import streamlit as st from faster_whisper import WhisperModel from pyannote.audio import Pipeline import pyannote.core from collections import defaultdict # Función para asignar hablantes a segmentos de transcripción def assign_speakers_to_segments(diarization, transcription_segments): speaker_segments = [] # Convertir diarización en un diccionario con los tiempos de inicio y fin de cada hablante diarization_dict = defaultdict(list) for segment, _, speaker in diarization.itertracks(yield_label=True): diarization_dict[speaker].append((segment.start, segment.end)) for transcription_segment in transcription_segments: start, end = transcription_segment.start, transcription_segment.end speakers_count = defaultdict(float) # Contar la duración de cada hablante dentro del segmento de transcripción for speaker, times in diarization_dict.items(): for seg_start, seg_end in times: # Calcular la intersección del tiempo de hablante con el segmento de transcripción overlap_start = max(start, seg_start) overlap_end = min(end, seg_end) overlap_duration = max(0, overlap_end - overlap_start) speakers_count[speaker] += overlap_duration # Elegir el hablante con la mayor duración total en el segmento if speakers_count: speaker = max(speakers_count, key=speakers_count.get) else: speaker = "Unknown" # Añadir el texto del segmento de transcripción y el hablante correspondiente speaker_segments.append((speaker, transcription_segment.text)) return speaker_segments # Función principal de la aplicación Streamlit def main(): logo_url = "https://wallnergroup.com/wp-content/uploads/2023/06/logo-wallner-2048x270.png" st.image(logo_url) st.markdown("""