From a8470c89a2291f6edd65bd8294f2e3212aff50e2 Mon Sep 17 00:00:00 2001 From: Rodrigo Nader Date: Tue, 3 Dec 2024 12:46:50 -0300 Subject: [PATCH] remove unnecessary function from youtube transcript (#4978) * remove unnecessary function from youtube transctipt * [autofix.ci] apply automated fixes --------- Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com> Co-authored-by: Eric Hare --- .../components/tools/youtube_transcripts.py | 37 ------------------- 1 file changed, 37 deletions(-) diff --git a/src/backend/base/langflow/components/tools/youtube_transcripts.py b/src/backend/base/langflow/components/tools/youtube_transcripts.py index a276b785e..04b671bc1 100644 --- a/src/backend/base/langflow/components/tools/youtube_transcripts.py +++ b/src/backend/base/langflow/components/tools/youtube_transcripts.py @@ -225,40 +225,3 @@ class YouTubeTranscriptsComponent(Component): except Exception as exc: # noqa: BLE001 # Using a specific error type for the return value return Data(data={"error": f"Failed to get YouTube transcripts: {exc!s}"}) - - def youtube_transcripts( - self, - url: str = "", - transcript_format: TranscriptFormat = TranscriptFormat.TEXT, - chunk_size_seconds: int = 120, - language: str = "", - translation: str = "", - ) -> Data | list[Data]: - """Helper method to handle transcripts outside of component calls. - - Args: - url: The YouTube URL to get transcripts from. - transcript_format: Format of transcripts ('text' or 'chunks'). - chunk_size_seconds: Size of each transcript chunk in seconds. - language: Comma-separated list of language codes. - translation: Target language for translation. - - Returns: - Data | list[Data]: Video transcripts as single Data or list of Data. - """ - if isinstance(transcript_format, str): - transcript_format = TranscriptFormat(transcript_format) - loader = YoutubeLoader.from_youtube_url( - url, - transcript_format=TranscriptFormat.TEXT - if transcript_format == TranscriptFormat.TEXT - else TranscriptFormat.CHUNKS, - chunk_size_seconds=chunk_size_seconds, - language=language.split(",") if language else ["en"], - translation=translation or None, - ) - - transcripts = loader.load() - if transcript_format == TranscriptFormat.TEXT and len(transcripts) > 0: - return Data(data={"transcript": transcripts[0].page_content}) - return [Data(data={"content": doc.page_content, "metadata": doc.metadata}) for doc in transcripts]