Buzz Live Recording does not work, but no information in logs

[robsiera]

Buzz Live Recording does not work, but no information in logs

[raivisdejus]

Ok, learned this about the GitHub. If you are signed into the GitHub, you can download build files from workflows.

[raivisdejus]

@robsiera Thanks for reporting this issue. It occurs when the model application is trying to use does not support the live transcription. Reading from previous settings is involved, so a workaround could be to change the model in the model selector or reinstalling the application and also removing it's settings when you uninstall. On Linux the settings will be in ~/.local/share/Buzz/

[robsiera]

Alas, your suggestions don't help.

a workaround could be to change the model in the model selector

In which dialog do you mean? Clicking on the microphone icon does not even bring up a dialog in my case.

reinstall the application and also removing it's settings when you uninstall

I tried that too. Same result.

Again, don't assume it is your application that is in the wrong, but my problem is, that the logs don't give a clue of what issue that Buzz seems to encounter.

BTW 1 - I'm using the Windows version.
BTW 2 - are the settings stored in the sqlite db?

[raivisdejus]

I assume you are referencing the logs from %USERPROFILE%\AppData\Local\Buzz\Buzz\Logs (Opens if you paste this in Windows explorer). Please share the log with us. If it is large you can also delete it and open the app again so that new log file is created and has only info from the last run.

From what I am Googling about Windows settings seem to be stored in Windows registry HKEY_CURRENT_USER\Software\Buzz\. Available via regedit.exe

[robsiera]

I just installed the nightly build, deleted my logfile, and was able to reproduce the issue.
This is what is now in the log:

[2024-06-06 17:15:57,569] settings.__init__:14 DEBUG -> settings filename: \HKEY_CURRENT_USER\Software\Buzz\OrganizationDefaults
[2024-06-06 17:16:21,445] model_loader.<module>:32 ERROR -> 
Traceback (most recent call last):
  File "buzz\model_loader.py", line 28, in <module>
  File "PyInstaller\loader\pyimod02_importers.py", line 419, in exec_module
  File "buzz\whisper_cpp.py", line 863, in <module>
  File "buzz\whisper_cpp.py", line 551, in __call__
ImportError: Could not load whisper.
[2024-06-06 17:16:21,445] model_loader.<module>:38 DEBUG -> Model root directory: C:\Users\MyUser\AppData\Local\Buzz\Buzz\Cache\models
[2024-06-06 17:16:21,496] locale.<module>:12 DEBUG -> UI locales ['en-BE', 'en-Latn-BE', 'en-GB', 'en-Latn-GB']
[2024-06-06 17:16:22,416] utils._find_ffmpeg_extension:114 DEBUG -> Loading FFmpeg6
[2024-06-06 17:16:22,416] utils._find_ffmpeg_extension:120 DEBUG -> Failed to load FFmpeg6 extension.
Traceback (most recent call last):
  File "torio\_extension\utils.py", line 116, in _find_ffmpeg_extension
  File "torio\_extension\utils.py", line 105, in _find_versionsed_ffmpeg_extension
  File "importlib\util.py", line 94, in find_spec
ModuleNotFoundError: No module named 'torio.lib'
[2024-06-06 17:16:22,416] utils._find_ffmpeg_extension:114 DEBUG -> Loading FFmpeg5
[2024-06-06 17:16:22,426] utils._find_ffmpeg_extension:120 DEBUG -> Failed to load FFmpeg5 extension.
Traceback (most recent call last):
  File "torio\_extension\utils.py", line 116, in _find_ffmpeg_extension
  File "torio\_extension\utils.py", line 105, in _find_versionsed_ffmpeg_extension
  File "importlib\util.py", line 94, in find_spec
ModuleNotFoundError: No module named 'torio.lib'
[2024-06-06 17:16:22,426] utils._find_ffmpeg_extension:114 DEBUG -> Loading FFmpeg4
[2024-06-06 17:16:22,426] utils._find_ffmpeg_extension:120 DEBUG -> Failed to load FFmpeg4 extension.
Traceback (most recent call last):
  File "torio\_extension\utils.py", line 116, in _find_ffmpeg_extension
  File "torio\_extension\utils.py", line 105, in _find_versionsed_ffmpeg_extension
  File "importlib\util.py", line 94, in find_spec
ModuleNotFoundError: No module named 'torio.lib'
[2024-06-06 17:16:22,426] utils._find_ffmpeg_extension:114 DEBUG -> Loading FFmpeg
[2024-06-06 17:16:22,426] utils._find_ffmpeg_extension:120 DEBUG -> Failed to load FFmpeg extension.
Traceback (most recent call last):
  File "torio\_extension\utils.py", line 116, in _find_ffmpeg_extension
  File "torio\_extension\utils.py", line 105, in _find_versionsed_ffmpeg_extension
  File "importlib\util.py", line 94, in find_spec
ModuleNotFoundError: No module named 'torio.lib'
[2024-06-06 17:16:22,741] settings.__init__:14 DEBUG -> settings filename: \HKEY_CURRENT_USER\Software\Buzz\OrganizationDefaults
[2024-06-06 17:16:22,745] buzz.main:60 DEBUG -> app_dir: C:\Program Files (x86)\Buzz\_internal
[2024-06-06 17:16:22,745] buzz.main:61 DEBUG -> log_dir: C:\Users\MyUser\AppData\Local\Buzz\Buzz\Logs
[2024-06-06 17:16:22,784] db._setup_db:38 DEBUG -> Database connection opened: C:\Users\MyUser\AppData\Local\Buzz\Buzz\Buzz.sqlite
[2024-06-06 17:16:22,796] settings.__init__:14 DEBUG -> settings filename: \HKEY_CURRENT_USER\Software\Buzz\OrganizationDefaults
[2024-06-06 17:16:22,817] settings.__init__:14 DEBUG -> settings filename: \HKEY_CURRENT_USER\Software\Buzz\OrganizationDefaults
[2024-06-06 17:16:22,828] file_transcriber_queue_worker.run:40 DEBUG -> Waiting for next transcription task
[2024-06-06 17:16:25,554] settings.__init__:14 DEBUG -> settings filename: \HKEY_CURRENT_USER\Software\Buzz\OrganizationDefaults

