Spaces:
Sleeping
Sleeping
Download server.py from versr-ai-browser/Verse: direct link, hf CLI and curl.
- Browser
- Download file 8.67 kB
-
https://huggingface.co/spaces/versr-ai-browser/Verse/resolve/main/server.py
- Command line
-
hf download hf://spaces/versr-ai-browser/Verse/server.py
-
curl -L -o server.py https://huggingface.co/spaces/versr-ai-browser/Verse/resolve/main/server.py
8.67 kB
| # pyrefly: ignore [missing-import] | |
| from dotenv import load_dotenv | |
| load_dotenv() | |
| import os | |
| # pyrefly: ignore [missing-import] | |
| from fastapi import FastAPI, HTTPException | |
| # pyrefly: ignore [missing-import] | |
| from fastapi.middleware.cors import CORSMiddleware | |
| # pyrefly: ignore [missing-import] | |
| from pydantic import BaseModel, Field | |
| # Import core RAG logic functions | |
| import core_logic | |
| app = FastAPI( | |
| title="AI Browser RAG Backend", | |
| description="Backend API for text parsing, embedding, retrieval, LLM querying, and AI actions", | |
| version="2.0.0" | |
| ) | |
| # Allow CORS for all origins since it will be deployed | |
| app.add_middleware( | |
| CORSMiddleware, | |
| allow_origins=["*"], | |
| allow_credentials=True, | |
| allow_methods=["*"], | |
| allow_headers=["*"], | |
| ) | |
| # --- Pydantic Schemas --- | |
| class ProcessHtmlRequest(BaseModel): | |
| html: str | |
| chunk_size: int = Field(default=50, ge=1) | |
| overlap: int = Field(default=10, ge=0) | |
| class ProcessHtmlResponse(BaseModel): | |
| chunks: list[str] | |
| class EmbedRequest(BaseModel): | |
| texts: list[str] | |
| class EmbedResponse(BaseModel): | |
| embeddings: list[list[float]] | |
| class RetrieveRequest(BaseModel): | |
| query: str | |
| chunks: list[str] | |
| embeddings: list[list[float]] | |
| top_k: int = Field(default=2, ge=1) | |
| class RetrievalResultItem(BaseModel): | |
| score: float | |
| index: int | |
| text: str | |
| class RetrieveResponse(BaseModel): | |
| results: list[RetrievalResultItem] | |
| class AnswerRequest(BaseModel): | |
| query: str | |
| context: str | |
| api_key: str | None = None | |
| class AnswerResponse(BaseModel): | |
| answer: str | |
| class UnifiedRagRequest(BaseModel): | |
| html: str | |
| query: str | |
| chunk_size: int = Field(default=50, ge=1) | |
| overlap: int = Field(default=10, ge=0) | |
| top_k: int = Field(default=2, ge=1) | |
| title: str = "" | |
| url: str = "" | |
| selected_text: str = "" | |
| api_key: str | None = None | |
| class UnifiedRagResponse(BaseModel): | |
| answer: str | |
| retrieved_chunks: list[RetrievalResultItem] | |
| # --- AI Action Schemas --- | |
| class TextRequest(BaseModel): | |
| text: str | |
| api_key: str | None = None | |
| class TranslateRequest(BaseModel): | |
| text: str | |
| target_language: str | |
| api_key: str | None = None | |
| class ExtractRequest(BaseModel): | |
| text: str | |
| data_type: str | |
| api_key: str | None = None | |
| class GeneratePostRequest(BaseModel): | |
| text: str | |
| platform: str | |
| api_key: str | None = None | |
| class RewriteRequest(BaseModel): | |
| text: str | |
| style: str | |
| api_key: str | None = None | |
| class ActionResponse(BaseModel): | |
| result: str | |
| # --- Route Handlers --- | |
| def read_root(): | |
| return {"status": "ok", "message": "AI Browser RAG Backend API is running.", "version": "2.0.0"} | |
| def process_html(payload: ProcessHtmlRequest): | |
| try: | |
| clean_text = core_logic.extract_clean_prose(payload.html) | |
| chunks = core_logic.chunk_text(clean_text, payload.chunk_size, payload.overlap) | |
| return ProcessHtmlResponse(chunks=chunks) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| def embed_chunks(payload: EmbedRequest): | |
| try: | |
| embeddings = core_logic.embed_texts(payload.texts) | |
| return EmbedResponse(embeddings=embeddings) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| def retrieve(payload: RetrieveRequest): | |
| try: | |
| if len(payload.chunks) != len(payload.embeddings): | |
