From 1af5f668558d993dc7ed93a70c9f68fe965893ff Mon Sep 17 00:00:00 2001
From: gilles <gil.soulier@gmail.com>
Date: Sun, 22 Feb 2026 12:11:58 +0100
Subject: [PATCH] fix(ai-service): gestion erreurs, health check, limite
 taille, run_in_executor

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
---
 ai-service/Dockerfile |  2 +-
 ai-service/main.py    | 49 ++++++++++++++++++++++++++++++-------------
 2 files changed, 36 insertions(+), 15 deletions(-)

diff --git a/ai-service/Dockerfile b/ai-service/Dockerfile
index 8f11fb8..b853bb5 100644
--- a/ai-service/Dockerfile
+++ b/ai-service/Dockerfile
@@ -1,4 +1,4 @@
-FROM python:3.11-slim
+FROM python:3.11.11-slim
 
 WORKDIR /app
 COPY requirements.txt .
diff --git a/ai-service/main.py b/ai-service/main.py
index 0cd96e8..3ec3d88 100644
--- a/ai-service/main.py
+++ b/ai-service/main.py
@@ -1,14 +1,16 @@
+import asyncio
 import io
 import os
 from typing import List
 
-from fastapi import FastAPI, File, UploadFile
-from PIL import Image
+from fastapi import FastAPI, File, HTTPException, UploadFile
+from PIL import Image, UnidentifiedImageError
 
 app = FastAPI(title="AI Plant Detection Service")
 
 _model = None
 MODEL_CACHE_DIR = os.environ.get("MODEL_CACHE_DIR", "/models")
+MAX_FILE_SIZE = 10 * 1024 * 1024  # 10 MB
 
 
 def get_model():
@@ -16,23 +18,16 @@ def get_model():
     if _model is None:
         from ultralytics import YOLO
         os.makedirs(MODEL_CACHE_DIR, exist_ok=True)
-        _model = YOLO("foduucom/plant-leaf-detection-and-classification")
+        try:
+            _model = YOLO("foduucom/plant-leaf-detection-and-classification")
+        except Exception as e:
+            raise RuntimeError(f"Impossible de charger le modèle YOLO: {e}") from e
     return _model
 
 
-@app.get("/health")
-def health():
-    return {"status": "ok"}
-
-
-@app.post("/detect")
-async def detect(file: UploadFile = File(...)):
-    data = await file.read()
-    img = Image.open(io.BytesIO(data)).convert("RGB")
-
+def _run_inference(img: Image.Image) -> list:
     model = get_model()
     results = model.predict(img, conf=0.25, iou=0.45, verbose=False)
-
     detections = []
     if results and results[0].boxes:
         boxes = results[0].boxes
@@ -45,3 +40,29 @@ async def detect(file: UploadFile = File(...)):
                 "confidence": round(conf, 3),
             })
     return detections
+
+
+@app.get("/health")
+def health():
+    return {"status": "ok", "model_loaded": _model is not None}
+
+
+@app.post("/detect")
+async def detect(file: UploadFile = File(...)):
+    data = await file.read()
+
+    if len(data) > MAX_FILE_SIZE:
+        raise HTTPException(status_code=413, detail="Fichier trop volumineux (max 10 MB)")
+
+    try:
+        img = Image.open(io.BytesIO(data)).convert("RGB")
+    except (UnidentifiedImageError, OSError) as e:
+        raise HTTPException(status_code=400, detail=f"Image invalide: {e}")
+
+    try:
+        loop = asyncio.get_event_loop()
+        detections = await loop.run_in_executor(None, _run_inference, img)
+    except RuntimeError as e:
+        raise HTTPException(status_code=503, detail=str(e))
+
+    return detections