Reconnaissance / app.py
Docfile's picture
Update app.py
93b0af7 verified
raw
history blame
10.2 kB
from flask import Flask, render_template, request, jsonify, send_from_directory
from flask_cors import CORS
from flask_limiter import Limiter
from flask_limiter.util import get_remote_address
from deepface import DeepFace
from werkzeug.utils import secure_filename
import os
import tempfile
import shutil
import uuid
import logging
import time
from datetime import datetime
from functools import wraps
import numpy as np
import cv2
from PIL import Image
import io
import threading
import queue
import hashlib
# Configuration du logging
logging.basicConfig(
filename='app.log',
level=logging.INFO,
format='%(asctime)s - %(levelname)s - %(message)s'
)
class FaceAnalysisApp:
def __init__(self):
self.app = Flask(__name__, static_folder='static')
self.setup_app()
def setup_app(self):
# Configuration de base
self.app.config['UPLOAD_FOLDER'] = 'static/uploads'
self.app.config['MAX_CONTENT_LENGTH'] = 16 * 1024 * 1024
self.app.config['ALLOWED_EXTENSIONS'] = {'png', 'jpg', 'jpeg', 'gif'}
self.app.config['SECRET_KEY'] = os.urandom(24)
# Initialisation des composants
CORS(self.app)
self.limiter = Limiter(
self.app,
key_func=get_remote_address,
default_limits=["200 per day", "50 per hour"]
)
# File d'attente pour le traitement asynchrone
self.task_queue = queue.Queue()
self.start_worker_thread()
# Cache pour les résultats
self.results_cache = {}
def start_worker_thread(self):
def worker():
while True:
task = self.task_queue.get()
if task is None:
break
try:
task()
except Exception as e:
logging.error(f"Error in worker thread: {str(e)}")
self.task_queue.task_done()
self.worker_thread = threading.Thread(target=worker, daemon=True)
self.worker_thread.start()
def timing_decorator(self, f):
@wraps(f)
def wrap(*args, **kwargs):
start = time.time()
result = f(*args, **kwargs)
end = time.time()
logging.info(f'{f.__name__} took {end-start:.2f} seconds to execute')
return result
return wrap
def validate_image(self, image_stream):
"""Valide et optimise l'image"""
try:
img = Image.open(image_stream)
# Vérification des dimensions
if img.size[0] > 2000 or img.size[1] > 2000:
img.thumbnail((2000, 2000), Image.LANCZOS)
# Conversion en RGB si nécessaire
if img.mode not in ('RGB', 'L'):
img = img.convert('RGB')
# Optimisation
output = io.BytesIO()
img.save(output, format='JPEG', quality=85, optimize=True)
output.seek(0)
return output
except Exception as e:
logging.error(f"Image validation error: {str(e)}")
raise ValueError("Invalid image format")
def process_face_detection(self, image_path):
"""Détection de visage avec mise en cache"""
image_hash = hashlib.md5(open(image_path, 'rb').read()).hexdigest()
if image_hash in self.results_cache:
return self.results_cache[image_hash]
try:
result = DeepFace.analyze(
img_path=image_path,
actions=['age', 'gender', 'race', 'emotion'],
enforce_detection=True
)
self.results_cache[image_hash] = result
return result
except Exception as e:
logging.error(f"Face detection error: {str(e)}")
raise
@timing_decorator
def verify_faces(self, image1_path, image2_path):
"""Comparaison des visages avec vérification approfondie"""
try:
# Vérification initiale de la présence de visages
face1 = cv2.imread(image1_path)
face2 = cv2.imread(image2_path)
if face1 is None or face2 is None:
raise ValueError("Unable to read one or both images")
result = DeepFace.verify(
img1_path=image1_path,
img2_path=image2_path,
enforce_detection=True,
model_name="VGG-Face"
)
# Enrichissement des résultats
result['timestamp'] = datetime.now().isoformat()
result['confidence_score'] = 1 - result.get('distance', 0)
