🗨️ ⚙️ comment-pipeline

Run workflows with comment commands

Getting Started

Check out the Example Workflow.

on: issue_comment

name: Example

jobs:
  parse-comment:
    runs-on: ubuntu-latest
    if: ${{ github.event.issue.pull_request }}
    outputs:
      command: ${{ steps.parse-comment.outputs.command }}
    steps:
      - uses: drewwyatt/comment-pipeline@v1
        id: parse-comment
        with:
          comment: ${{ toJSON(github.event.comment) }}
          commands: |
            /format
            /rebase
          github-token: ${{ secrets.GITHUB_TOKEN }}

  format:
    runs-on: ubuntu-latest
    needs: [parse-comment]
    if: ${{ needs.parse-comment.outputs.command == '/format' }}
    steps:
      - run: echo 'formatting...'

  rebase:
    runs-on: ubuntu-latest
    needs: [parse-comment]
    if: ${{ needs.read-comment.outputs.command == '/rebase' }}
    steps:
      - run: echo 'rebasing...'

Inputs

commands

Newline delimited list of commands that should trigger workflows.

e.g.

# just one
commands: build

# more than one
# (the slashes are not required that is totally up to your preference)
commands: |
  /format
  /rebase

⚠️ NOTE: comments must start with commands in order to trigger a workflow. SO:

If you passed /format and /rebase like above, the following comments would all trigger this action:

• "/format"
• "/format this please"
• "/rebase 🔥"

these would ✖️ not ✖️

• "rebase"
• "please /format this PR"

comment

The comment object (usually github.event.comment)

you probably always want to do exactly what you see in the example workflow.

comment: ${{ toJSON(github.event.comment) }}

github-token

Almost always should be ${{ secrets.GITHUB_TOKEN }}

reaction (optional - defaults to ":+1:")

Emoji reaction used to acknowledge comment.

Possible options:

• +1
• -1
• laugh
• confused
• heart
• hooray
• rocket
• eyes

Outputs

command (can be empty)

The matched command (from inputs.commands) - this will be an empty if there is no match.

This is most commonly used to determine if a job or step should run. e.g.

In the example workflow, if one left a comment "/format this please", outputs.command would be "/format" (because that is the exact command input match).

jobs:
  parse-comment:
    runs-on: ubuntu-latest
    if: ${{ github.event.issue.pull_request }}
    outputs:
      command: ${{ steps.parse-comment.outputs.command }}
    steps:
      - uses: actions/checkout@v2
      - uses: ./
        id: parse-comment
        with:
          comment: ${{ toJSON(github.event.comment) }}
          commands: |
            /format
            /rebase
          github-token: ${{ secrets.GITHUB_TOKEN }}

  format:
    runs-on: ubuntu-latest
    needs: [parse-comment]
    if: ${{ needs.parse-comment.outputs.command == '/format' }} # <------ SEE HERE
    steps:
      - run: echo 'formatting...'

arguments (can be empty)

This may be useful if you would like to specify additional information other than the command. You can see some examples of args in the tests, like:

Given the comment "/format this please" (using previous config examples above), outputs.arguments would be "this please".

Given a config where inputs.commands includes deploy, and a comment "deploy:stage", outputs.arguments would be ":stage".

invocation (can be empty)

The entire comment body (if a command match was found).