rating-b

Run locally

Copy .env.example to .env and update it with your local Postgres details:

cp .env.example .env

DJANGO_SECRET_KEY needs to be present but can have any value.

Create a virtual environment, install dependencies:

python -m venv ./venv
source ./venv/bin/activate
pip install -r requirements.txt

Commands

• python manage.py import_release N reads teams and players ratings for release_id N from API (e.g. 1443 stands for 2020-04-03 release) and dumps it to b.player_rating, b.team_rating and b.player_rating_by_tournament (top bonuses for each player).
• python manage.py calc_all_releases [--first_to_calc=2021-09-09] calculates all releases from first_to_calc until today.
• python manage.py calc_release YYYY-MM-DD reads previous release data from our DB (it must already exist) and creates new release for YYYY-MM-DD (this date must be Thursday for 2021+ and Friday for 2020-).

Project structure

The top directories are:

• dj -- core Django files.
• b -- Django files specific for rating B model.
  • models.py contains the descriptions of all tables in b schema;
  • management/commands/ contains commands that run rating-related computations.
• scripts -- functions that actually read the data from DB, compute ratings, and flush the results to the DB, and tests for them.

Deployment

rating-b runs on Fly. Deployment is defined in fly.toml and has one process: supercronic /app/crontab.

Supercronic is recommended by Fly for cron jobs: it loads environment variables, forwards job logs to stdout and stderr, and handles sigterm signals. We install it in the final section of Dockerfile.

Cron jobs are set up in crontab. For now, we recalculate all releases every three hours: this job takes about 30 minutes (as of April 2023), so there’s not much sense in making things more complicated.