Create app.py

app.py

@@ -0,0 +1,82 @@
+import shutil
+import torch
+from fastapi import FastAPI, UploadFile, File
+from PIL import Image
+from io import BytesIO
+from ultralytics import YOLO
+from huggingface_hub import hf_hub_download
+from transformers import AutoProcessor, AutoModelForCausalLM
+
+# --- 1. إعداد تطبيق FastAPI ---
+app = FastAPI(title="Object Detection & Captioning API")
+
+# --- 2. تحميل الموديلات (يتم لمرة واحدة عند بدء التشغيل) ---
+# تحميل موديل YOLO
+model_path = hf_hub_download(
+    repo_id="GradTeam/yolov26-objectDetection",
+    filename="best.pt"
+)
+yolo_model = YOLO(model_path)
+
+# تحميل موديل الوصف (GIT)
+device = "cuda" if torch.cuda.is_available() else "cpu"
+processor = AutoProcessor.from_pretrained("microsoft/git-large")
+git_model = AutoModelForCausalLM.from_pretrained("microsoft/git-large").to(device)
+
+# --- 3. الدوال المساعدة ---
+
+def get_yolo_detections(image_path):
+    results = yolo_model(image_path)
+    objects = []
+    for r in results:
+        boxes = r.boxes.xyxy.tolist()
+        classes = r.boxes.cls.tolist()
+        for box, cls in zip(boxes, classes):
+            name = yolo_model.names[int(cls)]
+            objects.append({
+                "name": name,
+                "box": [round(coord, 2) for coord in box] # تقريب الإحداثيات
+            })
+    return objects
+
+def get_image_caption(image_path, objects):
+    image = Image.open(image_path).convert("RGB")
+    names = [obj["name"] for obj in objects]
+    
+    # بناء الـ Prompt بناءً على الأجسام المكتشفة
+    text_prompt = "Objects detected: " + ", ".join(names) if names else "Describe this image."
+    
+    inputs = processor(images=image, text=text_prompt, return_tensors="pt").to(device)
+    generated_ids = git_model.generate(**inputs, max_length=50)
+    caption = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
+    return caption
+
+# --- 4. نقطة النهاية (API Endpoint) ---
+
+@app.post("/analyze")
+async def analyze_image(file: UploadFile = File(...)):
+    # حفظ الملف المرفوع مؤقتاً
+    temp_path = "temp_image.jpg"
+    with open(temp_path, "wb") as buffer:
+        shutil.copyfileobj(file.file, buffer)
+    
+    try:
+        # 1. تنفيذ كشف الأجسام
+        detected_objects = get_yolo_detections(temp_path)
+        
+        # 2. تنفيذ وصف الصورة بناءً على الأجسام
+        description = get_image_caption(temp_path, detected_objects)
+        
+        return {
+            "status": "success",
+            "detected_objects_count": len(detected_objects),
+            "objects": detected_objects,
+            "description": description
+        }
+    except Exception as e:
+        return {"status": "error", "message": str(e)}
+
+# --- 5. التشغيل (اختياري محلياً) ---
+if name == "__main__":
+    import uvicorn
+    uvicorn.run(app, host="0.0.0.0", port=7860)