Load the model.
-   */
-  async loadModel(model: Model) {
-    if (model.engine.toString() !== this.provider) return
-    events.emit(ModelEvent.OnModelReady, model)
-  }
-  /**
-   * Stops the model.
-   */
-  unloadModel(model: Model) {
-    if (model.engine && model.engine.toString() !== this.provider) return
-    events.emit(ModelEvent.OnModelStopped, {})
-  }
-
-  /**
-   * Headers for the inference request
-   */
-  override headers(): HeadersInit {
-    return {
-      'Authorization': `Bearer ${this.apiKey}`,
-      'api-key': `${this.apiKey}`,
-    }
-  }
-}

core/src/extensions/engines/AIEngine.ts

@@ -2,7 +2,8 @@ import { getJanDataFolderPath, joinPath } from '../../core'
 import { events } from '../../events'
 import { BaseExtension } from '../../extension'
 import { fs } from '../../fs'
-import { Model, ModelEvent } from '../../types'
+import { MessageRequest, Model, ModelEvent } from '../../types'
+import { EngineManager } from './EngineManager'
 
 /**
  * Base AIEngine
@@ -11,30 +12,71 @@ import { Model, ModelEvent } from '../../types'
 export abstract class AIEngine extends BaseExtension {
   // The inference engine
   abstract provider: string
-  // The model folder
-  modelFolder: string = 'models'
 
+  /**
+   * On extension load, subscribe to events.
+   */
+  override onLoad() {
+    this.registerEngine()
+
+    events.on(ModelEvent.OnModelInit, (model: Model) => this.loadModel(model))
+    events.on(ModelEvent.OnModelStop, (model: Model) => this.unloadModel(model))
+
+    this.prePopulateModels()
+  }
+
+  /**
+   * Defines models
+   */
   models(): Promise<Model[]> {
     return Promise.resolve([])
   }
 
   /**
-   * On extension load, subscribe to events.
+   * Registers AI Engines
    */
-  onLoad() {
+  registerEngine() {
-    this.prePopulateModels()
+    EngineManager.instance()?.register(this)
   }
 
+  /**
+   * Loads the model.
+   */
+  async loadModel(model: Model): Promise<any> {
+    if (model.engine.toString() !== this.provider) return Promise.resolve()
+    events.emit(ModelEvent.OnModelReady, model)
+    return Promise.resolve()
+  }
+  /**
+   * Stops the model.
+   */
+  async unloadModel(model?: Model): Promise<any> {
+    if (model?.engine && model.engine.toString() !== this.provider) return Promise.resolve()
+    events.emit(ModelEvent.OnModelStopped, model ?? {})
+    return Promise.resolve()
+  }
+
+  /*
+   * Inference request
+   */
+  inference(data: MessageRequest) {}
+
+  /**
+   * Stop inference
+   */
+  stopInference() {}
+
   /**
    * Pre-populate models to App Data Folder
    */
   prePopulateModels(): Promise<void> {
+    const modelFolder = 'models'
     return this.models().then((models) => {
       const prePoluateOperations = models.map((model) =>
         getJanDataFolderPath()
           .then((janDataFolder) =>
             // Attempt to create the model folder
-            joinPath([janDataFolder, this.modelFolder, model.id]).then((path) =>
+            joinPath([janDataFolder, modelFolder, model.id]).then((path) =>
               fs
                 .mkdir(path)
                 .catch()

core/src/extensions/engines/EngineManager.ts

@@ -0,0 +1,34 @@
+import { log } from '../../core'
+import { AIEngine } from './AIEngine'
+
+/**
+ * Manages the registration and retrieval of inference engines.
+ */
+export class EngineManager {
+  public engines = new Map<string, AIEngine>()
+
+  /**
+   * Registers an engine.
+   * @param engine - The engine to register.
+   */
+  register<T extends AIEngine>(engine: T) {
+    this.engines.set(engine.provider, engine)
+  }
+
+  /**
+   * Retrieves a engine by provider.
+   * @param provider - The name of the engine to retrieve.
+   * @returns The engine, if found.
+   */
+  get<T extends AIEngine>(provider: string): T | undefined {
+    return this.engines.get(provider) as T | undefined
+  }
+
+  static instance(): EngineManager | undefined {
+    return window.core?.engineManager as EngineManager
+  }
+}
+
+/**
+ * The singleton instance of the ExtensionManager.
+ */

