Update dependencies and add YouTube transcript support

package.json

@@ -1,7 +1,7 @@
 {
   "name": "pageassist",
   "displayName": "Page Assist - A Web UI for Local AI Models",
-  "version": "1.0.7",
+  "version": "1.0.8",
   "description": "Use your locally running AI models to assist you in your web browsing.",
   "author": "n4ze3m",
   "scripts": {
@@ -26,7 +26,7 @@
     "dayjs": "^1.11.10",
     "html-to-text": "^9.0.5",
     "langchain": "^0.1.9",
-    "lucide-react": "^0.340.0",
+    "lucide-react": "^0.350.0",
     "plasmo": "0.84.1",
     "property-information": "^6.4.1",
     "react": "18.2.0",
@@ -38,6 +38,7 @@
     "rehype-mathjax": "4.0.3",
     "remark-gfm": "3.0.1",
     "remark-math": "5.1.1",
+    "yt-transcript": "^0.0.2",
     "zustand": "^4.5.0"
   },
   "devDependencies": {

src/components/Sidepanel/Chat/empty.tsx

@@ -38,7 +38,7 @@ export const EmptySidePanel = () => {
     }
   }, [ollamaInfo])
 
-  const { setSelectedModel, selectedModel, chatMode, setChatMode } =
+  const { setSelectedModel, selectedModel, chatMode, setChatMode,  } =
     useMessage()
 
   return (

src/hooks/useMessage.tsx

@@ -20,9 +20,12 @@ import { getHtmlOfCurrentTab } from "~libs/get-html"
 import { PageAssistHtmlLoader } from "~loader/html"
 import { RecursiveCharacterTextSplitter } from "langchain/text_splitter"
 import { OllamaEmbeddings } from "@langchain/community/embeddings/ollama"
-import { createChatWithWebsiteChain, groupMessagesByConversation } from "~chain/chat-with-website"
+import {
+  createChatWithWebsiteChain,
+  groupMessagesByConversation
+} from "~chain/chat-with-website"
 import { MemoryVectorStore } from "langchain/vectorstores/memory"
-
+import { chromeRunTime } from "~libs/runtime"
 export type BotResponse = {
   bot: {
     text: string
@@ -134,11 +137,11 @@ export const useMessage = () => {
       url
     })
     const docs = await loader.load()
-    const chunkSize = await defaultEmbeddingChunkSize();
+    const chunkSize = await defaultEmbeddingChunkSize()
-    const chunkOverlap = await defaultEmbeddingChunkOverlap();
+    const chunkOverlap = await defaultEmbeddingChunkOverlap()
     const textSplitter = new RecursiveCharacterTextSplitter({
       chunkSize,
-      chunkOverlap,
+      chunkOverlap
     })
 
     const chunks = await textSplitter.splitDocuments(docs)
@@ -158,9 +161,10 @@ export const useMessage = () => {
   }
 
   const chatWithWebsiteMode = async (message: string) => {
-    const ollamaUrl = await getOllamaURL()
+    try {
+      let isAlreadyExistEmbedding: MemoryVectorStore
       const { html, url } = await getHtmlOfCurrentTab()
-    const isAlreadyExistEmbedding = keepTrackOfEmbedding[url]
+      isAlreadyExistEmbedding = keepTrackOfEmbedding[url]
       let newMessage: Message[] = [
         ...messages,
         {
@@ -179,13 +183,14 @@ export const useMessage = () => {
 
       const appendingIndex = newMessage.length - 1
       setMessages(newMessage)
+      const ollamaUrl = await getOllamaURL()
       const embeddingModle = await defaultEmbeddingModelForRag()
+
       const ollamaEmbedding = new OllamaEmbeddings({
         model: embeddingModle || selectedModel,
         baseUrl: cleanUrl(ollamaUrl)
       })
 
-
       const ollamaChat = new ChatOllama({
         model: selectedModel,
         baseUrl: cleanUrl(ollamaUrl)
@@ -212,10 +217,9 @@ export const useMessage = () => {
         retriever: vectorstore.asRetriever()
       })
 
-    try {
       const chunks = await chain.stream({
         question: sanitizedQuestion,
-        chat_history: groupMessagesByConversation(history),
+        chat_history: groupMessagesByConversation(history)
       })
       let count = 0
       for await (const chunk of chunks) {
@@ -258,7 +262,8 @@ export const useMessage = () => {
         {
           isBot: true,
           name: selectedModel,
-          message: `Something went wrong. Check out the following logs:
+          message: `Error in chat with website mode. Check out the following logs:
+          
 ~~~
 ${e?.message}
  ~~~

src/libs/get-html.ts

@@ -1,3 +1,4 @@
+
 const _getHtml = () => {
   const url = window.location.href
   const html = Array.from(document.querySelectorAll("script")).reduce(
@@ -29,3 +30,4 @@ export const getHtmlOfCurrentTab = async () => {
 
   return result
 }
+

src/loader/html.ts

@@ -2,6 +2,20 @@ import { BaseDocumentLoader } from "langchain/document_loaders/base"
 import { Document } from "@langchain/core/documents"
 import { compile } from "html-to-text"
 import { chromeRunTime } from "~libs/runtime"
+import { YtTranscript } from "yt-transcript"
+
+const YT_REGEX =
+  /(?:https?:\/\/)?(?:www\.)?(?:youtube\.com|youtu\.be)\/(?:watch\?v=)?([a-zA-Z0-9_-]+)/
+
+const isYoutubeLink = (url: string) => {
+  return YT_REGEX.test(url)
+}
+
+const getTranscript = async (url: string) => {
+  const ytTranscript = new YtTranscript({ url })
+  return await ytTranscript.getTranscript()
+}
+
 
