plopcas/emoji-issue-classifier

Emoji Issue Classifier

The Emoji Issue Classifier is a GitHub Action that automatically classifies issues by adding relevant emoji labels. It performs sentiment analysis, keyword matching, and machine learning-based classification to add relevant emoji labels to the issues.

Usage

1. Create a .github/workflows/emoji-classifier.yml file in your repository.
2. Add the following workflow configuration to the file:

# .github/workflows/issue-classifier.yml
name: 'Use Emoji Issue Classifier Action'

on:
  issues:
    types: [opened, edited]

jobs:
  classify:
    runs-on: ubuntu-latest

    steps:
    - name: Use Emoji Classifier
      uses: plopcas/emoji-issue-classifier@v1.1.0
      env:
        GITHUB_TOKEN: ${{ secrets.PAT }}

3. Configure the necessary environment variables:
   • GITHUB_TOKEN: GitHub token with repository access.
4. Customize the behavior of the Emoji Issue Classifier by modifying the Python script (main.py) based on your requirements.
5. Run the Emoji Issue Classifier by executing python main.py in the repository directory.

Test

In order to test the classifier you can:

1. Create a test.json file with the following content:

{
    "issue": {
      "number": 1
    }
}

2. Create a test.sh file with the following content:

#!/bin/bash
# test.sh

export GITHUB_TOKEN=YOUR_TOKEN
export GITHUB_REPOSITORY=YOUR_REPOSITORY
export GITHUB_EVENT_PATH=test/test.json
python classifier/main.py

3. Ensure that the bash script (test.sh) has the correct permissions to be executed.

chmod +x test.sh

4. Run run this bash script from the terminal with:

./test.sh

Customization

You can customize the behavior of the Emoji Issue Classifier by modifying the following components:

• Sentiment Analysis: adjust the sentiment thresholds and corresponding emojis in the get_sentiment_emoji function.
• Keyword Matching: add or modify keywords and corresponding emojis in the get_keyword_emoji function.
• Trained Model: train and save a new model using the provided train script and replace the model.pkl file.
• Label Color: change the label color in the add_label_to_issue function by modifying the hex color code.

Contributing

Contributions are welcome! If you have any ideas, suggestions, or bug reports, please open an issue or submit a pull request.

License

This project is licensed under the MIT License.