core/src/extensions/engines/LocalOAIEngine.ts

@@ -16,7 +16,7 @@ export abstract class LocalOAIEngine extends OAIEngine {
   /**
    * On extension load, subscribe to events.
    */
-  onLoad() {
+  override onLoad() {
     super.onLoad()
     // These events are applicable to local inference providers
     events.on(ModelEvent.OnModelInit, (model: Model) => this.loadModel(model))
@@ -26,10 +26,10 @@ export abstract class LocalOAIEngine extends OAIEngine {
   /**
    * Load the model.
    */
-  async loadModel(model: Model) {
+  override async loadModel(model: Model): Promise<void> {
     if (model.engine.toString() !== this.provider) return
-
+    const modelFolderName = 'models'
-    const modelFolder = await joinPath([await getJanDataFolderPath(), this.modelFolder, model.id])
+    const modelFolder = await joinPath([await getJanDataFolderPath(), modelFolderName, model.id])
     const systemInfo = await systemInformation()
     const res = await executeOnMain(
       this.nodeModule,
@@ -42,24 +42,22 @@ export abstract class LocalOAIEngine extends OAIEngine {
     )
 
     if (res?.error) {
-      events.emit(ModelEvent.OnModelFail, {
+      events.emit(ModelEvent.OnModelFail, { error: res.error })
-        ...model,
+      return Promise.reject(res.error)
-        error: res.error,
-      })
-      return
     } else {
       this.loadedModel = model
       events.emit(ModelEvent.OnModelReady, model)
+      return Promise.resolve()
     }
   }
   /**
    * Stops the model.
    */
-  unloadModel(model: Model) {
+  override async unloadModel(model?: Model): Promise<void> {
-    if (model.engine && model.engine?.toString() !== this.provider) return
+    if (model?.engine && model.engine?.toString() !== this.provider) return Promise.resolve()
-    this.loadedModel = undefined
 
-    executeOnMain(this.nodeModule, this.unloadModelFunctionName).then(() => {
+    this.loadedModel = undefined
+    return executeOnMain(this.nodeModule, this.unloadModelFunctionName).then(() => {
       events.emit(ModelEvent.OnModelStopped, {})
     })
   }

core/src/extensions/engines/OAIEngine.ts

@@ -34,7 +34,7 @@ export abstract class OAIEngine extends AIEngine {
   /**
    * On extension load, subscribe to events.
    */
-  onLoad() {
+  override onLoad() {
     super.onLoad()
     events.on(MessageEvent.OnMessageSent, (data: MessageRequest) => this.inference(data))
     events.on(InferenceEvent.OnInferenceStopped, () => this.stopInference())
@@ -43,12 +43,12 @@ export abstract class OAIEngine extends AIEngine {
   /**
    * On extension unload
    */
-  onUnload(): void {}
+  override onUnload(): void {}
 
   /*
    * Inference request
    */
-  inference(data: MessageRequest) {
+  override inference(data: MessageRequest) {
     if (data.model?.engine?.toString() !== this.provider) return
 
     const timestamp = Date.now()
@@ -114,7 +114,7 @@ export abstract class OAIEngine extends AIEngine {
   /**
    * Stops the inference.
    */
-  stopInference() {
+  override stopInference() {
     this.isCancelled = true
     this.controller?.abort()
   }

