⚡️ Comvent

Simple control for GitHub Actions to handle "comment event" - a building block for rich comment handling

| |

🌅 Contents

• Examples
• Action Inputs
• Comvent Configuration File
• Action Outputs

🚀 Examples

Comvent is being used in this repo as well. You can check out this issue and comment to see the actual Comvent offerings in action! 🏃

ChatBot Setup

It is pretty easy to provide simplistic ChatBot experience with PR / Issue comment.

You can add Comvent setup at the beginning of the issue_comment event, and use its output to run any step afterwards based on the patterns found. The below is an example to simply repsond to a comment.

.github/workflows/chatbot.yaml

- name: Handle with Comvent
  uses: rytswd/comvent@v0.3
  id: comvent
  with:
    token: ${{ secrets.GITHUB_TOKEN }}
    config-path: .github/comvent-chatbot.yaml

# The below only runs when Comvent finds matching comment.
# This assumes that Comvent config holding 'command-random' as a keyword.
- if: steps.comvent.outputs.command-random != ''
  name: Handle random event
  uses: actions/github-script@v3
  with:
    github-token: ${{ secrets.GITHUB_TOKEN }}
    # An example of responding to a comment.
    script: |
      const comment = `Command \`/random\` was triggered by the comment`;

      github.issues.createComment({
        issue_number: context.issue.number,
        owner: context.repo.owner,
        repo: context.repo.repo,
        body: comment
      })

#
# You can repeat similar handling based on the Comvent config.
#

Click to view full example

.github/workflows/chatbot.yaml

name: ChatBot
on:
  issue_comment:
    # This shouldn't be called for comment deletion
    types:
      - created
      - edited

jobs:
  chatbot:
    name: With latest code base
    runs-on: ubuntu-latest

    steps:
      - name: Handle with Comvent
        uses: rytswd/comvent@v0.3
        id: comvent
        with:
          token: ${{ secrets.GITHUB_TOKEN }}
          config-path: .github/comvent-chatbot.yaml

      # The below only runs when Comvent finds matching comment.
      # This assumes that Comvent config holding 'command-random' as a keyword.
      - if: steps.comvent.outputs.command-random != ''
        name: Handle random event
        uses: actions/github-script@v3
        with:
          github-token: ${{ secrets.GITHUB_TOKEN }}
          # An example of responding to a comment.
          script: |
            const comment = `Command \`/random\` was triggered by the comment`;

            github.issues.createComment({
              issue_number: context.issue.number,
              owner: context.repo.owner,
              repo: context.repo.repo,
              body: comment
            })

      #
      # You can repeat similar handling based on the Comvent config.
      #

As used above, Comvent configuration needs to define the keywords to find patterns.

.github/comvent-chatbot.yaml

---
version: 0.2.0

trigger: default

keywords:
  - name: command-random
    value: '^\/random$' # Regex which looks for a comment with '/random'
  - name: some-other-command
    value: 'some arbitrary regex setup'

  # You can have as many regex setup as you need.

Handle Abusive Content

You can add a simple automation to monitor abusive comments.

Using github-script, you can automatically remove such comment when Comvent finds one, while leaving a comment about the deletion.

.github/workflows/abuse-monitor.yaml

- name: Handle with Comvent
  uses: rytswd/comvent@v0.3
  id: comvent
  with:
    token: ${{ secrets.GITHUB_TOKEN }}
    config-path: .github/comvent-abuse-monitor.yaml

# The below only runs when Comvent finds matching comment.
# This assumes that Comvent config holding 'some-abusive-content' as a keyword.
- if: steps.comvent.outputs.some-abusive-content != ''
  name: Handle some abusive content
  uses: actions/github-script@v3
  with:
    github-token: ${{ secrets.GITHUB_TOKEN }}
    # An example of removing the abusive comment, and leaving a comment about the deletion.
    script: |
      github.issues.deleteComment({
        owner: context.repo.owner,
        repo: context.repo.repo,
        comment_id: ${{ github.event.comment.id }}
      })

      const comment = `Found abusive comment! 😰
      The comment was thus removed.`;

      github.issues.createComment({
        issue_number: context.issue.number,
        owner: context.repo.owner,
        repo: context.repo.repo,
        body: comment
      })

