get Youtube subtitles

yt-sbt.sh

#!/bin/bash

#sudo apt update
#sudo apt install ffmpeg
#pip install -U openai-whisper yt-dlp
#sudo chmod +x yt-sbt.sh

if [ $# -eq 0 ]; then
  echo "請提供 YouTube 影片 ID 作為參數。"
  exit 1
fi

video_id=$1
url="https://www.youtube.com/watch?v=$video_id"
output_file="${video_id}.txt"
langs=("en" "zh-TW" "zh")
subs_found=false

subs_file="/tmp/subs_$(date +%s%N).vtt"
audio_file="/tmp/audio_$(date +%s%N).mp3"
downloaded_subs_file="/tmp/downloaded_subs_$(date +%s%N).vtt"

for lang in "${langs[@]}"; do
  echo "正在嘗試下載 $lang 字幕..."
  if yt-dlp --write-subs --skip-download --sub-lang "$lang" --output "$subs_file" "$url"; then
    if [ -f "${subs_file}.${lang}.vtt" ]; then
      mv "${subs_file}.${lang}.vtt" "$downloaded_subs_file"
      subs_found=true
      break
    fi
  fi
done

if [ "$subs_found" = true ]; then
  echo "找到字幕，正在提取字幕文字..."
  sed '/^[0-9]/d' "$downloaded_subs_file" | sed '/^$/d' > "$output_file"
  echo "字幕已保存到 $output_file"
  rm "$downloaded_subs_file"
else
  echo "找不到字幕，正在下載音檔..."
  yt-dlp -f bestaudio --extract-audio --audio-format mp3 --output "$audio_file" "$url"
  echo "音檔下載完成，正在使用 Whisper 進行轉錄..."
  whisper --model small --output_dir /tmp --output_format txt "$audio_file"

  # 獲取轉錄結果的檔案路徑
  transcription_file="${audio_file%.*}.txt"
  if [ -f "$transcription_file" ]; then
    mv "$transcription_file" "$output_file"
    echo "轉錄已保存到 $output_file"
  else
    # 如果找不到轉錄結果檔案，顯示錯誤訊息並退出腳本
    echo "無法找到轉錄檔案。"
    exit 1
  fi
  rm "$audio_file"
fi