main.py - main - bioxakep/VoiceTrans

main.py
130 строк5 KB
bioxakep
refactor
13 авг 2025, 23:54
13 авг 2025, 23:546aa60b3
import argparse
import os

from transcribation import smart_transcriber, TranscribeResult
from utils_common import humanize_time, rename_sounds, get_phones
from voice_engines import VoskRecognizer


arg_parser = argparse.ArgumentParser()
arg_parser.add_argument(
    "sounds_dir_path",
    type=str,
    help="Директория со звуковыми записями",
)
arg_parser.add_argument(
    "-r",
    "--records_file_path",
    type=str,
    help="Файл со статистикой соединений",
    default=None,
)
arg_parser.add_argument(
    "-t",
    "--tags_file_path",
    type=str,
    help="Файл с поисковыми признаками (тэгами)",
    default=None,
)


def rename_and_transcribe(
    records_file_path: str,
    sounds_dir_path: str | None = None,
    tags_file_path: str | None = None,
):
    if not os.path.exists(records_file_path):
        raise FileExistsError("Файл со статистикой соединений не найден")
    vosk_recognizer = VoskRecognizer()
    if sounds_dir_path is None:
        sounds_dir_path = os.path.join(
            os.path.dirname(records_file_path),
            os.path.basename(records_file_path).split(".")[0],
        )
        if not os.path.exists(sounds_dir_path):
            raise FileExistsError("Директория со звуком не найдена")
        print(
            "Директория со звуком определена из имени файла со статистикой соединений"
        )
    trans_files_count = 0
    total_sound_duration: int = 0
    total_speech_duration: int = 0
    total_trans_time: int = 0
    for sound_file_path, abonent, contact, sound_dt in rename_sounds(
        billing_file_path=records_file_path,
        sounds_dir=sounds_dir_path,
    ):
        trans_result = smart_transcriber(
            audio_file_path=sound_file_path,
            recognizer=vosk_recognizer,
            tags_file_path=tags_file_path,
            abonent=abonent,
            contact=contact,
            sound_dt=sound_dt,
        )
        if isinstance(trans_result, TranscribeResult):
            total_trans_time += trans_result.trans_time
            total_speech_duration += trans_result.speech_duration
            total_sound_duration += trans_result.sound_duration
            trans_files_count += 1
    sound_duration_hum = humanize_time(total_sound_duration, clocks=False)
    speech_duration_hum = humanize_time(total_speech_duration, clocks=False)
    trans_time_hum = humanize_time(total_trans_time, clocks=False)
    print(
        f"Транскрибировано файлов: {trans_files_count}\n"
        f"Суммарная длительность аудиофайлов: {sound_duration_hum}\n"
        f"Суммарная длительность выделенной речи: {speech_duration_hum}\n"
        f"Общее время транскрибации: {trans_time_hum}"
    )


def just_transcribe(
    sounds_dir_path: str,
    tags_file_path: str | None = None,
):
    vosk_recognizer = VoskRecognizer()
    if not os.path.exists(sounds_dir_path):
        raise FileExistsError("Директория со звуком не найдена")
    trans_files_count = 0
    total_sound_duration: int = 0
    total_speech_duration: int = 0
    total_trans_time: int = 0
    for f in os.listdir(sounds_dir_path):
        abonent, contact = get_phones(f)
        file_path: str = os.path.join(sounds_dir_path, f)
        trans_result = smart_transcriber(
            audio_file_path=file_path,
            recognizer=vosk_recognizer,
            tags_file_path=tags_file_path,
            abonent=abonent,
            contact=contact,
        )
        if isinstance(trans_result, TranscribeResult):
            total_trans_time += trans_result.trans_time
            total_speech_duration += trans_result.speech_duration
            total_sound_duration += trans_result.sound_duration
            trans_files_count += 1
    sound_duration_hum = humanize_time(total_sound_duration, clocks=False)
    speech_duration_hum = humanize_time(total_speech_duration, clocks=False)
    trans_time_hum = humanize_time(total_trans_time, clocks=False)
    print(
        f"Транскрибировано файлов: {trans_files_count}\n"
        f"Суммарная длительность аудиофайлов: {sound_duration_hum}\n"
        f"Суммарная длительность выделенной речи: {speech_duration_hum}\n"
        f"Общее время транскрибации: {trans_time_hum}"
    )


if __name__ == "__main__":
    args = arg_parser.parse_args()
    if args.records_file_path:
        rename_and_transcribe(
            records_file_path=args.records_file_path,
            sounds_dir_path=args.sounds_dir_path,
            tags_file_path=args.tags_file_path,
        )
    else:
        just_transcribe(
            sounds_dir_path=args.sounds_dir_path,
            tags_file_path=args.tags_file_path,
        )