push-to-main.yml

on:
  push:
    branches:
      - main

env:
  TAG: "latest"

jobs:
  check-branch:
    runs-on: ubuntu-latest

    outputs:
      tag: ${{ steps.image-tag.outputs.tag }}

    steps:
    - uses: actions/checkout@v4
    - name: Find Image Tag
      id: image-tag
      env:
        BRANCH: ${{ github.ref_name }}
        COMMIT: ${{ github.sha }}
      run: |
          if [ "${{ github.event_name }}" == 'pull_request' ]; then
            echo "tag=pr-${{ github.event.number }}" >> "$GITHUB_OUTPUT"
          else
            if [ "$BRANCH" == "main" ]; then
                echo "tag=${{ env.TAG }}" >> "$GITHUB_OUTPUT"
            else
                echo "tag=$COMMIT" >> "$GITHUB_OUTPUT"
            fi
          fi

  check-change:
    runs-on: ubuntu-latest

    outputs:
      base: ${{ steps.filter.outputs.base }}
      modeling: ${{ steps.filter.outputs.modeling }}
      s3: ${{ steps.filter.outputs.s3 }}

    steps:
    - uses: actions/checkout@v4
    - uses: dorny/paths-filter@v3
      id: filter
      with:
        filters: |
          base:
            - 'pyproject.toml'
            - 'dockerfiles/Dockerfile.base'
            - '.github/workflows/build-push.yml'
          modeling:
            - 'src/**'
            - 'model_training/**'
            - 'hack/**'
            - '.github/workflows/train-model.yml'
          s3:
            - 'model_training/s3/**'

  build-push:
    needs: [check-change, check-branch]
    uses: ./.github/workflows/build-push.yml
    with:
      base_change: ${{ needs.check-change.outputs.base }}
      s3_change: ${{ needs.check-change.outputs.s3 }}
      image_repo: ${{ vars.IMAGE_REPO }}
      image_tag: ${{ needs.check-branch.outputs.tag }}
      push: true
    secrets:
      docker_username: ${{ secrets.BOT_NAME }}
      docker_password: ${{ secrets.BOT_TOKEN }}

  train-model:
    needs: [check-change, check-branch, build-push]
    if: ${{ needs.check-change.outputs.modeling == 'true' }}
    strategy:
      matrix:
        instance_type: [i3.metal]
    uses: ./.github/workflows/train-model.yml
    with:
      pipeline_name: std_v0.7.11
      instance_type: ${{ matrix.instance_type }}
      ami_id: 'ami-0e4d0bb9670ea8db0'
      github_repo: ${{ github.repository }}
      model_server_image: ${{ vars.IMAGE_REPO }}/kepler_model_server:${{ needs.check-branch.outputs.tag }}
      trainers: LogisticRegressionTrainer,ExponentialRegressionTrainer,SGDRegressorTrainer,GradientBoostingRegressorTrainer,XgboostFitTrainer
    secrets:
      self_hosted_github_token: ${{ secrets.GH_SELF_HOSTED_RUNNER_TOKEN }}
      aws_access_key_id: ${{ secrets.AWS_ACCESS_KEY_ID }}
      aws_secret_access_key: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
      aws_region: ${{ secrets.AWS_REGION }}