web_scrap_dynamic_csv_files.yml

name: update-dynamic-csv-data-files

on:
  workflow_dispatch:
  schedule:
    - cron: '0 0 2 * *' # “At 00:00 on day-of-month 2” (on the 2nd of the month)

jobs:
  build:
    runs-on: ubuntu-latest #ubuntu-20.04 # specifically calls the version number instead of latest to work on 'act'

    steps:
    - uses: actions/checkout@v4
    - name: Set up Python
      uses: actions/setup-python@v5
      with:
        python-version: '3.12'
        architecture: 'x64'
      env:
          AGENT_TOOLSDIRECTORY: /opt/hostedtoolcache

    - name: pip install requirements
      run: |
        export SKLEARN_ALLOW_DEPRECATED_SKLEARN_PACKAGE_INSTALL=True
        pip install --upgrade pip
        pip install -r requirements.txt
        pip install pydar

    - name: Web Scrap for New Changes - swath_coverage_by_time_position.csv
      working-directory: pydar
      run: |
        python updateCsvSwathCoverage.py

    - name: Web Scrap for New Changes - coradr_jpl_options.csv
      working-directory: pydar
      run: |
        python updateCsvCORADRJPLOptions.py

    - name: Web Scrap for New Changes - feature_name_details.csv
      working-directory: pydar
      run: |
        python updateCsvFeatureNameDetails.py

    - name: Check if script results in changes in files
      uses: tj-actions/verify-changed-files@v20
      id: verify-changed-files

    - name: If file changed, throw error to review changes
      if: steps.verify-changed-files.outputs.files_changed == 'true'
      run: |
        echo "Github Actions: review changes to data files"
        echo "Update README, Pytests, update .CSV"
        exit 1

    - name: If no changes to existing files are found, pass
      if: steps.verify-changed-files.outputs.files_changed == 'false'
      run: |
        echo "Github Actions: update csv scripts has no changes"