 export interface WebLoaderParams {
   html: string
@@ -21,6 +35,29 @@ export class PageAssistHtmlLoader
   }
 
   async load(): Promise<Document<Record<string, any>>[]> {
+    if (isYoutubeLink(this.url)) {
+      const transcript = await getTranscript(this.url)
+      if (!transcript) {
+        throw new Error("Transcript not found for this video.")
+      }
+
+      let text = ""
+
+      transcript.forEach((item) => {
+        text += item.text + " "
+      })
+
+
+      return [
+        {
+          metadata: {
+            source: this.url,
+            audio: { chunks: transcript }
+          },
+          pageContent: text
+        }
+      ]
+    }
     const htmlCompiler = compile({
       wordwrap: false
     })
@@ -30,6 +67,29 @@ export class PageAssistHtmlLoader
   }
 
   async loadByURL(): Promise<Document<Record<string, any>>[]> {
+    if (isYoutubeLink(this.url)) {
+      const transcript = await getTranscript(this.url)
+      if (!transcript) {
+        throw new Error("Transcript not found for this video.")
+      }
+
+      let text = ""
+
+      transcript.forEach((item) => {
+        text += item.text + " "
+      })
+
+
+      return [
+        {
+          metadata: {
+            source: this.url,
+            audio: { chunks: transcript }
+          },
+          pageContent: text
+        }
+      ]
+    }
     await chromeRunTime(this.url)
     const fetchHTML = await fetch(this.url)
     const html = await fetchHTML.text()

yarn.lock

@@ -4967,10 +4967,10 @@ lru-cache@^6.0.0:
   resolved "https://registry.yarnpkg.com/lru-cache/-/lru-cache-10.2.0.tgz#0bd445ca57363465900f4d1f9bd8db343a4d95c3"
   integrity sha512-2bIM8x+VAf6JT4bKAljS1qUWgMsqZRPGJS6FSahIMPVvctcNhyVp7AJu7quxOW9jwkryBReKZY5tY5JYv2n/7Q==
 
-lucide-react@^0.340.0:
+lucide-react@^0.350.0:
-  version "0.340.0"
+  version "0.350.0"
-  resolved "https://registry.yarnpkg.com/lucide-react/-/lucide-react-0.340.0.tgz#67a6fac6a5e257f2036dffae0dd94d6ccb28ce8e"
+  resolved "https://registry.yarnpkg.com/lucide-react/-/lucide-react-0.350.0.tgz#78b45342f4daff4535290e37b1ea7eb0961a3dab"
-  integrity sha512-mWzYhbyy2d+qKuKHh+GWElPwa+kIquTnKbmSLGWOuZy+bjfZCkYD8DQWVFlqI4mQwc4HNxcqcOvtQ7ZS2PwURg==
+  integrity sha512-5IZVKsxxG8Nn81gpsz4XLNgCAXkppCh0Y0P0GLO39h5iVD2WEaB9of6cPkLtzys1GuSfxJxmwsDh487y7LAf/g==
 
 magic-string@^0.30.0:
   version "0.30.6"
@@ -7772,6 +7772,13 @@ ws@^8.11.0:
   resolved "https://registry.yarnpkg.com/ws/-/ws-8.16.0.tgz#d1cd774f36fbc07165066a60e40323eab6446fd4"
   integrity sha512-HS0c//TP7Ina87TfiPUz1rQzMhHrl/SG2guqRcTOIUYD2q8uhUdNHZYJUaQ8aTGPzCh+c6oawMKW35nFl1dxyQ==
 
+xml-js@^1.6.11:
+  version "1.6.11"
+  resolved "https://registry.yarnpkg.com/xml-js/-/xml-js-1.6.11.tgz#927d2f6947f7f1c19a316dd8eea3614e8b18f8e9"
+  integrity sha512-7rVi2KMfwfWFl+GpPg6m80IVMWXLRjO+PxTq7V2CDhoGak0wzYzFgUY2m4XJ47OGdXd8eLE8EmwfAmdjw7lC1g==
+  dependencies:
+    sax "^1.2.4"
+
 xml-name-validator@^4.0.0:
   version "4.0.0"
   resolved "https://registry.yarnpkg.com/xml-name-validator/-/xml-name-validator-4.0.0.tgz#79a006e2e63149a8600f15430f0a4725d1524835"
@@ -7817,6 +7824,14 @@ yaml@^2.2.1, yaml@^2.3.4:
   resolved "https://registry.yarnpkg.com/yaml/-/yaml-2.3.4.tgz#53fc1d514be80aabf386dc6001eb29bf3b7523b2"
   integrity sha512-8aAvwVUSHpfEqTQ4w/KMlf3HcRdt50E5ODIQJBw1fQ5RL34xabzxtUlzTXVqc4rkZsPbvrXKWnABCD7kWSmocA==
 
+yt-transcript@^0.0.2:
+  version "0.0.2"
+  resolved "https://registry.yarnpkg.com/yt-transcript/-/yt-transcript-0.0.2.tgz#1c54aede89bb8a03bbca3ba58520dbbd9c828571"
+  integrity sha512-+cNRqW6tSQNDkQDVrWNT6hc6X2TnaQLvUJIepzn9r7XdEvPtUDkfsyhptW5+j0EPIEpnlsKyA/epCUrE4QKn2g==
+  dependencies:
+    axios "^1.6.7"
+    xml-js "^1.6.11"
+
 zod-to-json-schema@^3.22.3:
   version "3.22.4"
   resolved "https://registry.yarnpkg.com/zod-to-json-schema/-/zod-to-json-schema-3.22.4.tgz#f8cc691f6043e9084375e85fb1f76ebafe253d70"