# Copyright Advanced Micro Devices, Inc.
# SPDX-License-Identifier: MIT

name: Multi-Arch CI - Windows

on:
  workflow_call:
    inputs:
      build_config:
        type: string
        description: >-
          JSON object with build configuration for this platform. Fields:
          artifact_group, per_family_info, dist_amdgpu_families,
          build_variant_label, build_variant_cmake_preset,
          build_variant_suffix, build_pytorch,
          build_native_linux, test_python_packages_matrix,
          pytorch_build_matrix,
          prebuilt_stages, baseline_run_id.
      test_labels:
        type: string
      rocm_package_version:
        type: string
      test_type:
        type: string
      release_type:
        description: 'Release type: "ci", "dev", "nightly", or "prerelease".'
        type: string
        default: ci
      external_repo_config:
        description: "JSON config for external repo checkout (repository, ref, checkout_path, source_package)"
        type: string
        default: ""
      changed_projects:
        description: 'Comma-separated list of projects to test (e.g., "rocblas,hipblas" or "projects/rocblas"). Empty = run all tests.'
        type: string
        default: ""
      repository:
        description: "Repository to checkout. Defaults to github.repository."
        type: string
        default: ""
      ref:
        description: "Branch, tag, or SHA to checkout. Defaults to the triggering ref."
        type: string
        default: ""

permissions:
  contents: read

jobs:
  copy_prebuilt_stages:
    name: Copy Prebuilt Stages
    if: ${{ fromJSON(inputs.build_config).prebuilt_stages != '' && fromJSON(inputs.build_config).baseline_run_id != '' }}
    # TODO: Consider running on a Linux runner with --platform=windows to
    #   avoid Windows runner setup overhead (setup-python ~51s).
    runs-on: ${{ fromJSON(inputs.build_config).build_runs_on || 'azure-windows-scale-rocm' }}
    defaults:
      run:
        shell: bash
    permissions:
      contents: read
      id-token: write
    steps:
      - name: Checking out repository
        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
        with:
          repository: ${{ inputs.repository || github.repository }}
          ref: ${{ inputs.ref }}

      - name: Set up Python
        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
        with:
          python-version: "3.12"

      - name: Install artifact manager dependencies
        run: pip install boto3

      - name: Configure AWS credentials for artifact uploads
        uses: ./.github/actions/configure_aws_artifacts_credentials
        with:
          release_type: ${{ inputs.release_type }}

      - name: Copy prebuilt stage artifacts
        env:
          GITHUB_TOKEN: ${{ github.token }}
          RELEASE_TYPE: ${{ inputs.release_type }}
        run: |
          python build_tools/artifact_manager.py copy \
            --source-run-id=${{ fromJSON(inputs.build_config).baseline_run_id }} \
            --source-repository=${{ fromJSON(inputs.build_config).baseline_repository }} \
            --stage="${{ fromJSON(inputs.build_config).prebuilt_stages }}" \
            --amdgpu-families="${{ fromJSON(inputs.build_config).dist_amdgpu_families }}" \
            --expand-family-to-targets

  build_multi_arch_stages:
    name: Build Multi-Arch Stages
    needs: copy_prebuilt_stages
    if: ${{ !cancelled() && !failure() }}
    uses: ./.github/workflows/multi_arch_build_windows.yml
    secrets: inherit
    with:
      matrix_per_family_json: ${{ toJSON(fromJSON(inputs.build_config).per_family_info) }}
      dist_amdgpu_families: ${{ fromJSON(inputs.build_config).dist_amdgpu_families }}
      artifact_group: ${{ fromJSON(inputs.build_config).artifact_group }}
      build_variant_label: ${{ fromJSON(inputs.build_config).build_variant_label }}
      build_variant_cmake_preset: ${{ fromJSON(inputs.build_config).build_variant_cmake_preset }}
      build_variant_suffix: ${{ fromJSON(inputs.build_config).build_variant_suffix }}
      prebuilt_stages: ${{ fromJSON(inputs.build_config).prebuilt_stages }}
      build_runs_on: ${{ fromJSON(inputs.build_config).build_runs_on }}
      rocm_package_version: ${{ inputs.rocm_package_version }}
      release_type: ${{ inputs.release_type }}
      external_repo_config: ${{ inputs.external_repo_config }}
      repository: ${{ inputs.repository }}
      ref: ${{ inputs.ref }}
    permissions:
      contents: read
      id-token: write

  validate_artifact_structure:
    needs: [build_multi_arch_stages]
    name: Validate Artifact Structure
    if: ${{ !failure() && !cancelled() }}
    uses: ./.github/workflows/test_artifacts_structure.yml
    with:
      amdgpu_families: ${{ fromJSON(inputs.build_config).dist_amdgpu_families }}
      platform: windows
      release_type: ${{ inputs.release_type }}
      repository: ${{ inputs.repository }}
      ref: ${{ inputs.ref }}

