getTranscript

#!/usr/bin/env bash

URL="$1"
LANG="en"

if [ -z "$URL" ]; then
    echo "❌ Usage: $0 <YouTube URL>"
    exit 1
fi

TMP_DIR=$(mktemp -d)
TMP="$TMP_DIR/sub"

yt-dlp --write-auto-subs --sub-lang "$LANG" \
       --skip-download --sub-format vtt \
       -o "$TMP" "$URL" > /dev/null 2>&1

SUB_FILE=$(ls "$TMP_DIR"/*.vtt 2>/dev/null | head -n1)

if [ ! -f "$SUB_FILE" ]; then
    echo "❌ No subtitles found. Try checking the URL or language code."
    rm -rf "$TMP_DIR"
    exit 1
fi

CLEANED=$(mktemp)

awk '
    /^WEBVTT/    { next }
    /^NOTE/      { skip=1; next }
    /^Kind:/     { next }
    /^Language:/ { next }
    /^$/         { skip=0; next }
    skip         { next }
    /-->/        { next }
    {
        gsub(/<[^>]+>/, "")
        gsub(/&gt;/, ">")
        gsub(/&lt;/, "<")
        gsub(/&amp;/, "and")
        gsub(/&nbsp;/, " ")
        gsub(/\[[^\]]*\]/, "")
        gsub(/^[> \t]+/, "")
        gsub(/^[ \t]+|[ \t]+$/, "")
        gsub(/  +/, " ")
        if ($0 == "") next
        if ($0 != prev) {
            print $0
            prev = $0
        }
    }
' "$SUB_FILE" > "$CLEANED"

if command -v wl-copy >/dev/null && [ "$XDG_SESSION_TYPE" = "wayland" ]; then
    wl-copy < "$CLEANED"
    echo "✅ Copied using wl-copy (Wayland)"
elif command -v xclip >/dev/null; then
    xclip -selection clipboard < "$CLEANED"
    echo "✅ Copied using xclip (X11)"
elif command -v xsel >/dev/null; then
    xsel --clipboard < "$CLEANED"
    echo "✅ Copied using xsel"
else
    echo "⚠️  No clipboard tool found — printing to stdout:"
    cat "$CLEANED"
fi

rm -rf "$TMP_DIR"
rm -f "$CLEANED"