apunkt/SurfSense
1
0
Fork 0
mirror of https://github.com/MODSetter/SurfSense.git synced 2026-04-26 09:16:22 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge pull request #387 from nabthebest135/local-stt

Browse source 
local STT implementation with Faster-Whisper
This commit is contained in:
Rohan Verma 2025-10-15 14:08:09 -07:00 • committed by GitHub
commit 5ed9aa2b0b
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
6 changed files with 2721 additions and 2511 deletions
Download patch file Download diff file Expand all files Collapse all files

69
surfsense_backend/app/routes/documents_routes.py
View file

@ -784,25 +784,59 @@ async def process_file_in_background(
{"file_type": "audio", "processing_stage": "starting_transcription"},
)
# Open the audio file for transcription
with open(file_path, "rb") as audio_file:
# Determine STT service type
stt_service_type = "local" if app_config.STT_SERVICE and app_config.STT_SERVICE.startswith("local/") else "external"
# Check if using local STT service
if stt_service_type == "local":
# Use local Faster-Whisper for transcription
from app.services.stt_service import stt_service
try:
result = stt_service.transcribe_file(file_path)
transcribed_text = result.get("text", "")
if not transcribed_text:
raise ValueError("Transcription returned empty text")
# Add metadata about the transcription
transcribed_text = (
f"# Transcription of {filename}\n\n{transcribed_text}"
)
except Exception as e:
raise HTTPException(
status_code=422,
detail=f"Failed to transcribe audio file {filename}: {str(e)}"
) from e
await task_logger.log_task_progress(
log_entry,
f"Local STT transcription completed: {filename}",
{
"processing_stage": "local_transcription_complete",
"language": result.get("language"),
"confidence": result.get("language_probability"),
"duration": result.get("duration"),
},
)
else:
# Use LiteLLM for audio transcription
if app_config.STT_SERVICE_API_BASE:
transcription_response = await atranscription(
model=app_config.STT_SERVICE,
file=audio_file,
api_base=app_config.STT_SERVICE_API_BASE,
api_key=app_config.STT_SERVICE_API_KEY,
)
else:
transcription_response = await atranscription(
model=app_config.STT_SERVICE,
api_key=app_config.STT_SERVICE_API_KEY,
file=audio_file,
)
with open(file_path, "rb") as audio_file:
transcription_kwargs = {
"model": app_config.STT_SERVICE,
"file": audio_file,
"api_key": app_config.STT_SERVICE_API_KEY,
}
if app_config.STT_SERVICE_API_BASE:
transcription_kwargs["api_base"] = app_config.STT_SERVICE_API_BASE
transcription_response = await atranscription(**transcription_kwargs)
# Extract the transcribed text
transcribed_text = transcription_response.get("text", "")
# Extract the transcribed text
transcribed_text = transcription_response.get("text", "")
if not transcribed_text:
raise ValueError("Transcription returned empty text")
# Add metadata about the transcription
transcribed_text = (
@ -839,6 +873,7 @@ async def process_file_in_background(
"content_hash": result.content_hash,
"file_type": "audio",
"transcript_length": len(transcribed_text),
"stt_service": stt_service_type,
},
)
else: