Description
Thanks for the amazing tool and please forgive me if this has already been addressed and I missed it in my search. My understanding is that OpenAI Whisper and therefore this tool should be able to ingest video files directly. However, I can select only Audio files to ingest.
Describe the bug
Standalone version 0.19.5 on Windows 11 ingests only audio files, not video
To Reproduce
Steps to reproduce the behavior:
- Go to 'File'
- Click on 'Ingest files'
- File selector is fixed to "Audio files" and can't be changed to video or all files
Expected behavior
I expected to be able to choose video files.
Log file
2024-03-01 18:40:46,293 - StAI - DEBUG:
--------------
Platform: Windows 10
Platform version: 10.0.22631
OS: 10 10.0.22631 SP0 Multiprocessor Free ('', '', '')
running Python 3.10.11.final.0
-------------- (__main__.py:15)
2024-03-01 18:40:47,155 - StAI - DEBUG: All package requirements met. (__main__.py:59)
2024-03-01 18:40:47,192 - StAI - DEBUG: FFmpeg not found in current working directory. (storytoolkitai.py:671)
2024-03-01 18:40:47,192 - StAI - DEBUG: Looking for FFmpeg in env variable. (storytoolkitai.py:675)
2024-03-01 18:40:47,192 - StAI - DEBUG: FFMPEG_BINARY env variable is empty. Looking for FFmpeg in PATH. (storytoolkitai.py:682)
2024-03-01 18:40:47,192 - StAI - DEBUG: Checking ffmpeg binary: .\ffmpeg.EXE (storytoolkitai.py:694)
2024-03-01 18:40:47,214 - StAI - DEBUG: FFmpeg exit code: 1 (storytoolkitai.py:699)
2024-03-01 18:40:47,214 - StAI - DEBUG: FFmpeg found at .\ffmpeg.EXE (storytoolkitai.py:702)
2024-03-01 18:40:50,471 - StAI - DEBUG: Loading config file C:\Users\***\StoryToolkitAI\config.json. (storytoolkitai.py:319)
2024-03-01 18:40:50,639 - StAI - INFO: �[1m�[4mRunning StoryToolkitAI version 0.19.5 (standalone) (storytoolkitai.py:95)
2024-03-01 18:40:50,639 - StAI - DEBUG: Using cuda for Torch. (toolkit_ops.py:729)
2024-03-01 18:40:50,639 - StAI - DEBUG: Resolve API disabled via config. (toolkit_ops.py:192)
2024-03-01 18:40:50,640 - StAI - DEBUG: No items left in the queue. Try to ping the queue again later. (processing_queue.py:824)
2024-03-01 18:40:50,752 - StAI - DEBUG: Running with TK 8.6.12 (toolkit_ui.py:1057)
2024-03-01 18:40:50,863 - StAI - WARNING: A new standalone version of StoryToolkitAI is available. (toolkit_ui.py:1335)
2024-03-01 18:40:51,559 - StAI - INFO: Starting StoryToolkitAI GUI (toolkit_ui.py:2290)
2024-03-01 18:40:52,485 - StAI - DEBUG: Closing window: update_available (toolkit_ui.py:1745)
There's nothing else in the log because I can't add a video file.
Screenshots
System (please complete the following information):
- OS: Windows 11
- Python versions installed on machine: 3.10.11 (3.12 is also on the machine but not used by StAI)
- StoryToolkitAI Version 0.19.5
Additional context
I've been manually using ffmpeg to generate an audio file using this command I found on the Whisper discussions:
ffmpeg -i <recording> -ar 16000 -ac 1 -c:a pcm_s16le <output>.wav
It works fine, but I'd rather have a 1-step process than to extract audio for each project. I also tried to get around the file picker by putting my video in its own directory and then attempting to ingest the directory. The log file references itself but doesn't contain any helpful information:
2024-03-01 18:32:45,863 - StAI - DEBUG: Added queue id 1709335965.8637035 to queue history (processing_queue.py:63)
2024-03-01 18:32:56,959 - StAI - WARNING: Cannot start ingest. See log for more details. (toolkit_ui.py:12510)
2024-03-01 18:34:07,088 - StAI - DEBUG: Closing window: ingest-1709335965.8637035 (toolkit_ui.py:1745)
2024-03-01 18:39:10,788 - StAI - DEBUG:
Is there a way to directly ingest video files, or is this expected behavior? Thank you!