Create main.py

main.py

@@ -0,0 +1,238 @@
+
+from fastapi import FastAPI, UploadFile, File
+from fastapi.middleware.cors import CORSMiddleware
+from fastapi.responses import JSONResponse
+
+# Import both services
+try:
+    from split import (
+        app as ingestion_app,
+        ingest_pdf,
+        get_stats as get_ingestion_stats
+    )
+    print("✅ Loaded ingestion service (split.py)")
+except Exception as e:
+    print(f"⚠️ Warning: Could not load ingestion service: {e}")
+    ingestion_app = None
+
+try:
+    from query_service import (
+        app as query_app,
+        query_rag,
+        query_with_details,
+        get_stats as get_query_stats
+    )
+    print("✅ Loaded query service (query_service.py)")
+except Exception as e:
+    print(f"⚠️ Warning: Could not load query service: {e}")
+    query_app = None
+
+
+# Create main application
+app = FastAPI(
+    title="🚀 Multimodal RAG - Combined Service",
+    description="Unified service for PDF ingestion and multimodal querying",
+    version="1.0.0",
+    docs_url="/docs",
+    redoc_url="/redoc"
+)
+
+# Add CORS middleware
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+
+# ==================== HOME & HEALTH ENDPOINTS ====================
+
+@app.get("/")
+async def home():
+    """Main endpoint with service information"""
+    return {
+        "message": "✅ Multimodal RAG Service is running",
+        "status": "healthy",
+        "services": {
+            "ingestion": "available" if ingestion_app else "unavailable",
+            "query": "available" if query_app else "unavailable"
+        },
+        "endpoints": {
+            "ingestion": {
+                "ingest_pdf": "POST /ingest",
+                "ingestion_stats": "GET /ingest/stats"
+            },
+            "query": {
+                "query": "POST /query?question=YOUR_QUESTION&k=5",
+                "query_detailed": "POST /query/details?question=YOUR_QUESTION&k=5",
+                "query_stats": "GET /query/stats"
+            },
+            "documentation": {
+                "swagger": "/docs",
+                "redoc": "/redoc"
+            }
+        },
+        "features": [
+            "PDF text extraction",
+            "Table extraction", 
+            "Image extraction",
+            "Multimodal summarization",
+            "Vector similarity search",
+            "Context-aware answering"
+        ]
+    }
+
+
+@app.get("/health")
+async def health_check():
+    """Health check endpoint"""
+    return {
+        "status": "healthy",
+        "services": {
+            "ingestion": ingestion_app is not None,
+            "query": query_app is not None
+        }
+    }
+
+
+# ==================== INGESTION ENDPOINTS ====================
+
+@app.post("/ingest")
+async def ingest_document(file: UploadFile = File(...)):
+    """
+    Upload and ingest a PDF document
+    
+    This endpoint processes PDFs and extracts:
+    - Text content
+    - Tables
+    - Images
+    
+    All content is summarized and stored in the vectorstore.
+    """
+    if ingestion_app is None:
+        return JSONResponse(
+            status_code=503,
+            content={"error": "Ingestion service not available"}
+        )
+    
+    return await ingest_pdf(file)
+
+
+@app.get("/ingest/stats")
+async def ingestion_stats():
+    """Get ingestion service statistics"""
+    if ingestion_app is None:
+        return JSONResponse(
+            status_code=503,
+            content={"error": "Ingestion service not available"}
+        )
+    
+    return get_ingestion_stats()
+
+
+# ==================== QUERY ENDPOINTS ====================
+
+@app.post("/query")
+async def query_documents(question: str, k: int = 5):
+    """
+    Query the RAG system
+    
+    Args:
+        question: The question to ask
+        k: Number of documents to retrieve (default: 5)
+    
+    Returns:
+        Answer based on retrieved documents
+    """
+    if query_app is None:
+        return JSONResponse(
+            status_code=503,
+            content={"error": "Query service not available"}
+        )
+    
+    return await query_rag(question, k)
+
+
+@app.post("/query/details")
+async def query_documents_detailed(question: str, k: int = 5):
+    """
+    Query with detailed document information
+    
+    Args:
+        question: The question to ask
+        k: Number of documents to retrieve (default: 5)
+    
+    Returns:
+        Answer with detailed information about retrieved documents
+    """
+    if query_app is None:
+        return JSONResponse(
+            status_code=503,
+            content={"error": "Query service not available"}
+        )
+    
+    return await query_with_details(question, k)
+
+
+@app.get("/query/stats")
+async def query_stats():
+    """Get query service statistics"""
+    if query_app is None:
+        return JSONResponse(
+            status_code=503,
+            content={"error": "Query service not available"}
+        )
+    
+    return get_query_stats()
+
+
+# ==================== COMBINED STATS ENDPOINT ====================
+
+@app.get("/stats")
+async def combined_stats():
+    """Get combined statistics from both services"""
+    stats = {
+        "service": "combined",
+        "status": "healthy"
+    }
+    
+    try:
+        if query_app:
+            query_stats_data = get_query_stats()
+            stats["vectorstore"] = query_stats_data
+    except Exception as e:
+        stats["vectorstore_error"] = str(e)
+    
+    return stats
+
+
+
+
+@app.on_event("startup")
+async def startup_event():
+    """Run on application startup"""
+    print("\n" + "="*50)
+    print("🚀 Multimodal RAG Service Starting...")
+    print("="*50)
+    
+    if ingestion_app:
+        print("✅ Ingestion service: READY")
+    else:
+        print("⚠️  Ingestion service: NOT LOADED")
+    
+    if query_app:
+        print("✅ Query service: READY")
+    else:
+        print("⚠️  Query service: NOT LOADED")
+    
+    print("="*50)
+    print("📡 Service available at: http://0.0.0.0:7860")
+    print("📚 API Documentation: http://0.0.0.0:7860/docs")
+    print("="*50 + "\n")
+
+
+if __name__ == "__main__":
+    import uvicorn
+    uvicorn.run(app, host="0.0.0.0", port=7860)