diff --git a/python/voice_to_notes/ipc/handlers.py b/python/voice_to_notes/ipc/handlers.py index 35a98b5..c87f665 100644 --- a/python/voice_to_notes/ipc/handlers.py +++ b/python/voice_to_notes/ipc/handlers.py @@ -90,15 +90,40 @@ def make_diarize_handler() -> HandlerFunc: def make_diarize_download_handler() -> HandlerFunc: """Create a handler that downloads/validates the diarization model.""" + import os def handler(msg: IPCMessage) -> IPCMessage: payload = msg.payload hf_token = payload.get("hf_token") try: + import huggingface_hub + + # Disable pyannote telemetry (has a bug in v4.0.4) + os.environ.setdefault("PYANNOTE_METRICS_ENABLED", "false") from pyannote.audio import Pipeline - print("[sidecar] Downloading diarization model...", file=sys.stderr, flush=True) + # Persist token globally so ALL huggingface_hub downloads use auth. + # Setting env var alone isn't enough — pyannote's internal sub-downloads + # (e.g. PLDA.from_pretrained) don't forward the token= parameter. + # login() writes the token to ~/.cache/huggingface/token which + # huggingface_hub reads automatically for all downloads. + if hf_token: + os.environ["HF_TOKEN"] = hf_token + huggingface_hub.login(token=hf_token, add_to_git_credential=False) + + # Pre-download sub-models that pyannote loads internally. + # This ensures they're cached before Pipeline.from_pretrained + # tries to load them (where token forwarding can fail). + sub_models = [ + "pyannote/segmentation-3.0", + "pyannote/speaker-diarization-community-1", + ] + for model_id in sub_models: + print(f"[sidecar] Pre-downloading {model_id}...", file=sys.stderr, flush=True) + huggingface_hub.snapshot_download(model_id, token=hf_token) + + print("[sidecar] Downloading diarization pipeline...", file=sys.stderr, flush=True) pipeline = Pipeline.from_pretrained( "pyannote/speaker-diarization-3.1", token=hf_token, @@ -111,26 +136,23 @@ def make_diarize_download_handler() -> HandlerFunc: ) except Exception as e: error_msg = str(e) + print(f"[sidecar] Model download error: {error_msg}", file=sys.stderr, flush=True) # Make common errors more user-friendly - if "403" in error_msg and "gated" in error_msg.lower(): - # Extract which model needs access - if "segmentation" in error_msg: + if "403" in error_msg or "gated" in error_msg.lower(): + # Try to extract the specific model name from the error + import re + model_match = re.search(r"pyannote/[\w-]+", error_msg) + if model_match: + model_name = model_match.group(0) error_msg = ( - "Access denied for pyannote/segmentation-3.0. " - "Please visit huggingface.co/pyannote/segmentation-3.0 " - "and accept the license agreement." - ) - elif "speaker-diarization" in error_msg: - error_msg = ( - "Access denied for pyannote/speaker-diarization-3.1. " - "Please visit huggingface.co/pyannote/speaker-diarization-3.1 " - "and accept the license agreement." + f"Access denied for {model_name}. " + f"Please visit huggingface.co/{model_name} " + f"and accept the license agreement, then try again." ) else: error_msg = ( - "Access denied. Please accept the license agreements at: " - "huggingface.co/pyannote/speaker-diarization-3.1 and " - "huggingface.co/pyannote/segmentation-3.0" + "Access denied. Please accept the license agreements for all " + "required pyannote models on HuggingFace." ) elif "401" in error_msg: error_msg = "Invalid token. Please check your HuggingFace token." diff --git a/python/voice_to_notes/services/diarize.py b/python/voice_to_notes/services/diarize.py index c29fdd0..692b32f 100644 --- a/python/voice_to_notes/services/diarize.py +++ b/python/voice_to_notes/services/diarize.py @@ -2,15 +2,67 @@ from __future__ import annotations +import os +import subprocess import sys +import tempfile import time from dataclasses import dataclass, field +from pathlib import Path from typing import Any +# Disable pyannote telemetry — it has a bug in v4.0.4 where +# np.isfinite(None) crashes when max_speakers is not set. +os.environ.setdefault("PYANNOTE_METRICS_ENABLED", "false") + from voice_to_notes.ipc.messages import progress_message from voice_to_notes.ipc.protocol import write_message +def _ensure_wav(file_path: str) -> tuple[str, str | None]: + """Convert audio