Spaces:
Running
Running
File size: 10,412 Bytes
5c095ca 9b7c6ff 5c095ca 9b7c6ff 5951bbe 5c095ca 2798de4 5c095ca 2798de4 5c095ca 76e224e 5c095ca 9b7c6ff 99e045a 9b7c6ff 99e045a 5951bbe 99e045a 2798de4 99e045a 9b7c6ff 99e045a 9b7c6ff 5c095ca 5951bbe 5c095ca | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 | """
ResearchPilot FastAPI application.
STARTUP BEHAVIOR:
When the server starts, it loads ALL models into memory:
- BGE embedding model (~110MB)
- Cross-encoder re-ranker (~80MB)
- BM25 index (~40MB)
- Qdrant connection
This takes ~15 seconds once, then every request is fast.
This is called "warm start" - the model is always ready.
Without this, the first request after server restart
would take 20+ seconds. Unacceptable for production.
LIFESPAN PATTERN:
FastAPI's lifespan context manager runs code at startup
and shutdown. We use it to initialize the RAG pipeline
once and store it in app.state for all requests to share.
"""
import asyncio
import time
from contextlib import asynccontextmanager
from fastapi import FastAPI, HTTPException, Request
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import JSONResponse, StreamingResponse
from pydantic import BaseModel
import json
import os
from src.api.schemas import (
QueryRequest,
QueryResponse,
CitationSchema,
HealthResponse,
ErrorResponse,
)
class FeedbackRequest(BaseModel):
query: str
rating: int
thumbs: str | None = None
comment: str
model_used: str
citations_count: int
total_time_ms: float
from src.rag.pipeline import RAGPipeline, ConversationTurn
from src.utils.logger import setup_logger, get_logger
from config.settings import HF_API_KEY
setup_logger()
logger = get_logger(__name__)
# ---------------------------------------------------------
# LIFESPAN - runs at startup and shutdown
# ---------------------------------------------------------
@asynccontextmanager
async def lifespan(app: FastAPI):
"""
Initialize resources at startup, clean up at shutdown.
The 'yield' separates startup (before) from shutdown (after).
Everything before yield runs when server starts.
Everything after yield runs when server shuts down.
"""
# --------------- STARTUP ---------------
logger.info("ResearchPilot API starting up...")
start = time.time()
# Initialize RAG pipeline - loads all models into memory
# We store it on app.state so all request handlers can access it
app.state.rag_pipeline = RAGPipeline()
# Log the active model chain for deployment verification
from src.rag.llm_client import MultiModelClient
logger.info(f"Model chain: {MultiModelClient.MODEL_CHAIN}")
logger.info(f"HF_API_KEY configured: {bool(HF_API_KEY)}")
elapsed = time.time() - start
logger.info(f"API ready in {elapsed:.1f}s")
yield # Server is now running and handling requests
# --------------- SHUTDOWN ---------------
logger.info("ResearchPilot API shutting down...")
# ---------------------------------------------------------
# APP INITIALIZATION
# ---------------------------------------------------------
app = FastAPI(
title = "ResearchPilot API",
description = "Production RAG system for ML research paper Q&A",
version = "1.0.0",
lifespan = lifespan,
docs_url = "/docs", # Swagger UI at http://localhost:8000/docs
redoc_url = "/redoc", # ReDoc at http://localhost:8000/redoc
)
# CORS middleware — allows browser-based frontends to call this API
# Without this, a browser on localhost:3000 cannot call localhost:8000
app.add_middleware(
CORSMiddleware,
allow_origins = ["*"], # In production, restrict to your domain
allow_methods = ["*"],
allow_headers = ["*"],
)
# ---------------------------------------------------------
# EXCEPTION HANDLER
# ---------------------------------------------------------
@app.exception_handler(Exception)
async def global_exception_handler(request: Request, exc: Exception):
"""
Catch any unhandled exception and return a clean JSON error.
Without this, FastAPI returns a raw 500 error with no detail.
"""
logger.error(f"Unhandled exception on {request.url}: {exc}")
return JSONResponse(
status_code = 500,
content = {
"error": "Internal server error",
"detail": str(exc),
"code": 500,
}
)
# ---------------------------------------------------------
# ROUTES
# ---------------------------------------------------------
@app.get(
"/health",
response_model = HealthResponse,
summary = "Health check",
tags = ["System"],
)
async def health_check(request: Request) -> HealthResponse:
"""
Returns system health status.
Used by deployment platforms to verify the service is running.
Also useful for debugging - shows database sizes.
