feat: 增加url阅读能力,通过 jina.ai 实现

.gitignore

@@ -4,3 +4,4 @@ node_modules/
 /playwright-report/
 /blob-report/
 /playwright/.cache/
+.env

backend/api/routes/chat.py

@@ -4,20 +4,27 @@
 import json
 from typing import Dict, Optional
 from constants import SPEEKERS
-from utils import combine_audio, generate_dialogue, generate_podcast_info, generate_podcast_summary, get_pdf_text
+from utils import combine_audio, generate_dialogue, generate_podcast_info, generate_podcast_summary, get_link_text, get_pdf_text
 
 router = APIRouter()
 
 @router.post("/generate_transcript")
 async def generate_transcript(
     pdfFile: Optional[UploadFile] = File(None),
     textInput: str = Form(...),
+    mode: str = Form(...),
+    url: Optional[str] = Form(None),
     tone: str = Form(...),
     duration: str = Form(...),
     language: str = Form(...),
 
 ): 
-    pdfContent = await get_pdf_text(pdfFile)
+    pdfContent ="" 
+    if mode=='pdf':
+        pdfContent = await get_pdf_text(pdfFile)
+    else:
+        linkData = get_link_text(url)
+        pdfContent = linkData['text']
     new_text = pdfContent
     return StreamingResponse(generate_dialogue(new_text,textInput, tone, duration, language), media_type="application/json")
 
@@ -31,16 +38,28 @@ def test():
 def speeker():
     return JSONResponse(content=SPEEKERS)
 
+@router.get("/jina")
+def jina():
+    result = get_link_text("https://ui.shadcn.com/docs/components/select")
+    return JSONResponse(content=result)
+
 
 @router.post("/summarize")
 async def get_summary(
     textInput: str = Form(...),
     tone: str = Form(...),
     duration: str = Form(...),
     language: str = Form(...),
+    mode: str = Form(...),
+    url: Optional[str] = Form(None),
     pdfFile: Optional[UploadFile] = File(None)
 ):
-    pdfContent = await get_pdf_text(pdfFile)
+    pdfContent ="" 
+    if mode=='pdf':
+        pdfContent = await get_pdf_text(pdfFile)
+    else:
+        linkData = get_link_text(url)
+        pdfContent = linkData['text']
     new_text = pdfContent
     return StreamingResponse(
         generate_podcast_summary(
@@ -59,9 +78,17 @@ async def get_pod_info(
     tone: str = Form(...),
     duration: str = Form(...),
     language: str = Form(...),
+    mode: str = Form(...),
+    url: Optional[str] = Form(None),
     pdfFile: Optional[UploadFile] = File(None)
 ):
-    pdfContent = await get_pdf_text(pdfFile)
+    pdfContent ="" 
+    if mode=='pdf':
+        pdfContent = await get_pdf_text(pdfFile)
+    else:
+        linkData = get_link_text(url)
+        pdfContent = linkData['text']
+
     new_text = pdfContent[:100]
 
     return StreamingResponse(generate_podcast_info(new_text, textInput, tone, duration, language), media_type="application/json")

backend/main.py

@@ -16,7 +16,7 @@
 # 添加CORS中间件
 app.add_middleware(
     CORSMiddleware,
-    allow_origins=["https://ai.podcastlm.fun/"],
+    allow_origins=["*"],
     allow_credentials=True,
     allow_methods=["*"],
     allow_headers=["*"],

backend/utils.py

@@ -5,10 +5,10 @@
 import re
 import time
 import hashlib
-
 from typing import Any, Dict, Generator
 import uuid
 from openai import OpenAI
+import requests
 from fishaudio import fishaudio_tts
 from prompts import LANGUAGE_MODIFIER, LENGTH_MODIFIERS, PODCAST_INFO_PROMPT, QUESTION_MODIFIER, SUMMARY_INFO_PROMPT, SYSTEM_PROMPT, TONE_MODIFIER
 import json
@@ -254,6 +254,16 @@ def clear_pdf_cache():
     global pdf_cache
     pdf_cache.clear()
 
+def get_link_text(url: str):
+    """ 通过jina.ai 抓取url内容 """
+    url  = f"https://r.jina.ai/{url}"
+    headers = {}
+    headers['Authorization'] = 'Bearer jina_c1759c7f49e14ced990ac7776800dc44ShJNTXBCizzwjE7IMFYJ6LD960cG'
+    headers['Accept'] = 'application/json'
+    headers['X-Return-Format'] = 'text'
+    response = requests.get(url, headers=headers)
+    return response.json()['data']
+
 async def get_pdf_text(pdf_file: UploadFile):
     text = ""
     print(pdf_file)

frontend/.env.template

@@ -0,0 +1,2 @@
+BASE_URL=
+HOST_URL=

frontend/package.json

@@ -17,6 +17,7 @@
     "@radix-ui/react-select": "^2.1.2",
     "@radix-ui/react-slot": "^1.1.0",
     "@radix-ui/react-tabs": "^1.1.1",
+    "@radix-ui/react-toast": "^1.2.2",
     "@radix-ui/react-toggle": "^1.1.0",
     "@radix-ui/react-toggle-group": "^1.1.0",
     "axios": "^1.7.7",

frontend/src/App.tsx

@@ -5,6 +5,7 @@ import { useJsonData } from "./hooks/useJsonData";
 import { useStreamText } from './hooks/useStreamText';
 import { BASE_URL } from "./lib/constant";
 import MobileMenu from "./components/mobile-menu";
+import { Toaster } from "@/components/ui/toaster"
 
 function App() {
   const [isGenerating, setIsGenerating] = useState(false);
@@ -37,6 +38,7 @@ function App() {
   }
   return (
     <div className="h-screen flex flex-col overflow-hidden">
+       <Toaster />
       <main className="flex-grow flex bg-[rgb(245,245,245)] h-full">
         <Menu className="hidden md:flex "
           handleGenerate={handleGenerate}