result['processing_time'] = time.time()
return result
except Exception as e:
logging.error(f"Face verification error: {str(e)}")
raise
def setup_routes(self):
@self.app.route('/')
def index():
return render_template('index.html')
@self.app.route('/verify', methods=['POST'])
@self.limiter.limit("10 per minute")
def verify_faces_endpoint():
try:
if 'image1' not in request.files or 'image2' not in request.files:
return jsonify({'error': 'Two images are required'}), 400
image1 = request.files['image1']
image2 = request.files['image2']
# Validation des images
try:
image1_stream = self.validate_image(image1)
image2_stream = self.validate_image(image2)
except ValueError as e:
return jsonify({'error': str(e)}), 400
# Création des fichiers temporaires
with tempfile.TemporaryDirectory() as temp_dir:
image1_path = os.path.join(temp_dir, secure_filename(image1.filename))
image2_path = os.path.join(temp_dir, secure_filename(image2.filename))
# Sauvegarde des images optimisées
with open(image1_path, 'wb') as f:
f.write(image1_stream.getvalue())
with open(image2_path, 'wb') as f:
f.write(image2_stream.getvalue())
# Analyse des visages
result = self.verify_faces(image1_path, image2_path)
# Sauvegarde permanente si nécessaire
if result['verified']:
permanent_dir = os.path.join(self.app.static_folder, 'verified_faces')
os.makedirs(permanent_dir, exist_ok=True)
# Génération de noms uniques
timestamp = datetime.now().strftime('%Y%m%d_%H%M%S')
image1_name = f"face1_{timestamp}_{uuid.uuid4().hex[:8]}.jpg"
image2_name = f"face2_{timestamp}_{uuid.uuid4().hex[:8]}.jpg"
shutil.copy2(image1_path, os.path.join(permanent_dir, image1_name))
shutil.copy2(image2_path, os.path.join(permanent_dir, image2_name))
result['image1_url'] = f'/static/verified_faces/{image1_name}'
result['image2_url'] = f'/static/verified_faces/{image2_name}'
return jsonify(result)
except Exception as e:
logging.error(f"Verification endpoint error: {str(e)}")
return jsonify({'error': 'An internal error occurred'}), 500
@self.app.route('/analyze', methods=['POST'])
@self.limiter.limit("20 per minute")
def analyze_face_endpoint():
try:
if 'image' not in request.files:
return jsonify({'error': 'No image provided'}), 400
image = request.files['image']
# Validation de l'image
try:
image_stream = self.validate_image(image)
except ValueError as e:
return jsonify({'error': str(e)}), 400
# Traitement asynchrone
result_queue = queue.Queue()
def process_task():
try:
with tempfile.NamedTemporaryFile(suffix='.jpg', delete=False) as temp_file:
temp_file.write(image_stream.getvalue())
result = self.process_face_detection(temp_file.name)
result_queue.put(('success', result))
except Exception as e:
result_queue.put(('error', str(e)))
finally:
try:
os.unlink(temp_file.name)
except:
pass
self.task_queue.put(process_task)
# Attente du résultat avec timeout
try:
status, result = result_queue.get(timeout=30)
if status == 'error':
return jsonify({'error': result}), 500
return jsonify(result)
except queue.Empty:
return jsonify({'error': 'Processing timeout'}), 408
except Exception as e:
logging.error(f"Analysis endpoint error: {str(e)}")
return jsonify({'error': 'An internal error occurred'}), 500
@self.app.errorhandler(413)
def request_entity_too_large(error):
return jsonify({'error': 'File too large'}), 413
@self.app.errorhandler(429)
def ratelimit_handler(e):
return jsonify({'error': 'Rate limit exceeded'}), 429
def run(self, host='0.0.0.0', port=5000, debug=False):
self.setup_routes()
self.app.run(host=host, port=port, debug=debug)
if __name__ == '__main__':
app = FaceAnalysisApp()
app.run(debug=True)