assignbot 2.2.2

The preferences file

The users define their “review preferences” in a preferences file, formatted in yaml.

The “preferences file” is a yaml file constructed as follow :

username_1:
    max_auto_review_per_week: XXX
    max_auto_review_per_day: YYY
username_2:
    max_auto_review_per_week: ZZZ
    max_auto_review_per_day: WWW

where username_1 is the username of a gitlab user, and max_auto_review_per_week (max_auto_review_per_day) is the maximum of automatically assigned review per week (per day).

The S3 storage

Assignbot uses a CSV file to keep track of users assignments. This CSV file is stored in an S3 storage. When you execute the bot, you must provide the following environment variables:

• AWS_ACCESS_KEY_ID: your S3 key ID

• AWS_SECRET_ACCESS_KEY: your S3 secret access key

• S3_ENDPOINT_URL: your S3 endpoint url

• S3_BUCKET_NAME: the S3 bucket to be used

Executing the bot

Once assignbot installed, you can run it as follow :

$ GITLAB_URL="https://your.forge.org" \
  GITLAB_TOKEN="XXX-an-api-token" \
  S3_ENDPOINT_URL="https://your.s3.storage.fr" \
  S3_BUCKET_NAME="you_bucket_name" \
  AWS_SECRET_ACCESS_KEY="your_secret_access_key" \
  AWS_ACCESS_KEY_ID="your_access_key_id" \
  python3 -m assignbot ./users_preferences.yml

and the bot should start assigning merge requests to the users.

Use case example

To use this bot, you can add a new repository on your forge with the following .gitlab-ci.yml :

assign:
  stage: assign
  only:
    - schedules
  variables:
      # there is a gitlab bug in the validation of AWS variables. We work
      # around it while waiting for the correction.
      # see: https://gitlab.com/gitlab-org/gitlab/-/issues/215927
      AWS_ACCESS_KEY_ID: "$_AWS_ACCESS_KEY_ID"
      AWS_SECRET_ACCESS_KEY: "$_AWS_SECRET_ACCESS_KEY"
  script:
    - pip install assignbot
    - python -m assignbot ./users_preferences.yml

This job assumes that you have defined the appropriate environment variables, and that you have a users_preferences.yml file at the root of this repository.

Then, you can create a new “schedule job” in gitlab, to call this job periodically.