| raise HTTPException(status_code=400, detail="The number of chunks must match the number of embeddings.") | |
| results = core_logic.retrieve_top_chunks(payload.query, payload.chunks, payload.embeddings, payload.top_k) | |
| return RetrieveResponse(results=results) | |
| except HTTPException as he: | |
| raise he | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| def get_answer(payload: AnswerRequest): | |
| try: | |
| answer = core_logic.query_llm(payload.query, payload.context, payload.api_key) | |
| return AnswerResponse(answer=answer) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| def run_unified_rag(payload: UnifiedRagRequest): | |
| try: | |
| # Step 1: Clean and chunk HTML | |
| clean_text = core_logic.extract_clean_prose(payload.html) | |
| chunks = core_logic.chunk_text(clean_text, payload.chunk_size, payload.overlap) | |
| if not chunks: | |
| return UnifiedRagResponse(answer="No text chunks could be extracted from the HTML content.", retrieved_chunks=[]) | |
| # Step 2: Embed chunks | |
| embeddings = core_logic.embed_texts(chunks) | |
| # Step 3: Retrieve top chunks | |
| top_chunks = core_logic.retrieve_top_chunks(payload.query, chunks, embeddings, payload.top_k) | |
| # Step 4: Format context and query LLM | |
| context = "\n\n".join([f"[Source Chunk {item['index'] + 1}]: {item['text']}" for item in top_chunks]) | |
| answer = core_logic.query_llm( | |
| query=payload.query, | |
| context=context, | |
| title=payload.title, | |
| url=payload.url, | |
| selected_text=payload.selected_text, | |
| api_key=payload.api_key | |
| ) | |
| return UnifiedRagResponse(answer=answer, retrieved_chunks=top_chunks) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| # --- AI Action Endpoints --- | |
| def summarize(payload: TextRequest): | |
| try: | |
| result = core_logic.summarize_text(payload.text, payload.api_key) | |
| return ActionResponse(result=result) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| def explain(payload: TextRequest): | |
| try: | |
| result = core_logic.explain_text(payload.text, payload.api_key) | |
| return ActionResponse(result=result) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| def translate(payload: TranslateRequest): | |
| try: | |
| result = core_logic.translate_text(payload.text, payload.target_language, payload.api_key) | |
| return ActionResponse(result=result) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| def extract(payload: ExtractRequest): | |
| try: | |
| result = core_logic.extract_data(payload.text, payload.data_type, payload.api_key) | |
| return ActionResponse(result=result) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| def generate_post(payload: GeneratePostRequest): | |
| try: | |
| result = core_logic.generate_post(payload.text, payload.platform, payload.api_key) | |
| return ActionResponse(result=result) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| def rewrite(payload: RewriteRequest): | |
| try: | |
| result = core_logic.rewrite_text(payload.text, payload.style, payload.api_key) | |
| return ActionResponse(result=result) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| def key_points(payload: TextRequest): | |
| try: | |
| result = core_logic.list_key_points(payload.text, payload.api_key) | |
| return ActionResponse(result=result) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| def faqs(payload: TextRequest): | |
| try: | |
| result = core_logic.generate_faqs(payload.text, payload.api_key) | |
| return ActionResponse(result=result) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| def study_notes(payload: TextRequest): | |
| try: | |
| result = core_logic.create_study_notes(payload.text, payload.api_key) | |
| return ActionResponse(result=result) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| if __name__ == "__main__": | |
| # pyrefly: if __name__ == "__main__": | |
| import uvicorn | |
| import os | |
| port = int(os.environ.get("PORT", 8000)) | |
| uvicorn.run(app, host="0.0.0.0", port=port) | |