video-maker/backend/app/app/core/video_utils.py

import pandas as pd
from pathlib import Path
import subprocess
import shutil
import os
import chardet
import zipfile
from io import BytesIO
from translate import Translator
from chardet.universaldetector import UniversalDetector

DEFAULT_ENCODING = "utf-8"

def guess_codec(filenames: list) -> str:
  codec_detector = UniversalDetector()
  for filename in filenames:
    codec_detector.feed(filename.encode('cp437'))
    if codec_detector.done:
      break

  result = codec_detector.close()
  encoding = result.get("encoding")
  return encoding or DEFAULT_ENCODING
  
def check_zip(zip_filepath:str):
  path = Path(zip_filepath)
  with zipfile.ZipFile(str(path)) as zf:
    filenames = [x for x in zf.namelist() if not x.endswith('/')]
    result = guess_codec(filenames)
    true_filenames = [x.encode('cp437').decode(result) for x in zf.namelist() if not x.endswith('/')]
    # print(true_filenames)
    scenarios_files = [(x, i) for i, x in enumerate(true_filenames) if Path(x).suffix in [".xlsx", ".csv"] and not Path(x).name.startswith("._") and Path(x).stem != "style"]
    # print(scenarios_files)
    
    if len(scenarios_files) == 0:
      raise ValueError("no excel or csv file in zip.")
    if len(scenarios_files) > 1:
      raise ValueError("too many excel or csv file in zip.")
    f = zf.read(filenames[scenarios_files[0][1]])
    if Path(scenarios_files[0][0]).suffix == ".xlsx":
      table = pd.read_excel(BytesIO(f), dtype=object)
    elif Path(scenarios_files[0][0]).suffix == ".csv":
      table = pd.read_csv(BytesIO(f), dtype=object)
    table.reset_index(inplace=True)
    # print(table)
    
    stems = [Path(x).stem for x in true_filenames]
    for i in range(len(table)):
      # excel 裡的圖檔跟zip裡的檔案要一致
      if table.loc[i, ['素材']].isna().item():
        img =  table.loc[i, ['素材']].item()
        print(img)

        img_files = [x.strip() for x in img.split(',')]
        for img in img_files:
          print(img)
          n = stems.count(img)
          if n == 0:
            raise ValueError(f"{img}: no such media file in zip.")
          elif n > 1:
            raise ValueError(f'too many same name media files as {img} in zip')
      
      # 需要tts文字或音檔
      if table.loc[i, ['字幕']].isna().item():
        if table.loc[i, ['音檔']].isna().item():
          raise ValueError(f'text or voice file is needed at scene {i+1}.')
        voice_file = table.loc[i, ['音檔']].item()
        n = stems.count(voice_file)
        if n != 1:
          raise ValueError(f"voice file is can't find is zip at scene {i+1}.")
      
def update_zip(zip_path, lang):
    temp_zip_path = zip_path + ".tmp"

    with zipfile.ZipFile(zip_path, 'r') as zip_in, zipfile.ZipFile(temp_zip_path, 'w') as zip_out:
        for item in zip_in.infolist():
            with zip_in.open(item.filename) as src_file:
                if item.filename.split('.')[-1] == "xlsx":
                    table = pd.read_excel(src_file, dtype=object)
                    table = translate_table(table, lang)
                    table.to_excel(Path(item.filename).name ,sheet_name='Sheet_name_1')
                    zip_out.write(Path(item.filename).name, item.filename)
                    os.remove(Path(item.filename).name)
                elif item.filename.split('.')[-1] == "csv":
                    table = pd.read_csv(src_file, dtype=object)
                    table = translate_table(table, lang)
                    table.to_excel(Path(item.filename).name ,sheet_name='Sheet_name_1')
                    zip_out.write(Path(item.filename).name, item.filename)
                    os.remove(Path(item.filename).name)
                else:
                    # それ以外のファイルはそのままコピー
                    with zip_out.open(item.filename, 'w') as dst_file:
                        shutil.copyfileobj(src_file, dst_file)

    # 旧ZIPを削除し、新ZIPをリネーム
    os.remove(zip_path)
    os.rename(temp_zip_path, zip_path)
    
def translate_table(table, lang):
    translator= Translator(to_lang=lang)
    print(f"translate to {lang}")
    for i in range(len(table)):
        if (text:=table.loc[i, ['大標']].item()):
            print("大標:",text)
            translation = translator.translate(text)
            print("大標翻譯:",translation)
            table.loc[i, ['字幕']] = translation
        if (text:=table.loc[i, ['字幕']].item()):
            print('字幕:',text)
            translation = translator.translate(text)
            print('字幕翻譯:',translation)
            table.loc[i, ['字幕']] = translation
    return table