I'm really interested in this. What would be the best way to implement it?
I've been checking the github::calculate_total_modifications() method and the Github API docs and the only approach I found would be:

• Allow a 8th argument to be passed to the action (ignore_files), which would be a string with the comma-separated files to ignore.
• Inside calculate_total_modifications() define a variable totalLines = deteledLines + addedLines (current returned value)
• Loop over every file in the PR (using this endpoint GET /repos/:owner/:repo/pulls/:pull_number/files) and check if the filename matches one of the ignored_files. If it does, substract its (deletedLines + addedLines) from the totalLines.

Does this make sense? Am I missing anything?

It would be awesome to support this 🙌

@dowrow, your approach seems to completely make sense 🙂

The only thing I would suggest would be using a YAML sequence instead of a comma separated string for specifying the different files to ignore. That is, being able to specify it as files_to_ignore: ['package-lock.json', 'composer.lock'] instead of files_to_ignore: 'package-lock.json, composer.lock'.

Example of how the README.md complete example would look like:

name: labeler

on: [pull_request]

jobs:
  labeler:
    runs-on: ubuntu-latest
    name: Label the PR size
    steps:
      - uses: codelytv/pr-size-labeler@v1
        with:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          xs_max_size: '10'
          s_max_size: '100'
          m_max_size: '500'
          l_max_size: '1000'
          fail_if_xl: 'false'
          message_if_xl: 'This PR is so big! Please, split it 😊'
          files_to_ignore: ['package-lock.json', 'composer.lock']

What do you think?