| import re |
| from smolagents.tools import Tool |
|
|
|
|
| class YouTubeTranscriptTool(Tool): |
| name = "youtube_transcript" |
| description = ( |
| "Fetches the spoken transcript of a YouTube video. " |
| "Use this for ANY question about the content of a YouTube video — " |
| "what is said, shown, demonstrated, or described in it. " |
| "Pass the full YouTube URL or just the video ID." |
| ) |
| inputs = { |
| "url": { |
| "type": "string", |
| "description": ( |
| "The YouTube video URL (e.g. https://www.youtube.com/watch?v=XXXXXXXXXXX) " |
| "or bare video ID (e.g. XXXXXXXXXXX)." |
| ), |
| } |
| } |
| output_type = "string" |
|
|
| def __init__(self, *args, **kwargs): |
| self.is_initialized = False |
|
|
| def forward(self, url: str) -> str: |
| try: |
| from youtube_transcript_api import YouTubeTranscriptApi |
| from youtube_transcript_api._errors import ( |
| TranscriptsDisabled, |
| NoTranscriptFound, |
| VideoUnavailable, |
| CouldNotRetrieveTranscript, |
| ) |
| except ImportError: |
| return ( |
| "youtube-transcript-api is not installed. " |
| "Add 'youtube-transcript-api' to requirements.txt and rebuild." |
| ) |
|
|
| video_id = self._extract_video_id(url.strip()) |
| if not video_id: |
| return f"Could not extract a YouTube video ID from: {url}" |
|
|
| api = YouTubeTranscriptApi() |
|
|
| |
| |
| |
| try: |
| transcript_list = api.list(video_id) |
|
|
| try: |
| transcript = transcript_list.find_transcript( |
| ["en", "en-US", "en-GB"] |
| ) |
| except NoTranscriptFound: |
| |
| transcript = next(iter(transcript_list)) |
|
|
| entries = transcript.fetch() |
| |
| |
| try: |
| text = " ".join(e.text for e in entries) |
| except AttributeError: |
| text = " ".join(e["text"] for e in entries) |
|
|
| if len(text) > 12000: |
| text = text[:12000] + "\n\n[Transcript truncated at 12,000 characters]" |
|
|
| return text |
|
|
| except TranscriptsDisabled: |
| return ( |
| f"Transcripts are disabled for video {video_id}. " |
| "Try searching for the video title or description instead." |
| ) |
| except VideoUnavailable: |
| return f"Video {video_id} is unavailable or does not exist." |
| except NoTranscriptFound: |
| return ( |
| f"No transcript found for video {video_id}. " |
| "The video may not have captions." |
| ) |
| except CouldNotRetrieveTranscript as e: |
| return f"Could not retrieve transcript for {video_id}: {e}" |
| except Exception as e: |
| return f"Could not retrieve transcript for {video_id}: {e}" |
|
|
| @staticmethod |
| def _extract_video_id(url: str) -> str: |
| """Return the 11-character YouTube video ID from a URL or bare ID.""" |
| patterns = [ |
| r"(?:v=|youtu\.be/|embed/|shorts/)([A-Za-z0-9_-]{11})", |
| r"^([A-Za-z0-9_-]{11})$", |
| ] |
| for pattern in patterns: |
| match = re.search(pattern, url) |
| if match: |
| return match.group(1) |
| return "" |
