from fastapi import FastAPI, HTTPException from fastapi.middleware.cors import CORSMiddleware from fastapi.responses import HTMLResponse from pydantic import BaseModel from typing import List, Optional import logging from dataclasses import asdict import time from api_clients import SemanticScholarAPI from llm_providers import GroqExtractor logging.basicConfig(level=logging.INFO) logger = logging.getLogger(__name__) app = FastAPI(title="Aging Theory Analyzer API", version="1.0.0") app.add_middleware( CORSMiddleware, allow_origins=["*"], allow_credentials=True, allow_methods=["*"], allow_headers=["*"], ) class CollectRequest(BaseModel): queries: List[str] year_from: int = 2000 year_to: int = 2025 max_papers: int = 100 class ExtractRequest(BaseModel): papers: List[dict] llm_provider: str llm_api_key: Optional[str] = None class BatchExtractRequest(BaseModel): """Обработка одного батча статей""" papers: List[dict] llm_provider: str llm_api_key: str batch_number: int total_batches: int @app.get("/", response_class=HTMLResponse) async def root(): return """ Aging Theory Analyzer API

🧬 Aging Theory Analyzer API

✅ Backend is running on Hugging Face Spaces!

✨ Features:

📚 Available Endpoints:

GET /docs - Interactive API documentation (Swagger UI)
GET /api/health - Health check endpoint
POST /api/collect - Collect papers from Semantic Scholar
POST /api/extract/batch - Extract Q1-Q9 from one batch (10 papers)
POST /api/extract - Legacy: Extract all papers at once (not recommended for >50 papers)

🚀 Quick Start:

  1. Copy this URL:
  2. Paste in frontend "Backend URL" field
  3. Get free Groq API key: console.groq.com/keys
  4. Add queries and start analyzing!

Version 1.0.0 | Powered by FastAPI + Groq LLM (Llama 3.1 70B) + Semantic Scholar API

""" @app.get("/api/health") async def health(): """Health check endpoint""" return { "status": "ok", "version": "1.0.0", "timestamp": time.time(), "message": "Aging Theory Analyzer Backend is running", "features": { "paper_collection": "Semantic Scholar API", "llm_extraction": "Groq (Llama 3.1 70B)", "batch_processing": "10 papers per batch" } } @app.post("/api/collect") async def collect_papers(request: CollectRequest): """Collect papers from Semantic Scholar""" logger.info(f"Collection: {len(request.queries)} queries, {request.year_from}-{request.year_to}") all_papers = [] papers_per_query = request.max_papers // len(request.queries) if request.queries else request.max_papers api = SemanticScholarAPI() for query in request.queries: try: papers = await api.search_papers(query, request.year_from, request.year_to, papers_per_query) all_papers.extend(papers) logger.info(f"Collected {len(papers)} for '{query}'") except Exception as e: logger.error(f"Error: {e}") # Deduplicate seen = set() unique_papers = [] for paper in all_papers: key = paper.id or paper.title if key not in seen: seen.add(key) unique_papers.append(paper) logger.info(f"Collection complete: {len(unique_papers)} unique papers") return { "papers": [asdict(p) for p in unique_papers], "total": len(unique_papers), "api_calls": len(unique_papers) // 100 + 1, "recommended_batch_size": 10 } @app.post("/api/extract/batch") async def extract_batch(request: BatchExtractRequest): """ Extract Q1-Q9 from ONE batch of papers (recommended: 10 papers per batch) Frontend should call this endpoint multiple times for large datasets: - Split papers into batches of 10 - Call this endpoint for each batch - Update progress bar after each batch """ logger.info(f"Batch {request.batch_number}/{request.total_batches}: {len(request.papers)} papers with {request.llm_provider}") if not request.llm_api_key: raise HTTPException(status_code=400, detail="LLM API key required") extractor = GroqExtractor(request.llm_api_key) results = [] failed = [] for idx, paper in enumerate(request.papers, 1): try: result = await extractor.extract(paper) results.append(result) logger.info(f"Batch {request.batch_number}: [{idx}/{len(request.papers)}] ✓ {paper['title'][:50]}") except Exception as e: failed.append({ "paper_id": paper.get('id', 'unknown'), "paper_title": paper.get('title', 'unknown'), "error": str(e) }) logger.error(f"Batch {request.batch_number}: [{idx}/{len(request.papers)}] ✗ {str(e)[:100]}") return { "batch_number": request.batch_number, "total_batches": request.total_batches, "results": results, "processed": len(results), "failed": len(failed), "failed_papers": failed, "theories_found": sum(1 for r in results if r.get('q2') == 'Yes') } @app.post("/api/extract") async def extract_data(request: ExtractRequest): """ Legacy endpoint: Extract all papers at once ⚠️ NOT RECOMMENDED for large datasets (>50 papers) Use /api/extract/batch instead for better reliability """ logger.info(f"Extraction: {len(request.papers)} papers with {request.llm_provider}") if not request.llm_api_key: raise HTTPException(status_code=400, detail="API key required") # Warn if too many papers if len(request.papers) > 50: logger.warning(f"Large dataset detected ({len(request.papers)} papers). Consider using /api/extract/batch") extractor = GroqExtractor(request.llm_api_key) results = [] failed_count = 0 for idx, paper in enumerate(request.papers, 1): try: result = await extractor.extract(paper) results.append(result) logger.info(f"[{idx}/{len(request.papers)}] ✓") except Exception as e: failed_count += 1 logger.error(f"[{idx}] Failed: {e}") return { "results": results, "total": len(results), "theories_found": sum(1 for r in results if r.get('q2') == 'Yes'), "failed": failed_count, "warning": "Consider using /api/extract/batch for large datasets" if len(request.papers) > 50 else None }