  test_artifacts_per_family:
    needs: [copy_prebuilt_stages, build_multi_arch_stages]
    name: Test ${{ matrix.family_info.amdgpu_family }}
    if: ${{ !failure() && !cancelled() }}
    strategy:
      fail-fast: false
      matrix:
        family_info: ${{ fromJSON(inputs.build_config).per_family_info }}
    uses: ./.github/workflows/test_artifacts.yml
    with:
      # Use architecture-specific artifact group for fetching per-arch artifacts
      artifact_group: ${{ matrix.family_info.amdgpu_family }}
      amdgpu_families: ${{ matrix.family_info.amdgpu_family }}
      amdgpu_targets: ${{ matrix.family_info.amdgpu_targets }}
      test_runs_on: ${{ matrix.family_info.test-runs-on }}
      test_type: ${{ inputs.test_type }}
      test_labels: ${{ inputs.test_labels }}
      sanity_check_only_for_family: ${{ matrix.family_info.sanity_check_only_for_family }}
      changed_projects: ${{ inputs.changed_projects }}
      release_type: ${{ inputs.release_type }}
      repository: ${{ inputs.repository }}
      ref: ${{ inputs.ref }}

  build_python_packages:
    needs: [build_multi_arch_stages]
    name: Build Python Packages
    if: ${{ !failure() && !cancelled() }}
    uses: ./.github/workflows/build_windows_python_packages.yml
    with:
      artifact_group: ${{ fromJSON(inputs.build_config).artifact_group }}
      amdgpu_families: ${{ fromJSON(inputs.build_config).dist_amdgpu_families }}
      linux_amdgpu_families: ${{ fromJSON(inputs.build_config).linux_amdgpu_families }}
      windows_amdgpu_families: ${{ fromJSON(inputs.build_config).windows_amdgpu_families }}
      multiarch_index: true
      package_version: ${{ inputs.rocm_package_version }}
      release_type: ${{ inputs.release_type }}
      repository: ${{ inputs.repository }}
      ref: ${{ inputs.ref }}
    permissions:
      contents: read
      id-token: write

  test_python_packages_per_family:
    needs: [build_python_packages]
    name: Test Python
    if: ${{ !failure() && !cancelled() && toJSON(fromJSON(inputs.build_config).test_python_packages_matrix) != '[]' }}
    strategy:
      fail-fast: false
      matrix:
        include: ${{ fromJSON(inputs.build_config).test_python_packages_matrix }}
    uses: ./.github/workflows/test_rocm_wheels.yml
    with:
      amdgpu_family: ${{ matrix.amdgpu_family }}
      test_runs_on: ${{ matrix.test_runs_on }}
      # TODO: Simplify to just `needs.build_python_packages.outputs.package_find_links_url`
      # once kpack split is always enabled.
      package_find_links_url: >-
        ${{
          needs.build_python_packages.outputs.kpack_split == 'true'
          && needs.build_python_packages.outputs.package_find_links_url
          || format('{0}/{1}/index.html',
              needs.build_python_packages.outputs.package_find_links_url,
              matrix.amdgpu_family)
        }}
      python_version: ${{ matrix.python_version }}
      rocm_version: ${{ inputs.rocm_package_version }}
      container_image_name: ${{ matrix.container_image_name }}
      container_image_url: ${{ matrix.container_image_url }}
      repository: ${{ inputs.repository }}
      ref: ${{ inputs.ref }}

  # Multi-arch PyTorch build: one fat wheel covering every target, then
  # split into a host wheel + per-gfx device wheels via rocm-kpack.
  # KPACK_SPLIT_ARTIFACTS is always ON in multi-arch CI (FLAGS.cmake default),
  # so the split path is unconditional here. The split itself runs inline
  # in multi_arch_build_windows_pytorch_wheels_ci.yml based on
  # inputs.kpack_split.
  build_pytorch_wheel_fat:
    needs: [build_python_packages]
    name: Build PyTorch (fat + split)
    if: >-
      ${{
        !failure() &&
        !cancelled() &&
        fromJSON(inputs.build_config).build_pytorch == true &&
        toJSON(fromJSON(inputs.build_config).pytorch_build_matrix) != '[]'
      }}
    strategy:
      fail-fast: false
      matrix:
        include: ${{ fromJSON(inputs.build_config).pytorch_build_matrix }}
    uses: ./.github/workflows/multi_arch_build_windows_pytorch_wheels_ci.yml
    with:
      artifact_group: ${{ fromJSON(inputs.build_config).artifact_group }}
      python_version: ${{ matrix.python_version }}
      pytorch_git_ref: ${{ matrix.pytorch_git_ref }}
      # The reusable workflow expands this to gfx targets via
      # cmake/therock_amdgpu_targets.cmake.
      amdgpu_families: ${{ matrix.amdgpu_families }}
      kpack_split: ${{ needs.build_python_packages.outputs.kpack_split }}
      rocm_package_find_links_url: ${{ needs.build_python_packages.outputs.package_find_links_url }}
      rocm_version: ${{ inputs.rocm_package_version }}
      repository: ${{ inputs.repository }}
      ref: ${{ inputs.ref }}
    permissions:
      contents: read
      id-token: write