This is the proof of concept project for Vy.Vision RoarScore. This web application demonstrates an audience video with a heatmap overlay, a context video (if available), a score-over-time graph synchronized with the video, a radar chart of 10 emotions, and an example of a fan view for video scoreboards. Videos are selected using the Scene dropdown, and scoring profiles can be selected via the Profile dropdown.
Autenticated users can edit scenes and profiles. Authentication is username/password based, and user creation is handled in the Firebase console.
Data is stored in the same Firebase Firestore instance as roarscore-web but utilizes separate collections.
While this repo does include backend service processing code, it is not utilized and has been superseeded by the roarscore-service project. When an audience video is uploaded to a scene the roarscore-service project processes it like it would any other job.
- Firebase
- Auth (username/password)
- Firebase
- Storage
- Hosting
- Cloudflared
Local development uses an Express server to emulate firebase functions. The
development instance on magicmissile is using cloudflared to map to
poc.roarscore.ai.
Deployment to production hosting is superseeded by the roarscore-web project,
do not deploy to production without reviewing the firebase settings.