Repo Automation Bots

A collection of bots, based on probot, for performing common maintenance tasks across the open-source repos managed by Google on GitHub.

Implemented Bots

Name	Description	Install
auto-approve	Automatically approves and merges PRs matching user-specified configs	install
auto-label	Automatically labels issues and PRs with product, language, or directory based labels	install
blunderbuss	Assigns issues and PRs randomly to a specific list of users	install
cherry-pick-bot	Cherry-pick merged PRs between branches	install
conventional-commit-lint	PR checker that ensures that the commit messages follow conventionalcommits.org style	install
do-not-merge	PR checker that ensures the `do not merge` label is not present	install
failurechecker	Check for automation tasks, e.g., releases, that are in a failed state	install
flakybot	Listen on PubSub queue for broken builds, and open corresponding issues	install
generated-files-bot	PR checker to notify if you are modifying generated files	install
label-sync	Synchronize labels across organizations	install
license-header-lint	PR checker that ensures that source files contain valid license headers	install
merge-on-green	Merge a pull-request when all required checks have passed	install
policy	Check repo configuration against known rules	install
release-please	Proposes releases based on semantic version commits	install
release-trigger	Trigger releases jobs	install
repo-metadata-lint	Lint .repo-metadata.json files	install
snippet-bot	Check for mismatched region tags in PRs	install
sync-repo-settings	Synchronize repository settings from a centralized config	install
trusted-contribution	Allows Kokoro CI to trigger for trusted contributors	install

Development environment

You need to install node.js version 12 or higher.

To manage multiple Node.js versions, you can use nvm.

Running the app locally

Create a Proxy to Relay Webhooks

In order to forward to your local machine, you can use smee.io. Visit https://smee.io/new and create a proxy for relaying webhooks to your local web-service. After creating the proxy, you'll get the URL of the new proxy.

In the root directory of repo-automation-bots, run:

npm run proxy -- -u <URL-OF-PROXY>

Creating the Development Application

If it's your first time running your application, you should create a new GitHub application using the probot server:

cd packages/your-bot.
npm start.
visit: http://localhost:3000 and install.

Granting the Development Application permissions and events

By default there will be no permissions. Visit https://github.com/settings/installations, click configure, then 'app settings'.
Navigate to Permissions and Events. You likely need 'Repository > Pull Requests' for permissions.
You also will need to subscribe to events (bottom of page). For instance, if your bot responds to PR activity, the 'Events > Pull Request' should be enabled.

Install the bot on a repo

Follow the link to install the app and navigate to 'Install App', if installed on the organization you desire (likely yourself for testing), click the gear.
Under permissions ensure that there aren't pending requests to be approved
Under repository access select only select repositories. Select the repository you wish to test against.

Running Your Application

Once you've created your application, and installed it on some of your repos, start probot again, setting the following environment variables. Most can be found at github.com/settings/apps/{YOUR_APP}:

APP_ID: the ID, listed near the top, App ID: 12345
PRIVATE_KEY_PATH: path to App's private key, you can request a new one be created and downloaded at the bottom of the page.
- Alternatively, set the GitHub client ID and secret:
  - GITHUB_CLIENT_ID: client ID from the top of the page.
  - GITHUB_CLIENT_SECRET: client secret from the top of the page.
WEBHOOK_SECRET: secret key set in GitHub developer settings. Edit this to a known value in the settings page.

Environment variables set, run:

cd packages/your-bot.
npm start.

Running bots on a Cron

To run a bot on a schedule include a file in your bot's folder named cron whose content is valid unix -cron format. This will create a Cloud Scheduler Job which makes requests to your endpoint at the specified schedule.