browser audio detection phase 4

This commit is contained in:
Joren de Graaf
2026-10-02 21:45:24 +02:00
parent 1de4b8a0b3
commit 46ad514a70
20 changed files with 903 additions and 31 deletions
+16
View File
@@ -105,3 +105,19 @@ python manage.py compilemessages
```
The small amount of JavaScript UI text is translated by Django in `templates/js_translations.html` and read from `window.BOOMCAST_I18N`. After changing JavaScript, run `npm run build` and `python manage.py collectstatic --noinput` for production. Classifier class names, API keys, sensor tokens, H3 values and stored source/status values remain language-neutral.
## Browser Processing Beta
The administrator setting `browser_processing_beta_enabled` is off by default.
When enabled, the upload form shows **Process recording in this browser (Beta)**,
which also starts off. With the user toggle off, ordinary audio uses the
existing server upload and classifier. With it on, selected audio is checked
locally and selected video is converted to 16 kHz mono WAV before the local
check. Files the user submits still run through the server classifier.
Switching back to server mode cancels browser work and clears the local gate.
The server upload endpoint currently accepts audio, so a video selected in
server mode requires an audio file or browser mode.
Browser assets are loaded only after the user enables the toggle and selects a
file. Regenerate the self-hosted model using [the YAMNet build guide](scripts/BROWSER_YAMNET.md),
then run `npm run build` and `python manage.py collectstatic --noinput`.