core/src/extensions/engines/RemoteOAIEngine.ts

@@ -0,0 +1,26 @@
+import { OAIEngine } from './OAIEngine'
+
+/**
+ * Base OAI Remote Inference Provider
+ * Added the implementation of loading and unloading model (applicable to local inference providers)
+ */
+export abstract class RemoteOAIEngine extends OAIEngine {
+  // The inference engine
+  abstract apiKey: string
+  /**
+   * On extension load, subscribe to events.
+   */
+  override onLoad() {
+    super.onLoad()
+  }
+
+  /**
+   * Headers for the inference request
+   */
+  override headers(): HeadersInit {
+    return {
+      'Authorization': `Bearer ${this.apiKey}`,
+      'api-key': `${this.apiKey}`,
+    }
+  }
+}

core/src/extensions/engines/index.ts

@@ -2,3 +2,4 @@ export * from './AIEngine'
 export * from './OAIEngine'
 export * from './LocalOAIEngine'
 export * from './RemoteOAIEngine'
+export * from './EngineManager'

core/src/extensions/index.ts

@@ -32,4 +32,4 @@ export { HuggingFaceExtension } from './huggingface'
 /**
  * Base AI Engines.
  */
-export * from './ai-engines'
+export * from './engines'

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

@@ -91,15 +91,14 @@ export default class JanInferenceNitroExtension extends LocalOAIEngine {
     return super.loadModel(model)
   }
 
-  override unloadModel(model: Model): void {
+  override async unloadModel(model?: Model) {
-    super.unloadModel(model)
+    if (model?.engine && model.engine !== this.provider) return
-
-    if (model.engine && model.engine !== this.provider) return
 
     // stop the periocally health check
     if (this.getNitroProcesHealthIntervalId) {
       clearInterval(this.getNitroProcesHealthIntervalId)
       this.getNitroProcesHealthIntervalId = undefined
     }
+    return super.unloadModel(model)
   }
 }

web/containers/Providers/EventHandler.tsx

@@ -8,26 +8,17 @@ import {
   ExtensionTypeEnum,
   MessageStatus,
   MessageRequest,
-  Model,
   ConversationalExtension,
   MessageEvent,
   MessageRequestType,
   ModelEvent,
   Thread,
-  ModelInitFailed,
+  EngineManager,
 } from '@janhq/core'
 import { useAtomValue, useSetAtom } from 'jotai'
 import { ulid } from 'ulidx'
 
-import {
+import { activeModelAtom, stateModelAtom } from '@/hooks/useActiveModel'
-  activeModelAtom,
-  loadModelErrorAtom,
-  stateModelAtom,
-} from '@/hooks/useActiveModel'
-
-import { queuedMessageAtom } from '@/hooks/useSendChatMessage'
-
-import { toaster } from '../Toast'
 
 import { extensionManager } from '@/extension'
 import {
@@ -51,8 +42,6 @@ export default function EventHandler({ children }: { children: ReactNode }) {
   const activeModel = useAtomValue(activeModelAtom)
   const setActiveModel = useSetAtom(activeModelAtom)
   const setStateModel = useSetAtom(stateModelAtom)
-  const setQueuedMessage = useSetAtom(queuedMessageAtom)
-  const setLoadModelError = useSetAtom(loadModelErrorAtom)
 
   const updateThreadWaiting = useSetAtom(updateThreadWaitingForResponseAtom)
   const threads = useAtomValue(threadsAtom)
@@ -88,44 +77,11 @@ export default function EventHandler({ children }: { children: ReactNode }) {
     [addNewMessage]
   )
 
-  const onModelReady = useCallback(
-    (model: Model) => {
-      setActiveModel(model)
-      toaster({
-        title: 'Success!',
-        description: `Model ${model.id} has been started.`,
-        type: 'success',
-      })
-      setStateModel(() => ({
-        state: 'stop',
-        loading: false,
-        model: model.id,
-      }))
-    },
-    [setActiveModel, setStateModel]
-  )
-
   const onModelStopped = useCallback(() => {
-    setTimeout(() => {
+    setActiveModel(undefined)
-      setActiveModel(undefined)
+    setStateModel({ state: 'start', loading: false, model: '' })
-      setStateModel({ state: 'start', loading: false, model: '' })
-    }, 500)
   }, [setActiveModel, setStateModel])
 
