Webinterface for audio annotation
- Install django in the environment
- Run install.sh (be careful, it deletes the current database)
- Go into microfaune_annotator:
cd microfaune_annotator
- Run django dev server:
python manage.py runserver
The script install.sh deletes the current project database , recreates it and populates it with toy data for development purposes.
2 users are created (with the same password microfaune):
- admin
- annotator
The project has 3 custom models (cf annotate/models.py):
- Project: define an annotation project which contains a list of possible labels.
- AudioTrack: an object representing an audio track. It links an audio file, points to a Project and can have a prediction array.
- Annotation: an object representing an audio annotation. It points to an AudioTrack, to an user and can be marked as reviewed by another user.
The website has a simple architecture:
- The homepage lists all active projects and links to their pages
- A project page lists all the associated audio tracks and links to a page to annotate them
- The annotation page allows to view, create annotation and save annotation (TODO).