server.py

import os
import io
import json
import traceback
from uuid import uuid4

# --- FastAPI & Web server imports ---
import uvicorn
from fastapi import FastAPI, File, UploadFile, HTTPException, Security, Depends
from fastapi.security import APIKeyHeader
from fastapi.responses import JSONResponse

# --- ML & Data processing imports ---
import torch
from PIL import Image
import numpy as np

# --- استيراد من ملفات المشروع الخاصة بك ---
try:
    from model import build_interfuser_model
    from logic import (
        transform, InterfuserController, ControllerConfig,
        Tracker, WAYPOINT_SCALE_FACTOR
    )
except ImportError as e:
    print(f"Error importing from project files: {e}")
    print("Please ensure model.py and logic.py are in the same directory.")
    exit()

# ==============================================================================
#           1. إعدادات الخادم، النموذج، والأمان
# ==============================================================================
app = FastAPI(
    title="Interfuser Driving API (Secure & Stateful)",
    description="An API for driving commands with session management and API key authentication.",
    version="2.0.0"
)

# --- تحميل النموذج (يتم مرة واحدة عند بدء التشغيل) ---
MODEL_NAME = "interfuser_baseline"
WEIGHTS_PATH = os.path.join("weights", f"{MODEL_NAME}.pth")
DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")
MODEL_CONFIG = {
    "rgb_backbone_name": "r50", "embed_dim": 256, "direct_concat": True,
    'get': lambda key, default: MODEL_CONFIG.get(key, default)
}

print(f"Loading model '{MODEL_NAME}' on device '{DEVICE}'...")
if not os.path.exists(WEIGHTS_PATH):
    raise FileNotFoundError(f"Weights file not found at: {WEIGHTS_PATH}")

model = build_interfuser_model(MODEL_CONFIG)
model.load_state_dict(torch.load(WEIGHTS_PATH, map_location=DEVICE))
model.to(DEVICE)
model.eval()
print("✅ Model loaded successfully!")

# --- إدارة الجلسات والأمان ---
SESSIONS = {}  # قاموس لتخزين حالات الجلسات: {session_id: Tracker}
API_KEY_NAME = "X-API-KEY"
api_key_header = APIKeyHeader(name=API_KEY_NAME, auto_error=False)

# في تطبيق حقيقي، يجب أن تكون هذه المفاتيح في متغيرات بيئة أو خدمة إدارة أسرار
VALID_API_KEYS = {
    "your-super-secret-key-for-flutter-app", # مفتاح لتطبيق فلاتر
    "a-different-key-for-testing"          # مفتاح آخر للاختبار
}

async def get_api_key(api_key: str = Security(api_key_header)):
    """تبعية للتحقق من أن مفتاح الـ API صالح."""
    if api_key in VALID_API_KEYS:
        return api_key
    else:
        raise HTTPException(
            status_code=403, detail="Could not validate credentials or missing API Key"
        )


# ==============================================================================
#           2. تعريف نقاط نهاية الـ API (Endpoints)
# ==============================================================================

# --- حماية جميع نقاط النهاية باستخدام التبعية ---
# أي طلب لأي نقطة نهاية أدناه يجب أن يجتاز get_api_key أولاً
app.dependency_overrides[get_api_key] = get_api_key


@app.post("/sessions/create", summary="Create a new tracking session")
async def create_session(api_key: str = Depends(get_api_key)):
    """
    ينشئ جلسة تتبع جديدة ويعيد معرفًا فريدًا لها.
    هذه هي الخطوة الأولى قبل إرسال بيانات الإطارات.
    """
    session_id = str(uuid4())
    SESSIONS[session_id] = {"tracker": Tracker(), "frame_count": 0}
    print(f"New session created: {session_id}")
    return JSONResponse(content={"session_id": session_id})


