Upload api.py

api.py

@@ -246,24 +246,59 @@ def process_video_job(job_id: str):
                             })
                             saved_count += 1
                     else:
-                        # DeepFace fallback: no siempre devuelve boxes fácilmente; intentamos representaciones
+                        # DeepFace fallback con detección de bounding boxes vía Haar Cascade (OpenCV)
                         if DeepFace is None:
                             pass
                         else:
                             try:
-                                tmp_path = faces_root / f"frame_{frame_idx:06d}.jpg"
-                                cv2.imwrite(str(tmp_path), frame)
-                                reps = DeepFace.represent(img_path=str(tmp_path), model_name="Facenet512", enforce_detection=False)
-                                # reps puede ser lista de embeddings; no tenemos boxes -> guardamos frame completo como proxy
-                                for k, r in enumerate(reps or []):
-                                    emb = r.get("embedding") if isinstance(r, dict) else r
-                                    if emb is None:
-                                        continue
-                                    emb = np.array(emb, dtype=float)
-                                    emb = emb / (np.linalg.norm(emb) + 1e-9)
-                                    embeddings.append(emb.astype(float).tolist())
-                                    crops_meta.append({"file": tmp_path.name, "frame": frame_idx, "box": None})
-                                    saved_count += 1
+                                gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
+                                try:
+                                    haar_path = getattr(cv2.data, 'haarcascades', None) or ''
+                                    face_cascade = cv2.CascadeClassifier(os.path.join(haar_path, 'haarcascade_frontalface_default.xml'))
+                                except Exception:
+                                    face_cascade = None
+                                boxes_haar = []
+                                if face_cascade is not None and not face_cascade.empty():
+                                    faces_haar = face_cascade.detectMultiScale(gray, scaleFactor=1.1, minNeighbors=5, minSize=(40, 40))
+                                    for (x, y, w, h) in faces_haar:
+                                        top, left, bottom, right = max(0, y), max(0, x), min(frame.shape[0], y+h), min(frame.shape[1], x+w)
+                                        boxes_haar.append((top, right, bottom, left))
+                                if boxes_haar:
+                                    for (top, right, bottom, left) in boxes_haar:
+                                        crop = frame[top:bottom, left:right]
+                                        if crop.size == 0:
+                                            continue
+                                        fn = f"face_{frame_idx:06d}_{saved_count:03d}.jpg"
+                                        crop_path = faces_root / fn
+                                        cv2.imwrite(str(crop_path), crop)
+                                        reps = DeepFace.represent(img_path=str(crop_path), model_name="Facenet512", enforce_detection=False)
+                                        for r in (reps or []):
+                                            emb = r.get("embedding") if isinstance(r, dict) else r
+                                            if emb is None:
+                                                continue
+                                            emb = np.array(emb, dtype=float)
+                                            emb = emb / (np.linalg.norm(emb) + 1e-9)
+                                            embeddings.append(emb.astype(float).tolist())
+                                            crops_meta.append({
+                                                "file": fn,
+                                                "frame": frame_idx,
+                                                "box": [int(top), int(right), int(bottom), int(left)],
+                                            })
+                                            saved_count += 1
+                                else:
+                                    # Fallback mínimo: emb del frame completo
+                                    tmp_path = faces_root / f"frame_{frame_idx:06d}.jpg"
+                                    cv2.imwrite(str(tmp_path), frame)
+                                    reps = DeepFace.represent(img_path=str(tmp_path), model_name="Facenet512", enforce_detection=False)
+                                    for r in (reps or []):
+                                        emb = r.get("embedding") if isinstance(r, dict) else r
+                                        if emb is None:
+                                            continue
+                                        emb = np.array(emb, dtype=float)
+                                        emb = emb / (np.linalg.norm(emb) + 1e-9)
+                                        embeddings.append(emb.astype(float).tolist())
+                                        crops_meta.append({"file": tmp_path.name, "frame": frame_idx, "box": None})
+                                        saved_count += 1
                             except Exception as _e_df:
                                 print(f"[{job_id}] DeepFace fallback error: {_e_df}")
                 frame_idx += 1