Click to view full example

.github/workflows/abuse-monitor.yaml

name: Monitor Abusive Comments
on:
  issue_comment:
    # This shouldn't be called for comment deletion
    types:
      - created
      - edited

jobs:
  abuse-monitor:
    name: With latest code base
    runs-on: ubuntu-latest

    steps:
      - name: Handle with Comvent
        uses: rytswd/comvent@v0.3
        id: comvent
        with:
          token: ${{ secrets.GITHUB_TOKEN }}
          config-path: .github/comvent-abuse-monitor.yaml

      # The below only runs when Comvent finds matching comment.
      # This assumes that Comvent config holding 'some-abusive-content' as a keyword.
      - if: steps.comvent.outputs.some-abusive-content != ''
        name: Handle some abusive content
        uses: actions/github-script@v3
        with:
          github-token: ${{ secrets.GITHUB_TOKEN }}
          # An example of removing the abusive comment, and leaving a comment about the deletion.
          script: |
            github.issues.deleteComment({
              owner: context.repo.owner,
              repo: context.repo.repo,
              comment_id: ${{ github.event.comment.id }}
            })

            const comment = `Found abusive comment! 😰
            The comment was thus removed.`;

            github.issues.createComment({
              issue_number: context.issue.number,
              owner: context.repo.owner,
              repo: context.repo.repo,
              body: comment
            })

As used above, Comvent configuration needs to define the keywords to find patterns.

.github/comvent-abuse-monitor.yaml

---
version: 0.2.0

trigger: default

keywords:
  - name: some-abusive-content
    value: 'some abusive content' # Find abusive comment

🧪 Action Inputs

Name	Description	IsRequired
token	GITHUB_TOKEN or Personal Access Token with repo scope	Yes
config-path	Path to Comvent configuration file. You can find more about the configuration below. Defaults to .github/comvent-setup.yaml. Regardless of your GitHub Action setup, this assumes the current directory is the root of your repository.
config-check-only	A flag to run only Comvent configuration setup. Used for testing only.

⚙️ Comvent Configuration File

Comvent uses a dedicated YAML file for its configuration.

Example

The below is the copy of .github/comvent-setup.yaml.

---
# version determines the supported values and how they are handled. The current
# latest is 0.2.0, and if not speccified, it infers the latest version.
version: 0.2.0

# trigger can accept either `default` or `specific`.
# - default:  all comments will be handled by Comvent, except for those from
#             users listed under users.inactive list below.
# - specific: only comments made by specified users under users.active list
#             below gets handled by Comvent.
trigger: default

# users is consisted of `active` or `inactive` user list. Providing a list
# which does not match the `trigger` setup above is simply ignored.
users:
  active:
    - rytswd # This is no-op, as the `trigger` above is set to `default`
  inactive:
    - random-user
    - another-user

# keywords is a list comprised of `name` and `value` fields.
# - name:  used for Comvent output, and thus should not include whitespace.
# - value: regex value to search for. The search takes place for each line in
#          the comment.
keywords:
  - name: command-random
    value: '^\/random$' # Regex which looks for a comment with '/random'

You can find a few examples in .github/ directory.

Configuration Details

Key	Description	Default
version	The version of configuration spec. The latest is 0.2.0. This may not be in line with the released version as of now, until Comvent releases v1.0	0.2.0
trigger	When to handle Comvent. default means event triggered by anyone would be handled by Comvent, and specific means only specified users' comments would be handled.	default
users	Stanza with active or inactive with list of GitHub user accounts. active list is only used for specific trigger setup, and inactive is for default trigger setup.
keywords	Keywords to process comment based on. Each value is regex searched in comment, for each line.

🍸 Action Outputs

Comvent provides the outputs with which you can tweak your Action worrkflow.

As the outputs will be based on the Comvent Configuration File provided to Comvent, the below table is based on the following simple configuration as an example.

.github/comvent-config.yaml

---
version: 0.2.0

trigger: default

keywords:
  - name: command-random
    value: '^\/random$'
  - name: some-other-command
    value: 'some arbitrary regex setup'
  - name: command-echo
    value: '^\/echo (.+)$'

Name	Description
comvent-found-any-match	Special keyword provided by Comvent by default. If any keyword is matched, this will provide found as output.
(Example) command-random	When the RegExp condition is matched, the entire line will be returned as the output. With the above example, this will be /random.
(Example) some-other-command	When the RegExp condition is matched, the entire line will be returned as the output. With the above example, this will be something like Test test - some arbitrary regex setup - test test.
(Example) command-echo	When the RegExp condition with group is matched, the first group match will be returned, and if the first group is an empty string, it would instead return the entire line. For example, with the above example, if the comment was /echo this is a test message., the output will be this is a test message..