Autobids Portal

This is a webapp for interacting with the autobids system.

Setup

Non-python requirements

To use all functionality, you will need:

• Apptainer, with the following images:
  • docker://khanlab/tar2bids
  • docker://khanlab/cfmm2tar
  • docker://khanlab/gradcorrect (along with a private coefficients file)
• A mail server compatible with Flask-Mail.
• A redis server.
• A PostgreSQL server.
• A git repository containing custom heuristics.

Installation instructions

1. Set up a virtual environment.
2. pip install -e ., or poetry install if your environment has poetry.
3. Adjust your configuration by making a copy of .envtemplate and changing any relevant variables.
4. Export all necessary environment variables (set -a && . .env && set +a does the job if you keep them in a .env file).
5. Set up the db: flask db upgrade.

Running the application

Local/Development

The easiest way to get a local development system running is to use docker compose.

1. docker-compose up --build
2. The first time and after any DB model changes, run your migrations by sshing into the autobidsportal container (docker -it exec {container_id} /bin/bash) and running flask db upgrade.
3. The ssh setup for archiving between containers in docker-compose can be a little finnicky. If those steps fail, this is a first place to look. Usually sshing into the relevant containers and running ssh-keyscan and friends can help resolve this.

Production

• A production server (i.e. not the Flask development server invoked with flask run) should be used to serve the autobids portal.
• A service manager should also be used to manage the rq workers that execute asynchronous tasks.
• The operational CLI commands (i.e. flask check_pis, flask run-all-cfmm2tar, etc.) should be run on a regular basis: See crontab.example for an example of how this can be configured.