Create app.py

app.py

@@ -0,0 +1,103 @@
+from flask import Flask, request, jsonify, make_response
+from flask_cors import CORS
+import numpy as np
+import librosa
+import io
+import os
+
+app = Flask(__name__)
+CORS(app, resources={r"/api/*": {"origins": "*"}})
+
+@app.after_request
+def after_request(response):
+    response.headers.add('Access-Control-Allow-Origin', '*')
+    response.headers.add('Access-Control-Allow-Headers', '*')
+    response.headers.add('Access-Control-Allow-Methods', 'GET,POST,OPTIONS')
+    return response
+
+@app.route('/', methods=['GET'])
+def home():
+    return jsonify({'message': 'AcoustiTrack API - CNN-1D VAD + GCC-PHAT DOA', 'status': 'running'})
+
+@app.route('/api/health', methods=['GET', 'OPTIONS'])
+def health():
+    if request.method == 'OPTIONS':
+        return make_response('', 204)
+    return jsonify({'status': 'ok', 'model': 'CNN-1D VAD + GCC-PHAT DOA'})
+
+@app.route('/api/analyze', methods=['POST', 'OPTIONS'])
+def analyze():
+    if request.method == 'OPTIONS':
+        return make_response('', 204)
+    try:
+        if 'audio' not in request.files:
+            return jsonify({'error': 'Aucun fichier audio'}), 400
+        
+        audio_file = request.files['audio']
+        audio_bytes = audio_file.read()
+        
+        y, sr = librosa.load(io.BytesIO(audio_bytes), sr=16000, mono=False)
+        
+        if y.ndim == 1:
+            y = np.tile(y, (4, 1))
+        
+        n_channels = y.shape[0] if y.ndim > 1 else 1
+        duration = y.shape[-1] / sr
+        
+        # ============================================
+        # VOTRE MODELE CNN-1D ICI
+        # Chargez votre modele et faites l'inference
+        # ============================================
+        vad_prediction = "voice"
+        vad_confidence = 0.962
+        doa_angle = -25.5
+        
+        n_frames = max(1, int(duration * 10))
+        vad_timeline = []
+        for i in range(n_frames):
+            vad_timeline.append({
+                'start': round(i * 0.1, 2),
+                'end': round((i + 1) * 0.1, 2),
+                'label': 'voice' if np.random.random() > 0.3 else 'silence',
+                'confidence': round(0.85 + np.random.random() * 0.15, 3)
+            })
+        
+        doa_trajectory = []
+        for i in range(n_frames):
+            doa_trajectory.append({
+                'time': round(i * 0.1, 2),
+                'angle': round(doa_angle + np.random.randn() * 5, 1)
+            })
+        
+        voice_frames = sum(1 for f in vad_timeline if f['label'] == 'voice')
+        
+        return jsonify({
+            'success': True,
+            'metadata': {
+                'channels': int(n_channels),
+                'sampleRate': int(sr),
+                'duration': round(duration, 2),
+                'samples': int(y.shape[-1])
+            },
+            'vad': {
+                'prediction': vad_prediction,
+                'confidence': vad_confidence,
+                'timeline': vad_timeline
+            },
+            'doa': {
+                'angle': doa_angle,
+                'confidence': 0.89,
+                'trajectory': doa_trajectory
+            },
+            'metrics': {
+                'snr': 18.5,
+                'voiceRatio': round(voice_frames / len(vad_timeline), 2),
+                'meanDoa': round(float(np.mean([d['angle'] for d in doa_trajectory])), 1)
+            }
+        })
+    except Exception as e:
+        return jsonify({'error': str(e)}), 500
+
+if __name__ == '__main__':
+    port = int(os.environ.get('PORT', 7860))
+    app.run(host='0.0.0.0', port=port)