validate_datasets.yml

name: validate_datasets

on:
  push:
    branches: [master, main]
    tags: ['*']
  pull_request:
    types: [opened, synchronize, reopened, edited]
    branches: [master, main, 'bep*']
  schedule:
    - cron: "0 4 * * 1"
  workflow_dispatch:

concurrency:
  group: ${{ github.ref }}
  cancel-in-progress: true

defaults:
  run:
    shell: bash

jobs:
  # Prepare the matrix dynamically based on whether a bids-specification PR is referenced
  prepare-matrix:
    runs-on: ubuntu-latest
    outputs:
      matrix: ${{ steps.set-matrix.outputs.matrix }}
      spec_pr: ${{ steps.find-pr.outputs.spec_pr }}
      val_pr: ${{ steps.find-pr.outputs.val_pr }}
    steps:
    - uses: actions/checkout@v6
      with:
        fetch-depth: 0  # Get full history for commit message scanning

    - name: Check for bids-specification PR reference
      id: find-pr
      run: |
        # Check PR description if this is a PR
        if [ "${{ github.event_name }}" == "pull_request" ]; then
          cat << 'EOF' > /tmp/pr_body.txt
        ${{ github.event.pull_request.body }}
        EOF
          SPEC_PR=$(grep -oP 'BIDS Specification PR:\s*(https://github\.com/bids-standard/bids-specification/pulls?/|bids-standard/bids-specification#)*\K[0-9]+' /tmp/pr_body.txt | head -1 || true)
          [ -n "$SPEC_PR" ] && echo "spec_pr=$SPEC_PR" >> $GITHUB_OUTPUT || :
          VAL_PR=$(grep -oP 'BIDS Validator PR:\s*(https://github\.com/bids-standard/bids-validator/pulls?/|bids-standard/bids-validator#)*\K[0-9]+' /tmp/pr_body.txt | head -1 || true)
          [ -n "$VAL_PR" ] && echo "val_pr=$VAL_PR" >> $GITHUB_OUTPUT || :
        fi

    - name: Set matrix
      id: set-matrix
      run: |
        EXTRA_ITEM=''
        if [ -n "${{ steps.find-pr.outputs.spec_pr }}${{ steps.find-pr.outputs.val_pr }}" ]; then
          EXTRA_ITEM=', "dev-prs"'
        fi
        echo "matrix=[\"stable\", \"main\", \"dev\"${EXTRA_ITEM}]" >> $GITHUB_OUTPUT

  build:
    needs: prepare-matrix
    strategy:
      fail-fast: false
      matrix:
        platform: [ubuntu-latest, macos-latest, windows-latest]
        bids-validator: ${{ fromJson(needs.prepare-matrix.outputs.matrix) }}

    runs-on: ${{ matrix.platform }}

    env:
      SPEC_PR: ${{ needs.prepare-matrix.outputs.spec_pr }}
      VAL_PR: ${{ needs.prepare-matrix.outputs.val_pr }}
      TZ: Europe/Berlin
      FORCE_COLOR: 1

    steps:
    - uses: actions/checkout@v6

    - uses: denoland/setup-deno@v2
      with:
        deno-version: v2.x

    - name: Install BIDS validator (stable)
      if: matrix.bids-validator == 'stable'
      run: |
          deno install -Agf -n bids-validator jsr:@bids/validator

    - name: Install BIDS validator (main)
      if: matrix.bids-validator == 'main'
      run: |
        # If unmerged validator PRs are needed for testing, you can use
        # https://github.com/<FORK>/bids-validator/raw/<BRANCH>/bids-validator/src/bids-validator.ts
        deno install -Agf https://github.com/bids-standard/bids-validator/raw/deno-build/bids-validator.js

    - name: Install BIDS validator (dev)
      if: ( matrix.bids-validator == 'dev' ) || ( matrix.bids-validator == 'dev-prs' && env.VAL_PR == '' )
      run: |
        git clone -b dev https://github.com/bids-standard/bids-validator/ ../bids-validator
        cd ../bids-validator
        deno compile -A -o $HOME/.deno/bin/bids-validator src/bids-validator.ts

    - name: Install BIDS validator (dev-prs with VAL_PR)
      if: matrix.bids-validator == 'dev-prs' && env.VAL_PR != ''
      shell: bash
      env:
        GH_TOKEN: ${{ github.token }}
      run: |
        # Fetch PR info using gh cli (handles auth automatically)
        fork_url=$(gh api repos/bids-standard/bids-validator/pulls/${{ env.VAL_PR }} --jq '.head.repo.html_url')
        branch=$(gh api repos/bids-standard/bids-validator/pulls/${{ env.VAL_PR }} --jq '.head.ref')
        echo "I: Received fork_url=$fork_url branch=$branch"
        if [ "$fork_url" = "null" ] || [ -z "$fork_url" ]; then
          echo "E: Failed to get PR info for VAL_PR=${{ env.VAL_PR }}"
          exit 1
        fi
        # Perform the same installation as in "dev"
        git clone -b "$branch" "$fork_url" ../bids-validator
        cd ../bids-validator
        deno compile -A -o $HOME/.deno/bin/bids-validator src/bids-validator.ts

    - name: Display versions and environment information
      run: |
        echo $TZ
        date
        which deno && echo "deno\n----" && deno --version || true
        echo "node\n----"; node --version
        echo "npm\n----"; npm --version
        echo "bids-validator"; bids-validator --version

    - name: Check that no large files are present
      if: "matrix.bids-validator == 'stable'"
      run: |
        echo "Checking for big files ..."
        found=`find . -not -path "./.git*" -type f -size +1M`
        if [ "$found" == "" ]
        then
            echo "No big files present, great!"
        else
            echo "Found big files:"
            echo "$found"
            exit 1;
        fi

    - name: Skip stable validation for datasets with unreleased validator features
      # Replace ${EMPTY} with dataset patterns, when this is needed
      # Reset to "for DS in ${EMPTY}; ..." after a validator release
      run: for DS in ${EMPTY}; do touch $DS/.SKIP_VALIDATION; done
      if: matrix.bids-validator == 'stable'

    - name: Skip main validation for datasets with unreleased spec features
      # Replace ${EMPTY} with dataset patterns, when this is needed
      # Reset to "for DS in ${EMPTY}; ..." after a spec release
      run: for DS in ${EMPTY}; do touch $DS/.SKIP_VALIDATION; done
      if: matrix.bids-validator != 'dev'

    - name: Set BIDS_SCHEMA variable for dev version
      if: matrix.bids-validator == 'dev'
      # When proposing new features, schema changes may be necessary.
      # Update this URL to the schema.json from PRs to the spec, when needed.
      # If this variable is unset, dev will generally track the latest development
      # release of https://jsr.io/@bids/schema
      run: echo BIDS_SCHEMA=https://bids-specification.readthedocs.io/en/latest/schema.json >> $GITHUB_ENV

    - name: Set BIDS_SCHEMA variable for dev-prs version
      if: matrix.bids-validator == 'dev-prs' && env.SPEC_PR != ''
      # Use the readthedocs PR preview build for the schema
      run: |
        echo "Using schema from bids-specification PR #${{ env.SPEC_PR }}"
        echo BIDS_SCHEMA=https://bids-specification--${{ env.SPEC_PR }}.org.readthedocs.build/en/${{ env.SPEC_PR }}/schema.json >> $GITHUB_ENV

    - name: Validate all BIDS datasets using bids-validator
      run: |
        cat ./run_tests.sh
        ./run_tests.sh