luciano/markitdown
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fallback to _transcribe_audio

Browse source
This commit is contained in:
Ji Zhang 2025-02-11 17:30:39 -08:00
parent 8301427ab5
commit b8927e5e65
1 changed files with 15 additions and 7 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
packages/markitdown/src/markitdown/converters/_wav_converter.py
View file

@ -1,7 +1,10 @@
import logging
from typing import Union from typing import Union
from ._base import DocumentConverter, DocumentConverterResult from ._base import DocumentConverter, DocumentConverterResult
from ._media_converter import MediaConverter from ._media_converter import MediaConverter
logger = logging.getLogger(__name__)
# Optional Transcription support # Optional Transcription support
IS_AUDIO_TRANSCRIPTION_CAPABLE = False IS_AUDIO_TRANSCRIPTION_CAPABLE = False
IS_WHISPER_CAPABLE = False IS_WHISPER_CAPABLE = False
@ -82,13 +85,18 @@ class WavConverter(MediaConverter):
) )
def _transcribe_with_whisper(self, local_path: str, client) -> str: def _transcribe_with_whisper(self, local_path: str, client) -> str:
"""Transcribe audio using OpenAI's Whisper model.""" """Transcribe audio using OpenAI's Whisper model, falling back to speech_recognition if it fails."""
try:
with open(local_path, "rb") as audio_file: with open(local_path, "rb") as audio_file:
transcription = client.audio.transcriptions.create( transcription = client.audio.transcriptions.create(
model="whisper-1", model="whisper-1",
file=audio_file file=audio_file
) )
return transcription.text.strip() return transcription.text.strip()
except Exception as e:
logger.warning(f"Whisper transcription attempt failed: {str(e)}")
logger.info("Falling back to speech_recognition...")
return self._transcribe_audio(local_path)
def _transcribe_audio(self, local_path) -> str: def _transcribe_audio(self, local_path) -> str:
recognizer = sr.Recognizer() recognizer = sr.Recognizer()