-  const onModelInitFailed = useCallback(
-    (res: ModelInitFailed) => {
-      console.error('Failed to load model: ', res.error.message)
-      setStateModel(() => ({
-        state: 'start',
-        loading: false,
-        model: res.id,
-      }))
-      setLoadModelError(res.error.message)
-      setQueuedMessage(false)
-    },
-    [setStateModel, setQueuedMessage, setLoadModelError]
-  )
-
   const updateThreadTitle = useCallback(
     (message: ThreadMessage) => {
       // Update only when it's finished
@@ -274,7 +230,10 @@ export default function EventHandler({ children }: { children: ReactNode }) {
 
       // 2. Update the title with the result of the inference
       setTimeout(() => {
-        events.emit(MessageEvent.OnMessageSent, messageRequest)
+        const engine = EngineManager.instance()?.get(
+          messageRequest.model?.engine ?? activeModelRef.current?.engine ?? ''
+        )
+        engine?.inference(messageRequest)
       }, 1000)
     }
   }
@@ -283,23 +242,16 @@ export default function EventHandler({ children }: { children: ReactNode }) {
     if (window.core?.events) {
       events.on(MessageEvent.OnMessageResponse, onNewMessageResponse)
       events.on(MessageEvent.OnMessageUpdate, onMessageResponseUpdate)
-      events.on(ModelEvent.OnModelReady, onModelReady)
-      events.on(ModelEvent.OnModelFail, onModelInitFailed)
       events.on(ModelEvent.OnModelStopped, onModelStopped)
     }
-  }, [
+  }, [onNewMessageResponse, onMessageResponseUpdate, onModelStopped])
-    onNewMessageResponse,
-    onMessageResponseUpdate,
-    onModelReady,
-    onModelInitFailed,
-    onModelStopped,
-  ])
 
   useEffect(() => {
     return () => {
       events.off(MessageEvent.OnMessageResponse, onNewMessageResponse)
       events.off(MessageEvent.OnMessageUpdate, onMessageResponseUpdate)
+      events.off(ModelEvent.OnModelStopped, onModelStopped)
     }
-  }, [onNewMessageResponse, onMessageResponseUpdate])
+  }, [onNewMessageResponse, onMessageResponseUpdate, onModelStopped])
   return <Fragment>{children}</Fragment>
 }

web/extension/ExtensionManager.ts

@@ -1,6 +1,6 @@
 /* eslint-disable @typescript-eslint/no-explicit-any */
 
-import { BaseExtension, ExtensionTypeEnum } from '@janhq/core'
+import { AIEngine, BaseExtension, ExtensionTypeEnum } from '@janhq/core'
 
 import Extension from './Extension'
 
@@ -8,14 +8,26 @@ import Extension from './Extension'
  * Manages the registration and retrieval of extensions.
  */
 export class ExtensionManager {
+  // Registered extensions
   private extensions = new Map<string, BaseExtension>()
 
+  // Registered inference engines
+  private engines = new Map<string, AIEngine>()
+
   /**
    * Registers an extension.
    * @param extension - The extension to register.
    */
   register<T extends BaseExtension>(name: string, extension: T) {
     this.extensions.set(extension.type() ?? name, extension)
+
+    // Register AI Engines
+    if ('provider' in extension && typeof extension.provider === 'string') {
+      this.engines.set(
+        extension.provider as unknown as string,
+        extension as unknown as AIEngine
+      )
+    }
   }
 
   /**
@@ -29,6 +41,15 @@ export class ExtensionManager {
     return this.extensions.get(type) as T | undefined
   }
 
+  /**
+   * Retrieves a extension by its type.
+   * @param engine - The engine name to retrieve.
+   * @returns The extension, if found.
+   */
+  getEngine<T extends AIEngine>(engine: string): T | undefined {
+    return this.engines.get(engine) as T | undefined
+  }
+
   /**
    * Loads all registered extension.
    */

web/hooks/useActiveModel.ts

@@ -1,12 +1,13 @@
 import { useCallback, useEffect, useRef } from 'react'
 
