transcript

#!/bin/bash

# transcript - Get YouTube video transcript for free
# Usage: transcript [--all] <url>

GET_ALL=0
LANG="en" # Default language
URL=""

# Parse arguments
while [[ $# -gt 0 ]]; do
  case $1 in
    --all)
      GET_ALL=1
      shift
      ;;
    --lang)
      LANG="$2"
      shift
      shift
      ;;
    http://*|https://*)
      URL="$1"
      shift
      ;;
    *)
      echo "Unknown argument: $1"
      shift
      ;;
  esac
done

if [ -z "$URL" ]; then
    echo "Usage: transcript [--all] <url>"
    exit 1
fi

# Resolve script directory (handling symlinks)
SOURCE="${BASH_SOURCE[0]}"
while [ -h "$SOURCE" ]; do
  DIR="$( cd -P "$( dirname "$SOURCE" )" >/dev/null 2>&1 && pwd )"
  SOURCE="$(readlink "$SOURCE")"
  [[ $SOURCE != /* ]] && SOURCE="$DIR/$SOURCE"
done
SCRIPT_DIR="$( cd -P "$( dirname "$SOURCE" )" >/dev/null 2>&1 && pwd )"

# Prefer local binary, then python module, then system command
if [ -f "$SCRIPT_DIR/yt-dlp" ]; then
    YTDLP="$SCRIPT_DIR/yt-dlp"
    IMPERSONATE=""
elif python3 -m yt_dlp --version &> /dev/null; then
    YTDLP="python3 -m yt_dlp"
    # Enable impersonation if using the python module (more likely to have curl_cffi)
    IMPERSONATE="--impersonate Chrome"
elif command -v yt-dlp &> /dev/null; then
    YTDLP="yt-dlp"
    IMPERSONATE=""
else
    echo "yt-dlp not found. Downloading latest standalone binary..." >&2
    CURL_URL="https://github.com/yt-dlp/yt-dlp/releases/latest/download/yt-dlp"
    DOWNLOAD_PATH="$SCRIPT_DIR/yt-dlp"
    
    if curl -L "$CURL_URL" -o "$DOWNLOAD_PATH"; then
        chmod +x "$DOWNLOAD_PATH"
        YTDLP="$DOWNLOAD_PATH"
        IMPERSONATE=""
        echo "yt-dlp downloaded successfully to $SCRIPT_DIR." >&2
    else
        echo "Error: Failed to download yt-dlp. Please install it manually."
        exit 1
    fi
fi

if ! command -v jq &> /dev/null; then
    echo "Error: jq is not installed. Please install it first."
    exit 1
fi

TEMP_DIR=$(mktemp -d)
trap 'rm -rf "$TEMP_DIR"' EXIT

TITLE=""
DESCRIPTION=""

if [ $GET_ALL -eq 1 ]; then
    # echo "Fetching metadata..." >&2
    TITLE=$($YTDLP $IMPERSONATE --get-title "$URL" 2>/dev/null)
    DESCRIPTION=$($YTDLP $IMPERSONATE --get-description "$URL" 2>/dev/null)
fi

# Download auto-generated subtitles in VTT format
# Strategy: Try requested lang first, then try with cookies.
# Only fallback to other languages if the user didn't specify a strict custom language (default en).

# echo "Fetching transcript (Attempt 1: $LANG)..." >&2
$YTDLP $IMPERSONATE --write-auto-subs --sub-lang "$LANG" --skip-download \
    --output "$TEMP_DIR/sub.%(ext)s" "$URL" > /dev/null 2>&1

# If download failed, try with cookies
if [ -z "$(find "$TEMP_DIR" -name "sub.*.vtt" | head -n 1)" ]; then
    # echo "Attempt 2: $LANG with cookies..." >&2
    $YTDLP $IMPERSONATE --write-auto-subs --sub-lang "$LANG" --skip-download \
        --cookies-from-browser "chromium:/home/lewis/.config/chromium-chatbot" \
        --output "$TEMP_DIR/sub.%(ext)s" "$URL" > /dev/null 2>&1
fi

# Fallbacks (Only if using default English, otherwise respect user choice)
if [ "$LANG" == "en" ]; then
    # If STILL failed, fetch RU (high success rate for this specific video and others)
    if [ -z "$(find "$TEMP_DIR" -name "sub.*.vtt" | head -n 1)" ]; then
        # echo "Attempt 3: RU..." >&2
        $YTDLP $IMPERSONATE --write-auto-subs --sub-lang ru --skip-download \
            --output "$TEMP_DIR/sub.%(ext)s" "$URL" > /dev/null 2>&1
    fi

    # If STILL failed, fetch any available language
    if [ -z "$(find "$TEMP_DIR" -name "sub.*.vtt" | head -n 1)" ]; then
        # echo "Attempt 4: Any language..." >&2
        $YTDLP $IMPERSONATE --write-auto-subs --sub-lang ".*" --skip-download \
            --output "$TEMP_DIR/sub.%(ext)s" "$URL" > /dev/null 2>&1
    fi
fi

# Find the downloaded subtitle file
SUB_FILE=$(find "$TEMP_DIR" -name "sub.*.vtt" | head -n 1)

if [ ! -f "$SUB_FILE" ]; then
    # Try manual subtitles (standard attempt)
    "$YTDLP" --write-subs --skip-download --output "$TEMP_DIR/sub.%(ext)s" "$URL" > /dev/null 2>&1
    SUB_FILE=$(find "$TEMP_DIR" -name "sub.*.vtt" | head -n 1)
    
    # Retry manual subtitles with cookies if needed
    if [ ! -f "$SUB_FILE" ]; then
         "$YTDLP" --write-subs --skip-download \
            --cookies-from-browser "chromium:/home/lewis/.config/chromium-chatbot" \
            --output "$TEMP_DIR/sub.%(ext)s" "$URL" > /dev/null 2>&1
         SUB_FILE=$(find "$TEMP_DIR" -name "sub.*.vtt" | head -n 1)
    fi
fi

if [ -f "$SUB_FILE" ]; then
    # Process VTT to plain text as a single paragraph
    TRANSCRIPT=$(sed -E \
        -e '/^[0-9]{2}:[0-9]{2}:[0-9]{2}\.[0-9]{3} --> [0-9]{2}:[0-9]{2}:[0-9]{2}\.[0-9]{3}.*/d' \
        -e 's/<[^>]*>//g' \
        -e '/^(WEBVTT|Kind:|Language:)/d' \
        -e '/^[[:space:]]*$/d' \
        "$SUB_FILE" | uniq | tr '\n' ' ' | sed 's/  */ /g' | sed 's/^ //' | sed 's/ $//')
    
    if [ $GET_ALL -eq 1 ]; then
        # Output as JSON
        jq -n \
            --arg title "$TITLE" \
            --arg description "$DESCRIPTION" \
            --arg transcript "$TRANSCRIPT" \
            '{title: $title, description: $description, transcript: $transcript}'
    else
        # Output as plain text
        echo "$TRANSCRIPT"
    fi
else
    echo "Error: Could not retrieve transcript for $URL"
    exit 1
fi