huggingface-nightly.yml

name: huggingface-nightly-build-and-publish

on:
  schedule:
    # Nightly at 03:17 UTC
    - cron: "17 3 * * *"
  workflow_dispatch:
    inputs:
      upload_to:
        description: "Where to upload (none/testpypi/pypi)"
        required: true
        default: "testpypi"
        type: choice
        options:
          - none
          - testpypi
          - pypi
      skip_existing:
        description: "Skip already-uploaded versions"
        required: true
        default: true
        type: boolean

permissions:
  contents: read

jobs:
  build:
    runs-on: ubuntu-latest
    strategy:
      fail-fast: false
      matrix:
        repo:
          - id: nemotron-page-elements-v3
            url: https://huggingface.co/nvidia/nemotron-page-elements-v3
            project_subdir: ""
          - id: nemotron-table-structure-v1
            url: https://huggingface.co/nvidia/nemotron-table-structure-v1
            project_subdir: ""
          - id: nemotron-graphic-elements-v1
            url: https://huggingface.co/nvidia/nemotron-graphic-elements-v1
            project_subdir: ""
          - id: llama-nemotron-embed-1b-v2
            url: https://huggingface.co/nvidia/llama-nemotron-embed-1b-v2
            # Repo layout matches nemotron-ocr-v1: Python project is nested under a same-named subdir.
            project_subdir: llama-nemotron-embed-1b-v2
          - id: llama-nemotron-rerank-1b-v2
            url: https://huggingface.co/nvidia/llama-nemotron-rerank-1b-v2
            project_subdir: ""

    steps:
      - name: Checkout orchestrator repo
        uses: actions/checkout@v4

      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: "3.12"

      - name: Install git-lfs
        run: |
          sudo apt-get update
          sudo apt-get install -y git-lfs
          git lfs install

      - name: Decide upload target
        id: target
        shell: bash
        run: |
          set -euo pipefail
          # Default for scheduled runs: testpypi
          upload_to="testpypi"
          skip_existing="true"
          if [[ "${{ github.event_name }}" == "workflow_dispatch" ]]; then
            upload_to="${{ inputs.upload_to }}"
            skip_existing="${{ inputs.skip_existing }}"
          fi
          echo "upload_to=${upload_to}" >> "$GITHUB_OUTPUT"
          echo "skip_existing=${skip_existing}" >> "$GITHUB_OUTPUT"

      - name: Build (and maybe upload)
        env:
          TEST_PYPI_API_TOKEN: ${{ secrets.TEST_PYPI_API_TOKEN }}
          PYPI_API_TOKEN: ${{ secrets.PYPI_API_TOKEN }}
        shell: bash
        run: |
          set -euo pipefail

          upload_flag=""
          repo_url="https://test.pypi.org/legacy/"
          token_env="TEST_PYPI_API_TOKEN"
          if [[ "${{ steps.target.outputs.upload_to }}" == "none" ]]; then
            upload_flag=""
          else
            upload_flag="--upload"
          fi
          if [[ "${{ steps.target.outputs.upload_to }}" == "pypi" ]]; then
            repo_url="https://upload.pypi.org/legacy/"
            token_env="PYPI_API_TOKEN"
          fi

          skip_existing_flag=""
          if [[ "${{ steps.target.outputs.skip_existing }}" == "true" ]]; then
            skip_existing_flag="--skip-existing"
          fi

          python ci/scripts/nightly_build_publish.py \
            --repo-id "${{ matrix.repo.id }}" \
            --repo-url "${{ matrix.repo.url }}" \
            --work-dir ".work" \
            --dist-dir "dist-out" \
            --project-subdir "${{ matrix.repo.project_subdir }}" \
            ${upload_flag} \
            --repository-url "${repo_url}" \
            --token-env "${token_env}" \
            ${skip_existing_flag}

      - name: Upload build artifacts (GH Actions)
        if: always()
        uses: actions/upload-artifact@v4
        with:
          name: dist-${{ matrix.repo.id }}
          path: dist-out/${{ matrix.repo.id }}/*

  build_ocr_cuda:
    # nemotron-ocr-v1 needs nvcc/CUDA headers to build its extension.
    # Build with Python 3.12 to match upstream package constraints and
    # avoid producing an extension for the wrong Python ABI.
    runs-on: ubuntu-latest
    container:
      # Build extension with CUDA 13 toolchain (nvcc/headers from devel image).
      image: nvidia/cuda:13.0.0-devel-ubuntu24.04
    steps:
      - name: Install system deps (git, lfs, build tools)
        shell: bash
        run: |
          set -euo pipefail
          apt-get update
          apt-get install -y --no-install-recommends \
            ca-certificates \
            git \
            git-lfs \
            build-essential \
            ninja-build \
            python3 \
            python3-venv \
            python3-pip
          git lfs install