@app.post("/predict/{session_id}", summary="Run a single frame prediction for a session")
async def predict(
    session_id: str,
    rgb_image: UploadFile = File(..., description="Front-facing RGB camera image."),
    measurements_json: UploadFile = File(..., description="JSON file with vehicle measurements."),
    api_key: str = Depends(get_api_key)
):
    """
    يشغل التنبؤ لإطار واحد ضمن جلسة موجودة.
    يستخدم الـ Tracker المستمر الخاص بالجلسة لتتبع الأجسام عبر الزمن.
    """
    if session_id not in SESSIONS:
        raise HTTPException(status_code=404, detail="Session not found. Please create a new session.")

    session_data = SESSIONS[session_id]
    tracker = session_data["tracker"]
    session_data["frame_count"] += 1
    current_frame = session_data["frame_count"]

    try:
        # --- قراءة ومعالجة المدخلات ---
        image_bytes = await rgb_image.read()
        measurements_string = await measurements_json.read()
        rgb_pil = Image.open(io.BytesIO(image_bytes)).convert("RGB")
        m_dict = json.loads(measurements_string)

        # --- تجهيز التنسورات للنموذج ---
        front_tensor = transform(rgb_pil).unsqueeze(0).to(DEVICE)
        dummy_tensor = torch.zeros_like(front_tensor)
        measurements_tensor = torch.tensor([[
            m_dict.get(k, 0.0) for k in ['x', 'y', 'theta', 'speed', 'steer', 'throttle', 'brake', 'command', 'is_junction', 'should_brake']
        ]], dtype=torch.float32).to(DEVICE)
        target_point_tensor = torch.tensor([[0.0, 100.0]], dtype=torch.float32).to(DEVICE)

        inputs = {
            'rgb': front_tensor, 'rgb_left': dummy_tensor, 'rgb_right': dummy_tensor,
            'rgb_center': dummy_tensor, 'lidar': dummy_tensor,
            'measurements': measurements_tensor, 'target_point': target_point_tensor
        }

        # --- تشغيل النموذج والتحكم ---
        with torch.no_grad():
            outputs = model(inputs)
            traffic, waypoints, is_junction, traffic_light, stop_sign, _ = outputs

        traffic_np = traffic[0].detach().cpu().numpy().reshape(20, 20, -1)
        waypoints_np = waypoints[0].detach().cpu().numpy() * WAYPOINT_SCALE_FACTOR
        
        pos = [m_dict.get('x', 0.0), m_dict.get('y', 0.0)]
        theta = m_dict.get('theta', 0.0)

        # استخدام Tracker المستمر الخاص بالجلسة
        updated_traffic = tracker.update_and_predict(traffic_np.copy(), pos, theta, current_frame)
        
        controller = InterfuserController(ControllerConfig())
        steer, throttle, brake, _ = controller.run_step(
            m_dict.get('speed', 5.0), waypoints_np, is_junction.sigmoid()[0,1].item(),
            traffic_light.sigmoid()[0,0].item(), stop_sign.sigmoid()[0,1].item(), updated_traffic
        )

        # --- بناء وإرجاع الاستجابة ---
        control_commands = {"steer": float(steer), "throttle": float(throttle), "brake": bool(brake)}
        return JSONResponse(content={"status": "success", "control_commands": control_commands})

    except Exception as e:
        print(traceback.format_exc())
        raise HTTPException(status_code=500, detail=f"An internal error occurred: {str(e)}")


@app.delete("/sessions/{session_id}", summary="Delete a tracking session")
async def delete_session(session_id: str, api_key: str = Depends(get_api_key)):
    """
    يحذف جلسة تتبع لتحرير الموارد على الخادم.
    """
    if session_id in SESSIONS:
        del SESSIONS[session_id]
        print(f"Session deleted: {session_id}")
        return JSONResponse(content={"message": "Session deleted successfully."})
    raise HTTPException(status_code=404, detail="Session not found.")


# ==============================================================================
#           3. نقطة بداية تشغيل الخادم
# ==============================================================================
if __name__ == "__main__":
    print("--- Interfuser API Server ---")
    print("API documentation will be available at http://127.0.0.1:8000/docs")
    uvicorn.run(app, host="0.0.0.0", port=8000)