generacion_clusters_video_nuevo.py

from pathlib import Path
from sklearn.cluster import DBSCAN
import numpy as np
import json

class DataHub:
    def __init__(self, video_analysis_json_path: str):
        print("DataHub inicializando con JSON:", video_analysis_json_path)
        self.video = json.loads(Path(video_analysis_json_path).read_text(encoding='utf-8'))

class get_face_clusters:
    def __init__(self, data: DataHub):
        self.data = data
    
    def get_clusters(self, eps: float, min_samples: int):
        caras = self.data.video.get("caras", {})

        embeddings_caras = []

        for cara in caras:
            embeddings_caras.append(cara['embeddings'])

        X = np.array(embeddings_caras)

        clustering = DBSCAN(eps=eps, min_samples=min_samples, metric='euclidean').fit(X)
        labels_caras = clustering.labels_
        print(labels_caras)

        return labels_caras
    
class get_voices_clusters:
    def __init__(self, data: DataHub):
        self.data = data
    
    def get_clusters(self, eps: float, min_samples: int):
        voices = self.data.video.get("voices", {})

        embeddings_voices = []

        for voice in voices:
            embeddings_voices.append(voice['embeddings'])

        X = np.array(embeddings_voices)

        clustering = DBSCAN(eps=eps, min_samples=min_samples, metric='euclidean').fit(X)
        labels_voices = clustering.labels_
        print(labels_voices)

        return labels_voices

class get_scene_clusters:
    def __init__(self, data: DataHub):
        self.data = data
    
    def get_clusters(self, eps: float, min_samples: int):
        scenes = self.data.video.get("escenas", {})

        embeddings_scenes = []

        for scene in scenes:
            embeddings_scenes.append(scene['embeddings'])

        X = np.array(embeddings_scenes)

        clustering = DBSCAN(eps=eps, min_samples=min_samples, metric='euclidean').fit(X)
        labels_scenes = clustering.labels_
        print(labels_scenes)

        return labels_scenes