      - name: Checkout orchestrator repo
        uses: actions/checkout@v4

      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: "3.12"

      - name: Decide upload target
        id: target
        shell: bash
        run: |
          set -euo pipefail
          # Default for scheduled runs: testpypi
          upload_to="testpypi"
          skip_existing="true"
          if [[ "${{ github.event_name }}" == "workflow_dispatch" ]]; then
            upload_to="${{ inputs.upload_to }}"
            skip_existing="${{ inputs.skip_existing }}"
          fi
          echo "upload_to=${upload_to}" >> "$GITHUB_OUTPUT"
          echo "skip_existing=${skip_existing}" >> "$GITHUB_OUTPUT"

      - name: Build nemotron-ocr-v1 (and maybe upload)
        env:
          TEST_PYPI_API_TOKEN: ${{ secrets.TEST_PYPI_API_TOKEN }}
          PYPI_API_TOKEN: ${{ secrets.PYPI_API_TOKEN }}
          CUDA_HOME: /usr/local/cuda
          BUILD_CPP_EXTENSION: "1"
          BUILD_CPP_FORCE: "1"
          TORCH_CUDA_ARCH_LIST: "8.0;8.6;8.9;9.0;10.0;12.0+PTX"
          PIP_EXTRA_INDEX_URL: "https://download.pytorch.org/whl/cu130"
        shell: bash
        run: |
          set -euo pipefail

          upload_flag=""
          repo_url="https://test.pypi.org/legacy/"
          token_env="TEST_PYPI_API_TOKEN"
          if [[ "${{ steps.target.outputs.upload_to }}" == "none" ]]; then
            upload_flag=""
          else
            upload_flag="--upload"
          fi
          if [[ "${{ steps.target.outputs.upload_to }}" == "pypi" ]]; then
            repo_url="https://upload.pypi.org/legacy/"
            token_env="PYPI_API_TOKEN"
          fi

          skip_existing_flag=""
          if [[ "${{ steps.target.outputs.skip_existing }}" == "true" ]]; then
            skip_existing_flag="--skip-existing"
          fi

          python --version
          python ci/scripts/nightly_build_publish.py \
            --repo-id "nemotron-ocr-v1" \
            --repo-url "https://huggingface.co/nvidia/nemotron-ocr-v1" \
            --work-dir ".work" \
            --dist-dir "dist-out" \
            --project-subdir "nemotron-ocr" \
            --build-no-isolation \
            --venv-pip-install "hatchling" \
            --venv-pip-install "setuptools>=68" \
            --venv-pip-install "torch==2.9.1" \
            --venv-pip-install "torchvision==0.24.1" \
            --build-env "BUILD_CPP_EXTENSION=1" \
            --build-env "BUILD_CPP_FORCE=1" \
            ${upload_flag} \
            --repository-url "${repo_url}" \
            --token-env "${token_env}" \
            ${skip_existing_flag}

          python - <<'PY'
          import sys
          import zipfile
          from pathlib import Path

          wheels = sorted(Path("dist-out/nemotron-ocr-v1").glob("*.whl"))
          if not wheels:
              raise SystemExit("No wheel found in dist-out/nemotron-ocr-v1")

          py_tag = f"cpython-{sys.version_info.major}{sys.version_info.minor}"
          matches = []
          for wheel in wheels:
              with zipfile.ZipFile(wheel) as zf:
                  names = zf.namelist()
              for name in names:
                  if "nemotron_ocr_cpp/_nemotron_ocr_cpp." in name:
                      matches.append((wheel.name, name))

          if not matches:
              raise SystemExit("Built wheel is missing nemotron_ocr_cpp extension artifact")

          if not any(py_tag in ext for _, ext in matches):
              formatted = "\n".join(f"{w}: {ext}" for w, ext in matches)
              raise SystemExit(
                  f"Built extension ABI does not match runner Python ({py_tag}). Found:\n{formatted}"
              )
          print("Verified nemotron_ocr_cpp extension ABI:", py_tag)
          PY

      - name: Upload build artifacts (GH Actions)
        if: always()
        uses: actions/upload-artifact@v4
        with:
          name: dist-nemotron-ocr-v1
          path: dist-out/nemotron-ocr-v1/*