Skip to content
compass

GitHub Action

IssueWhiz

v1 Latest version

IssueWhiz

compass

IssueWhiz

Automated Issue Triaging

Installation

Copy and paste the following snippet into your .yml file.

 
              
- name: IssueWhiz

              
  uses: pierotofy/issuewhiz@v1
Learn more about this action in pierotofy/issuewhiz

Choose a version

IssueWhiz - Automated Issue Triaging

License GitHub Issues GitHub Pull Requests GitHub Stars GitHub Forks GitHub Release

IssueWhiz

IssueWhiz automates the triaging of issues in your repositories. It helps streamline the process of categorizing issues and enforcing project's guidelines using customizable and flexible rules.

Features

  • Automatically label, close and comment on newly opened issues.
  • Rules defined using boolean logic coupled with variables computed with LLM-based text classification.
  • Save time and reduce manual triaging efforts.
  • Easily customizable to fit specific needs.

In a nutshell, you can define boolean questions using natural language, such as:

  • A: Does this text look like a software bug report?
  • B: Is this about a frontend problem?
  • C: Is this about a backend problem?

And define rules to act on such questions:

  • if A and B --> Add label bug frontend
  • if A and C --> Add label bug backend
  • if A --> Add label bug
  • else --> Add comment Thanks for opening an issue! We will triage this shortly.

Usage

  • Create a .github/workflows/issuewhiz.yml in your repository:
name: Issue Triage
on:
  issues:
    types:
      - opened
jobs:
  issue_triage:
    runs-on: ubuntu-latest
    permissions:
      issues: write
    steps:
      - uses: pierotofy/issuewhiz@v1
        with:
          # Your OpenAI token (https://platform.openai.com/api-keys)
          # Added to : https://github.com/<repo>/<name>/settings/secrets actions
          openAI: ${{ secrets.OPENAI_TOKEN }}

          # GitHub Token with write access to repository issues
          # (you can leave this unless you want to use a different user that "github-actions")
          ghToken: ${{ secrets.GITHUB_TOKEN }}

          # Optional filter, any line in an issue that begins with "#" 
          # will be discarded prior to LLM evaluation 
          filter: |
            - "#"
          
          # Define your boolean questions, which must evaluate to yes/no or true/false
          # <VAR>: <QUESTION>
          variables: |
            - A: "A question about using a software or seeking guidance on doing something?"
            - B: "Contains a suggestion for an improvement or a feature request?"
          
          # Boolean expressions evaluating to true will execute the actions listed next to it
          # Evaluation continues until there are no more expressions, or a "stop: true" action is found
          logic: |
            - "A and (not B)": [label: question, stop: true]
            - "B and (not A)": [label: enhancement, stop: true]
            - "(not A) and (not B)": [comment: "I'm not sure how to classify this one! I will close the issue", close: true]

          # Optional signature to append to each message posted by the bot
          signature: "p.s. I'm just an automated script, not a human being."

Expressions

You can reference all of your variables in the boolean expressions, as well as the special body, title and title_lowercase variables, which contains the text of the issue body and title respectively. This can be useful for doing classical regex matches alongside LLM evaluation.

For example, to match the substring "bug:" in the issue title:

logic: |
  - 'title ~= "bug:"': [stop: true]

You can use all the expressions supported by Filtrex:

There are only 2 types: numbers and strings. Numbers may be floating point or integers. Boolean logic is applied on the truthy value of values (e.g. any non-zero number is true, any non-empty string is true, otherwise false).

Values Description
43, -1.234 Numbers
"hello" String
foo, a.b.c, 'foo-bar' External data variable defined by application (may be numbers or strings)
Numeric arithmetic Description
x + y Add
x - y Subtract
x * y Multiply
x / y Divide
x % y Modulo
x ^ y Power
Comparisons Description
x == y Equals
x < y Less than
x <= y Less than or equal to
x > y Greater than
x >= y Greater than or equal to
x ~= y Regular expression match
x in (a, b, c) Equivalent to (x == a or x == b or x == c)
x not in (a, b, c) Equivalent to (x != a and x != b and x != c)
Boolean logic Description
x or y Boolean or
x and y Boolean and
not x Boolean not
x ? y : z If boolean x, value y, else z
( x ) Explicit operator precedence
Built-in functions Description
abs(x) Absolute value
ceil(x) Round floating point up
floor(x) Round floating point down
log(x) Natural logarithm
max(a, b, c...) Max value (variable length of args)
min(a, b, c...) Min value (variable length of args)
random() Random floating point from 0.0 to 1.0
round(x) Round floating point
sqrt(x) Square root

Actions

Action Description
comment Add a comment to the issue
close Close the issue
stop Stop executing actions
label Assign a label to the issue

Need more? Help us.

Roadmap

We welcome contributions!

Here's some ideas:

  • Adding support for more LLM backend APIs like LLAMA
  • Add support for more actions

Or propose something by opening a pull request!

License

AGPLv3

Build

npm i -g @vercel/ncc
ncc build index.js