I checked the content of the folder C:\Users\MyUser\AppData\Local\Buzz\Buzz\Cache\models. That folder is empty.

Transcritions work. But they seem to search for models elsewhere.
Namely here C:\Users\MyUser\.cache\whisper (and here C:\Users\MyUser\.cache\huggingface\hub?).

[2024-06-06 17:24:50,788] file_transcriber_queue_worker.run:56 DEBUG -> Starting next transcription task
[2024-06-06 17:24:50,807] whisper_file_transcriber.transcribe:47 DEBUG -> Starting whisper file transcription, task = FileTranscriptionTask(transcription_options=TranscriptionOptions(language='nl', task=<Task.TRANSCRIBE: 'transcribe'>, model=TranscriptionModel(model_type=<ModelType.WHISPER: 'Whisper'>, whisper_model_size=<WhisperModelSize.MEDIUM: 'medium'>, hugging_face_model_id='DTAI-KULeuven/robbert-2023-dutch-large'), word_level_timings=False, temperature=(0.0, 0.2, 0.4, 0.6, 0.8, 1.0), initial_prompt='', openai_access_token=''), file_transcription_options=FileTranscriptionOptions(file_paths=['C:/Users/MyUser/Documents/Sound Recordings/Working/Done/dries.mp3'], url=None, output_formats={<OutputFormat.TXT: 'txt'>}), model_path='C:\\Users\\MyUser\\.cache\\whisper\\medium.pt', id=85732229, uid=UUID('c7dea163-1871-4686-921a-986edfa58a18'), segments=[], status=None, error=None, queued_at=None, started_at=None, completed_at=None, output_directory=None, source=<Source.FILE_IMPORT: 'file_import'>, file_path='C:/Users/MyUser/Documents/Sound Recordings/Working/Done/dries.mp3', url=None, fraction_downloaded=0.0)

[2024-06-06 17:29:23,671] whisper_file_transcriber.transcribe:70 DEBUG -> whisper process completed with code = 0, time taken = 0:04:32.863902, number of segments = 68
[2024-06-06 17:29:23,672] settings.__init__:14 DEBUG -> settings filename: \HKEY_CURRENT_USER\Software\Buzz\OrganizationDefaults
[2024-06-06 17:29:23,672] file_transcriber.write_output:107 DEBUG -> Writing transcription output, path = C:/Users/MyUser/Documents/Sound Recordings/Working/Done\dries (transcribed oon 06-Jun-2024 17-29-23).txt, output format = OutputFormat.TXT, number of segments = 68
[2024-06-06 17:29:23,673] file_transcriber_queue_worker.run:40 DEBUG -> Waiting for next transcription task
[2024-06-06 17:29:23,673] file_transcriber.write_output:136 DEBUG -> Written transcription output

Finally you might want to know this:

[raivisdejus]

Whisper (regular whisper) models are currently stored (cached) in .cache\whisper other models will be stored in Buzz\Cache\models.

Curious how you installed the "nightly" build? Cloned repo and ran manual build or is there some place where one may download nightly builds?