How devrathiyer/ballotproof works
Project type and purpose
• BallotProof is a full‑stack web application built on Django. It delivers a privacy‑preserving ballot validation service by shifting all image analysis into the user’s browser (via OpenCV.js) while the backend handles template provisioning, address lookups, and storage of annotation metadata.
File structure
• Top level
– manage.py / generate.py: Django CLI and sample‐ballot generator (PIL).
– Pipfile / Procfile: Python deps and Heroku startup.
– annotation/, primitives/, screenshots/: data assets for templates, primitives, and examples.
• ballotproof/ (Django project)
– settings.py, wsgi.py, urls.py: project config and top‐level routing.
• checker/ (Django “checker” app)
– static/checker/js (OpenCV.js), css, img: front‑end assets.
– templates/checker: HTML pages (index, app, no_annotation).
– views.py, forms.py, urls.py: endpoint logic for address form, template selection.
• annotator/ (Django “annotator” app)
– models.py, admin.py, migrations: ballot template metadata.
– views.py, urls.py: admin/external endpoints for uploading/managing annotation masks.
Extract architecture details from README
• Client‑side image analysis only (OpenCV.js).
• Address → Geocodio API → district → select template.
• Ballot templates/masks hosted on AWS S3.
• Django backend on Heroku; Postgres DB for template metadata.