Update main.py

main.py

@@ -2,14 +2,19 @@
 from __future__ import annotations
 import os, time, uuid, logging
 from typing import List, Optional, Dict, Any, Tuple
-import requests
+
 import numpy as np
+import requests
 from fastapi import FastAPI, BackgroundTasks, Header, HTTPException
 from pydantic import BaseModel, Field
 from qdrant_client import QdrantClient
 from qdrant_client.http.models import VectorParams, Distance, PointStruct
 
-logging.basicConfig(level=logging.INFO)
+# ---------- logging ----------
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(levelname)s:%(name)s:%(message)s"
+)
 LOG = logging.getLogger("remote_indexer")
 
 # ---------- ENV ----------
@@ -25,7 +30,10 @@ if not HF_TOKEN:
     LOG.warning("HF_API_TOKEN manquant — le service refusera /index et /query.")
 
 # ---------- Clients ----------
-qdr = QdrantClient(url=QDRANT_URL, api_key=QDRANT_API if QDRANT_API else None)
+try:
+    qdr = QdrantClient(url=QDRANT_URL, api_key=QDRANT_API if QDRANT_API else None)
+except Exception as e:
+    LOG.warning(f"Qdrant client init: {e}")
 
 # ---------- Pydantic ----------
 class FileIn(BaseModel):
@@ -45,7 +53,7 @@ class QueryRequest(BaseModel):
     query: str
     top_k: int = 6
 
-# ---------- Jobs store (en mémoire) ----------
+# ---------- Jobs store (mémoire) ----------
 JOBS: Dict[str, Dict[str, Any]] = {}  # {job_id: {"status": "...", "logs": [...], "created": ts}}
 
 # ---------- Utils ----------
@@ -62,13 +70,11 @@ def _post_embeddings(batch: List[str]) -> Tuple[np.ndarray, int]:
     r.raise_for_status()
     data = r.json()
     arr = np.array(data, dtype=np.float32)
-    # arr: [batch, dim]   (sentence-transformers)
-    #  ou  [batch, tokens, dim]  -> mean pooling
+    # [batch, dim] (sentence-transformers) ou [batch, tokens, dim] -> mean-pooling
     if arr.ndim == 3:
         arr = arr.mean(axis=1)
     if arr.ndim != 2:
         raise RuntimeError(f"Unexpected embeddings shape: {arr.shape}")
-    # normalisation
     norms = np.linalg.norm(arr, axis=1, keepdims=True) + 1e-12
     arr = arr / norms
     return arr.astype(np.float32), size
@@ -115,8 +121,7 @@ def run_index_job(job_id: str, req: IndexRequest):
         LOG.info(f"[{job_id}] Index start project={req.project_id} files={len(req.files)}")
         _append_log(job_id, f"Start project={req.project_id} files={len(req.files)}")
 
-        # premier batch pour récupérer la dimension
-        # on prépare un mini lot
+        # warmup pour dimension
         warmup = []
         for f in req.files[:1]:
             warmup.append(next(_chunk_with_spans(f.text, req.chunk_size, req.overlap))[2])
@@ -126,15 +131,8 @@ def run_index_job(job_id: str, req: IndexRequest):
         _ensure_collection(col, dim)
         _append_log(job_id, f"Collection ready: {col} (dim={dim})")
 
-        points_buffer: List[PointStruct] = []
         point_id = 0
 
-        def flush_points():
-            nonlocal points_buffer
-            if points_buffer:
-                qdr.upsert(collection_name=col, points=points_buffer)
-                points_buffer = []
-
         # boucle fichiers
         for fi, f in enumerate(req.files, 1):
             chunks, metas = [], []
@@ -167,7 +165,6 @@ def run_index_job(job_id: str, req: IndexRequest):
                 total_chunks += len(chunks)
                 _append_log(job_id, f"file {fi}/{len(req.files)}: +{len(chunks)} chunks (total={total_chunks}) ~{sz/1024:.1f}KiB")
 
-        flush_points()
         _append_log(job_id, f"Done. chunks={total_chunks}")
         _set_status(job_id, "done")
         LOG.info(f"[{job_id}] Index finished. chunks={total_chunks}")
@@ -179,6 +176,10 @@ def run_index_job(job_id: str, req: IndexRequest):
 # ---------- API ----------
 app = FastAPI()
 
+@app.get("/")
+def root():
+    return {"ok": True, "service": "remote-indexer", "docs": "/health, /index, /status/{job_id}, /query, /wipe"}
+
 @app.get("/health")
 def health():
     return {"ok": True}
@@ -217,7 +218,6 @@ def query(req: QueryRequest, x_auth_token: Optional[str] = Header(default=None))
     for p in res:
         pl = p.payload or {}
         txt = pl.get("text")
-        # hard cap snippet size
         if txt and len(txt) > 800:
             txt = txt[:800] + "..."
         out.append({"path": pl.get("path"), "chunk": pl.get("chunk"), "start": pl.get("start"), "end": pl.get("end"), "text": txt})
@@ -232,3 +232,10 @@ def wipe_collection(project_id: str, x_auth_token: Optional[str] = Header(defaul
         return {"ok": True}
     except Exception as e:
         raise HTTPException(400, f"wipe failed: {e}")
+
+# ---------- Entrypoint (respecte $PORT des Spaces) ----------
+if __name__ == "__main__":
+    import uvicorn
+    port = int(os.getenv("PORT", "7860"))
+    LOG.info(f"===== Application Startup on PORT {port} =====")
+    uvicorn.run(app, host="0.0.0.0", port=port)