Algorithms Database

The Video Query Algorithms project uses PostgresSQL for it database needs. In the Video Query project, the database is seeded and managed through Django Migrations. These migrations are currently a part of the Video Query Api repository.

To set your database up, please follow the instructions located in the API wiki.

Seeding Feature data

In dev, you might find a need to seed Feature data in your database. For example, if you are not ready to create Features from your own video dataset, you can run the following to seed some data for evaluation. Note: This only works in python 3.6 and above because of changes in os.scandir.

python load_db.py ../data/features/stock-video-clips_features/  --base_url "<api url>"

After loading the features, a video record must also be created in the database with the name of the video, e.g. DowntownBrooklynDrive_480p