Usage of annotation tool requires basic knowledge of Django.
To initialize the project create an environment, install requirements.txt, execute python manage.py makemigrations, and then python manage.py migrate. This will create sqlite3 database for your annotator. In case you want to use another database type, edit settings according to django docs.
To use the tool you need to create superuser and add possible labels for texts with django admin.
You can upload data for annotation via django shell or using django migration (the proper way). In any case consider using bulk_create for speed. To download annotated data use python manage.py dumpdata mydata.json.
The main object - article for annotation. It is required to have html_id, text or readability html - ra_summary, a title - ra_title, and be associated with particular annotator from django.contrib.auth.models.User - bee.
Types for article. This model has one-to-one relations with Article and many-to-many relations with FakeType (not the best model name). All annotated Articles are stored in Classified
Model for types of manipulation or whatever else your annotation is about. Each one has label and long description
Comments for articles. Each feedback has one-to-one relation with Article.