"""
pipeline = request.app.state.rag_pipeline
# Get Qdrant collection size
qdrant_size = pipeline.retriever.hybrid_retriever.qdrant.get_collection_size()
# Get BM25 index size
bm25_size = len(pipeline.retriever.hybrid_retriever.bm25.chunk_ids)
return HealthResponse(
status = "healthy",
model = "zai-org/GLM-5.1",
vector_db_size = qdrant_size,
bm25_index_size = bm25_size,
version = "1.0.0",
)
@app.post(
"/query/stream",
summary = "Stream query research papers",
tags = ["RAG"],
)
async def stream_query_papers(
request: Request,
query_input: QueryRequest,
):
import asyncio
pipeline = request.app.state.rag_pipeline
async def async_generator():
"""
Wraps the synchronous pipeline.stream_query() generator in an
async-friendly way using a thread + asyncio.Queue so we never
block the FastAPI event loop.
"""
loop = asyncio.get_event_loop()
queue: asyncio.Queue = asyncio.Queue()
SENTINEL = object()
def run_sync():
try:
for chunk in pipeline.stream_query(
question = query_input.question,
history = [ConversationTurn(role=t.role, content=t.content, citations=t.citations) for t in query_input.history],
top_k = query_input.top_k,
filter_category = query_input.filter_category,
filter_year_gte = query_input.filter_year_gte,
):
loop.call_soon_threadsafe(queue.put_nowait, chunk)
except Exception as e:
logger.error(f"Stream pipeline error: {e}", exc_info=True)
error_event = f'data: {json.dumps({"error": str(e)})}\n\n'
loop.call_soon_threadsafe(queue.put_nowait, error_event)
finally:
loop.call_soon_threadsafe(queue.put_nowait, SENTINEL)
import threading
thread = threading.Thread(target=run_sync, daemon=True)
thread.start()
while True:
item = await queue.get()
if item is SENTINEL:
break
yield item
return StreamingResponse(
async_generator(),
media_type="text/event-stream",
headers={
"Cache-Control": "no-cache",
"X-Accel-Buffering": "no",
}
)
@app.post(
"/feedback",
summary = "Submit feedback",
tags = ["System"],
)
async def submit_feedback(feedback: FeedbackRequest):
os.makedirs("logs", exist_ok=True)
with open("logs/feedback.jsonl", "a", encoding="utf-8") as f:
f.write(json.dumps(feedback.model_dump()) + "\n")
return {"status": "ok"}
@app.post(
"/query",
response_model = QueryResponse,
summary = "Query research papers",
tags = ["RAG"],
)
async def query_papers(
request: Request,
query_input: QueryRequest,
) -> QueryResponse:
"""
Submit a natural language question about ML research.
The system retrieves relevant paper excerpts and generates
a grounded answer with citations.
- **question**: Your research question (3-500 characters)
- **top_k**: Number of paper chunks to retrieve (1-20, default 5)
- **filter_category**: Filter by ArXiv category (e.g. cs.LG)
- **filter_year_gte**: Only include papers from this year onwards
"""
pipeline = request.app.state.rag_pipeline
logger.info(
f"Query received: '{query_input.question[:60]}' "
f"[top_k={query_input.top_k}]"
)
# Run the RAG pipeline in a thread pool
# WHY asyncio.to_thread:
# Our RAG pipeline is CPU-bound (not async).
# Running it directly in an async handler would BLOCK
# the entire FastAPI event loop - no other requests
# could be processed while one query is running.
# asyncio.to_thread runs it in a separate thread,
# keeping the event loop free for other requests.
try:
response = await asyncio.to_thread(
pipeline.query,
query_input.question,
[ConversationTurn(role=t.role, content=t.content, citations=t.citations) for t in query_input.history],
query_input.top_k,
query_input.filter_category,
query_input.filter_year_gte,
)
except Exception as e:
logger.error(f"RAG pipeline error: {e}")
raise HTTPException(status_code=500, detail=str(e))
# Convert RAGResponse dataclass to API schema
citations = [
CitationSchema(
paper_id = c.get("paper_id", ""),
title = c.get("title", ""),
authors = c.get("authors", []),
published_date = c.get("published_date", ""),
arxiv_url = c.get("arxiv_url", ""),
)
for c in response.citations
]
return QueryResponse(
answer = response.answer,
citations = citations,
query = response.query,
chunks_used = len(response.retrieved_chunks),
retrieval_time_ms = response.retrieval_time_ms,
generation_time_ms = response.generation_time_ms,
total_time_ms = response.total_time_ms,
has_context = response.has_context,
)
@app.get(
"/",
summary = "API root",
tags = ["System"],
)
async def root():
"""API root - confirms service is running."""
return {
"service": "ResearchPilot API",
"version": "1.0.0",
"docs": "/docs",
"health": "/health",
} |