to 16kHz mono WAV if needed. + + pyannote.audio v4.0.4 has a bug where its AudioDecoder returns + duration=None for some formats (FLAC, etc.), causing crashes. + Converting to WAV ensures the duration header is always present. + + Returns: + (path_to_use, temp_path_or_None) + If conversion was needed, temp_path is the WAV file to clean up. + """ + ext = Path(file_path).suffix.lower() + if ext == ".wav": + return file_path, None + + tmp = tempfile.NamedTemporaryFile(suffix=".wav", delete=False) + tmp.close() + try: + subprocess.run( + [ + "ffmpeg", "-y", "-i", file_path, + "-ar", "16000", "-ac", "1", "-c:a", "pcm_s16le", + tmp.name, + ], + check=True, + capture_output=True, + ) + print( + f"[sidecar] Converted {ext} to WAV for diarization", + file=sys.stderr, + flush=True, + ) + return tmp.name, tmp.name + except (subprocess.CalledProcessError, FileNotFoundError) as e: + # ffmpeg not available or failed — try original file and hope for the best + print( + f"[sidecar] WAV conversion failed ({e}), using original file", + file=sys.stderr, + flush=True, + ) + os.unlink(tmp.name) + return file_path, None + + @dataclass class SpeakerSegment: """A time span assigned to a speaker.""" @@ -40,14 +92,19 @@ class DiarizeService: if self._pipeline is not None: return self._pipeline - import os - print("[sidecar] Loading pyannote diarization pipeline...", file=sys.stderr, flush=True) # Use token from argument, fall back to environment variable if not hf_token: hf_token = os.environ.get("HF_TOKEN") or os.environ.get("HUGGING_FACE_HUB_TOKEN") or None + # Persist token globally so ALL huggingface_hub sub-downloads use auth. + # Pyannote has internal dependencies that don't forward the token= param. + if hf_token: + os.environ["HF_TOKEN"] = hf_token + import huggingface_hub + huggingface_hub.login(token=hf_token, add_to_git_credential=False) + models = [ "pyannote/speaker-diarization-3.1", "pyannote/speaker-diarization", @@ -118,8 +175,27 @@ class DiarizeService: if max_speakers is not None: kwargs["max_speakers"] = max_speakers + # Convert to WAV to work around pyannote v4.0.4 duration bug + audio_path, temp_wav = _ensure_wav(file_path) + + print( + f"[sidecar] Running diarization on {audio_path} with kwargs: {kwargs}", + file=sys.stderr, + flush=True, + ) + # Run diarization - diarization = pipeline(file_path, **kwargs) + try: + raw_result = pipeline(audio_path, **kwargs) + finally: + if temp_wav: + os.unlink(temp_wav) + + # pyannote 4.0+ returns DiarizeOutput; older versions return Annotation directly + if hasattr(raw_result, "speaker_diarization"): + diarization = raw_result.speaker_diarization + else: + diarization = raw_result # Convert pyannote output to our format result = DiarizationResult() diff --git a/python/voice_to_notes/services/pipeline.py b/python/voice_to_notes/services/pipeline.py index 7ca728e..b0e148b 100644 --- a/python/voice_to_notes/services/pipeline.py +++ b/python/voice_to_notes/services/pipeline.py @@ -127,15 +127,17 @@ class PipelineService: hf_token=hf_token, ) except Exception as e: + import traceback print( f"[sidecar] Diarization failed, falling back to transcription-only: {e}", file=sys.stderr, flush=True, ) + traceback.print_exc(file=sys.stderr) write_message( progress_message( request_id, 80, "pipeline", - "Diarization unavailable, using transcription only..." + f"Diarization failed ({e}), using transcription only..." ) ) diff --git a/src/lib/components/SettingsModal.svelte b/src/lib/components/SettingsModal.svelte index 85ed628..0b95d09 100644 --- a/src/lib/components/SettingsModal.svelte +++ b/src/lib/components/SettingsModal.svelte @@ -118,12 +118,14 @@

Speaker detection uses pyannote.audio models hosted on HuggingFace. You must accept the license for each model:

  1. Create a free account at openUrl('https://huggingface.co/join')}>huggingface.co
    2. -
  2. Accept the license on each of these pages: +
  3. Accept the license on all three of these pages:
    4. diff --git a/src/lib/components/SpeakerManager.svelte b/src/lib/components/SpeakerManager.svelte index a3c773a..e6e92f6 100644 --- a/src/lib/components/SpeakerManager.svelte +++ b/src/lib/components/SpeakerManager.svelte @@ -1,5 +1,6 @@