-import { events, Model, ModelEvent } from '@janhq/core'
+import { EngineManager, Model } from '@janhq/core'
 import { atom, useAtom, useAtomValue, useSetAtom } from 'jotai'
 
 import { toaster } from '@/containers/Toast'
 
 import { LAST_USED_MODEL_ID } from './useRecommendedModel'
 
+import { extensionManager } from '@/extension'
 import { downloadedModelsAtom } from '@/helpers/atoms/Model.atom'
 import { activeThreadAtom } from '@/helpers/atoms/Thread.atom'
 
@@ -38,19 +39,13 @@ export function useActiveModel() {
       (stateModel.model === modelId && stateModel.loading)
     ) {
       console.debug(`Model ${modelId} is already initialized. Ignore..`)
-      return
+      return Promise.resolve()
     }
 
     let model = downloadedModelsRef?.current.find((e) => e.id === modelId)
 
-    // Switch between engines
+    await stopModel().catch()
-    if (model && activeModel && activeModel.engine !== model.engine) {
-      stopModel()
-      // TODO: Refactor inference provider would address this
-      await new Promise((res) => setTimeout(res, 1000))
-    }
 
-    // TODO: incase we have multiple assistants, the configuration will be from assistant
     setLoadModelError(undefined)
 
     setActiveModel(undefined)
@@ -68,7 +63,8 @@ export function useActiveModel() {
         loading: false,
         model: '',
       }))
-      return
+
+      return Promise.reject(`Model ${modelId} not found!`)
     }
 
     /// Apply thread model settings
@@ -83,15 +79,52 @@ export function useActiveModel() {
     }
 
     localStorage.setItem(LAST_USED_MODEL_ID, model.id)
-    events.emit(ModelEvent.OnModelInit, model)
+    const engine = EngineManager.instance()?.get(model.engine)
+    return engine
+      ?.loadModel(model)
+      .then(() => {
+        setActiveModel(model)
+        setStateModel(() => ({
+          state: 'stop',
+          loading: false,
+          model: model.id,
+        }))
+        toaster({
+          title: 'Success!',
+          description: `Model ${model.id} has been started.`,
+          type: 'success',
+        })
+      })
+      .catch((error) => {
+        console.error('Failed to load model: ', error)
+        setStateModel(() => ({
+          state: 'start',
+          loading: false,
+          model: model.id,
+        }))
+
+        toaster({
+          title: 'Failed!',
+          description: `Model ${model.id} failed to start.`,
+          type: 'success',
+        })
+        setLoadModelError(error)
+      })
   }
 
   const stopModel = useCallback(async () => {
     if (activeModel) {
       setStateModel({ state: 'stop', loading: true, model: activeModel.id })
-      events.emit(ModelEvent.OnModelStop, activeModel)
+      const engine = EngineManager.instance()?.get(activeModel.engine)
+      await engine
+        ?.unloadModel(activeModel)
+        .catch()
+        .then(() => {
+          setActiveModel(undefined)
+          setStateModel({ state: 'start', loading: false, model: '' })
+        })
     }
-  }, [activeModel, setStateModel])
+  }, [activeModel, setActiveModel, setStateModel])
 
   return { activeModel, startModel, stopModel, stateModel }
 }

web/hooks/useSendChatMessage.ts

@@ -11,13 +11,12 @@ import {
   ExtensionTypeEnum,
   Thread,
   ThreadMessage,
-  events,
   Model,
   ConversationalExtension,
-  MessageEvent,
   InferenceEngine,
   ChatCompletionMessageContentType,
   AssistantTool,
+  EngineManager,
 } from '@janhq/core'
 import { atom, useAtom, useAtomValue, useSetAtom } from 'jotai'
 
