Dinghy

Dinghy, a GitHub activity digest tool.

Dinghy uses the GitHub GraphQL API to find recent activity on releases, issues and pull requests, and writes a compact HTML digest like this.

Sample Digest

Here's a sample of a Dinghy digest reporting on some PSF repos: black, requests, and PEPs.

Getting Started

1. Install dinghy:

   $ python -m pip install dinghy

2. To run dinghy you will need a GitHub personal access token. The scopes you need to assign to it depend on what repos you'll be accessing. If you are only accessing public repos, then you don't need any scopes. If you will be accessing any private repos, then you need the "repo" scope. Create a token and define the GITHUB_TOKEN environment variable with the value:

   $ export GITHUB_TOKEN=ghp_Y2oxDn9gHJ3W2NcQeyJsrMOez

3. Then run dinghy with a GitHub URL:

   $ dinghy https://github.com/Me/MyProject
   Wrote digest: digest.html

   You will have a digest of the repo's last week of activity in digest.html. It will look something like this.

   You can also write a YAML configuration file to digest multiple sources, or with different time periods:

   $ dinghy my-dinghy-config.yaml
   Wrote digest: proj1.html
   Wrote digest: proj2-daily.html
   Wrote digest: proj2-weekly.html

   Extra arguments specify which digests to write:

   $ dinghy my-dinghy-config.yaml proj1.html
   Wrote digest: proj1.html

Configuration

Dinghy configuration is read from a YAML file (dinghy.yaml by default). Here's an example:

digests:
  - digest: lastweek.html
    title: My projects last week
    since: 1 week
    items:
      - https://github.com/orgs/myorg/projects/17
      - https://github.com/orgs/anotherorg/projects/8
      - https://github.com/myorg/myrepo/pulls

  - digest: hotnews.html
    title: Today's news
    since: 1 day
    items:
      - url: https://github.com/orgs/anotherorg/projects/8
        home_repo: anotherorg/wg
      - https://github.com/myorg/churnchurn/issues

  - digest: all_prs.html
    since: 1 day
    items:
      - search: org:myorg is:pr
        title: MyOrg pull requests

defaults:
  ignore_users:
    - app-user
    - fake-bot

The digests clause is a list of digests to produce. The defaults clause sets defaults for the digest options in the rest of the file. Each digests clause specifies what to digest:

• The digest setting is the HTML digest file to write.
• The since setting indicates how far back to look for activity. It can use units of weeks, days, hours, minutes and seconds, and can also be abbreviated, like 1d6h. Using since: forever will include all activity regardless of when it happened. If since is omitted, it defaults to one week. You can specify --since=<SINCE> on the dinghy command line to provide an explicit value.
• The items setting is a list of things to report on, specified in a few different ways:
  • The url setting is a GitHub URL, in a number of forms:
    • An organization project URL will report on the issues and pull requests in the project. Your GitHub token will need the "read:project" scope.
    • A URL to a repo will report on the issues, pull requests and releases in the repo.
    • A URL to a repo's issues will report on the issues in the repo.
    • A URL to a repo's pull requests will report on the pull requests in the repo.
    • A URL to a repo's releases will report on the releases in the repo.
    • Any of these URLs can point to a GitHub Enterprise installation instead of https://github.com.
  • The search setting can specify a GitHub search query to find issues or pull requests. The query will have an updated: term added to it to account for the since: setting.
  • If an item only needs to specify a GitHub URL, then it can simply be the URL string.
• The optional title setting will be used to construct the title and main header of the HTML page.
• The template setting is the name of a Jinja2 template file to use to produce the digest. It defaults to "digest.html.j2", which is packaged with dinghy. The data passed to the template is under-specified; if you want to write a template of your own, model it on the built-in digest.html.j2.

• For GitHub Enterprise, you can specify api_root, which is the URL to build on for GraphQL API requests. It defaults to "https://api.github.com/graphql".

Items can have additional options:

• By default, no activity is reported for bot users. If you want to include them, use include_bots: true.
• Some applications perform actions using real user accounts, but you'd like to ignore them anyway. You can list those user names that should be ignored in the ignore_users setting.
• Digests can have an explicit title set with the title setting.
• Options for organization projects include:
  • home_repo is the owner/repo of the repo in which most issues will be created. Issues in other repos will have the repo indicated in the digest.

Daily Publishing

The sample digest is published daily using a GitHub Action from its own repo: nedbat/dinghy_sample. You can use it as a starting point for your own publishing.

Contributors

Thanks to all who have helped:

• Ned Batchelder
• Bill Mill
• Doug Hellmann
• Henry Gessau
• Lucas Taylor
• Quentin Pradet
• Simon de Vlieger