RTI STAR (Statistical Traffic Analysis Report)

STAR is a Python/Flask-based web application created to help examine traffic stop records for evidence of racial disproportionality and is based on the Veil of Darkness method developed by Grogger & Ridgeway (2006).

Quickstart

You need a default_config.py file under star/. See star/default_config.py-sample for an example.

Docker

Run the following commands to have your environment in a Docker container.

docker-compose build
docker-compose up

The MANAGE_ARG parameter is passed to the python manage.py command (defaults to server)

Or, run the following commands to bootstrap your environment.

pip install -r requirements.txt
python3 ./manage.py server

Shell

To open the interactive shell, run:

python3 ./manage.py shell

By default, you will have access to app.

Running Tests

To run all tests, run:

py.test

Deployment

Note: You need credentials on cds-mallard and to be able to ssh into it. You will also need the cfs-production's user/host and the key saved in ~/.ssh/cfs-production-admin.pem.

To deploy the app, do the following steps:

• ssh into mallard: ssh <USER>@<CDS_MALLARD_HOST>

• sudo su gitlab-runner

• ssh into cfs-production: ssh -i ~/.ssh/cfs-production-admin.pem <CFS_PRODUCTION_USER>@<CFS_PRODUCTION_HOST>

• Go to the rti-star directory: cd rti-star

• Review the current image's tag in the docker-compose.yml file: cat docker-compose.yml

  • Look for the image specified and see which tag rti-star has:

  services:
  app:
      image: rti-star:<TAG_NUMBER>
  ...

• Confirm there is a backup of this image in the backup folder: cd backup && ls

  • There should be a rti-star<TAG_NUMBER>.tar.gz
  • If there is not, then from the rti-star directory run docker save rti-star:<TAG_NUMBER> | gzip > backup/rti-star<TAG_NUMBER>.tar.gz

• From your local environment, build the image if you haven't already: docker-compose build

• give the image the appropriate name and tag: docker tag star_app rti-star:<NEW_TAG_NUMBER>

  • (NEW_TAG_NUMBER = 1 + TAG_NUMBER)

• save the updated docker image: docker save rti-star:<NEW_TAG_NUMBER> | gzip > rti-star<NEW_TAG_NUMBER>.tar.gz

• scp the new image to mallard: scp path/to/rti-star<NEW_TAG_NUMBER>.tar.gz <USER>@<CDS_MALLARD_HOST>:~/

• ssh into mallard: ssh <USER>@<CDS_MALLARD_HOST>

• scp the new image to cfs-production: sudo scp -i ~/.ssh/cfs-production-admin.pem rti-star<NEW_TAG_NUMBER>.tar.gz <CFS_PRODUCTION_USER>@<CFS_PRODUCTION_HOST>:~/rti-star/backup

• sudo su gitlab-runner

• ssh into cfs-production: ssh -i ~/.ssh/cfs-production-admin.pem <CFS_PRODUCTION_USER>@<CFS_PRODUCTION_HOST>

• Go to the rti-star directory: cd rti-star

• Load the new image: docker load --input backup/rti-star<NEW_TAG_NUMBER>.tar.gz

• docker-compose down

• Update the docker-compose.yml file to the new tag: vi docker-compose.yml:

  services:
  app:
      image: rti-star:<NEW_TAG_NUMBER>
  ...

• docker-compose up - d

Licensing

This software is released under the BSD 3-Clause License with the exception of files under the star/static/vendor/ directory. Please read LICENSE.txt for details.

All files under star/static/vendor/ are licensed under their respective license agreements.