@@ -65,7 +64,6 @@ export default function useSendChatMessage() {
   const currentMessages = useAtomValue(getCurrentChatMessagesAtom)
   const selectedModel = useAtomValue(selectedModelAtom)
   const { activeModel, startModel } = useActiveModel()
-  const setQueuedMessage = useSetAtom(queuedMessageAtom)
   const loadModelFailed = useAtomValue(loadModelErrorAtom)
 
   const modelRef = useRef<Model | undefined>()
@@ -78,6 +76,7 @@ export default function useSendChatMessage() {
   const [fileUpload, setFileUpload] = useAtom(fileUploadAtom)
   const setIsGeneratingResponse = useSetAtom(isGeneratingResponseAtom)
   const activeThreadRef = useRef<Thread | undefined>()
+  const setQueuedMessage = useSetAtom(queuedMessageAtom)
 
   const selectedModelRef = useRef<Model | undefined>()
 
@@ -142,10 +141,7 @@ export default function useSendChatMessage() {
       activeThreadRef.current.assistants[0].model.id
 
     if (modelRef.current?.id !== modelId) {
-      setQueuedMessage(true)
+      await startModel(modelId)
-      startModel(modelId)
-      await waitForModelStarting(modelId)
-      setQueuedMessage(false)
     }
     setIsGeneratingResponse(true)
     if (currentMessage.role !== ChatCompletionRole.User) {
@@ -160,7 +156,10 @@ export default function useSendChatMessage() {
           )
       }
     }
-    events.emit(MessageEvent.OnMessageSent, messageRequest)
+    const engine = EngineManager.instance()?.get(
+      messageRequest.model?.engine ?? selectedModelRef.current?.engine ?? ''
+    )
+    engine?.inference(messageRequest)
   }
 
   const sendChatMessage = async (message: string) => {
@@ -364,30 +363,20 @@ export default function useSendChatMessage() {
 
     if (modelRef.current?.id !== modelId) {
       setQueuedMessage(true)
-      startModel(modelId)
+      await startModel(modelId)
-      await waitForModelStarting(modelId)
       setQueuedMessage(false)
     }
     setIsGeneratingResponse(true)
-    events.emit(MessageEvent.OnMessageSent, messageRequest)
+
+    const engine = EngineManager.instance()?.get(
+      messageRequest.model?.engine ?? modelRequest.engine ?? ''
+    )
+    engine?.inference(messageRequest)
 
     setReloadModel(false)
     setEngineParamsUpdate(false)
   }
 
-  const waitForModelStarting = async (modelId: string) => {
-    return new Promise<void>((resolve) => {
-      setTimeout(async () => {
-        if (modelRef.current?.id !== modelId && !loadModelFailedRef.current) {
-          await waitForModelStarting(modelId)
-          resolve()
-        } else {
-          resolve()
-        }
-      }, 200)
-    })
-  }
-
   return {
     sendChatMessage,
     resendChatMessage,

web/screens/Chat/ErrorMessage/index.tsx

@@ -74,7 +74,8 @@ const ErrorMessage = ({ message }: { message: ThreadMessage }) => {
               </p>
               <ModalTroubleShooting />
             </div>
-          ) : loadModelError?.includes('EXTENSION_IS_NOT_INSTALLED') ? (
+          ) : loadModelError &&
+            loadModelError?.includes('EXTENSION_IS_NOT_INSTALLED') ? (
             <div
               key={message.id}
               className="flex w-full flex-col items-center text-center text-sm font-medium text-gray-500"

web/services/coreService.ts

@@ -1,3 +1,5 @@
+import { EngineManager } from '@janhq/core'
+
 import { appService } from './appService'
 import { EventEmitter } from './eventsService'
 import { restAPI } from './restService'
@@ -12,6 +14,7 @@ export const setupCoreServices = () => {
   if (!window.core) {
     window.core = {
       events: new EventEmitter(),
+      engineManager: new EngineManager(),
       api: {
         ...(window.electronAPI ? window.electronAPI : restAPI),
         ...appService,