[CI] Add multi-hardware wheel build and release workflow (#7312)

### What this PR does / why we need it?
Adds a scheduled CI workflow (schedule_release_code_and_wheel.yml) to
automatically build and release vllm-ascend source packages and binary
wheels for multiple Ascend hardware targets.

Key features:

1. Source release: Builds tar.gz sdist and uploads to PyPI on version
tag push
2. Multi-hardware wheel builds: Supports three hardware targets in
parallel:
2.1 A2 (Ascend 910B): x86_64 + ARM64, Python 3.10 / 3.11
2.2 A3 (Ascend 910C): x86_64 + ARM64, Python 3.10 / 3.11
2.3 310P: x86_64 + ARM64, Python 3.10 / 3.11
3. Wheel repair: Uses auditwheel to produce manylinux-compatible wheels,
excluding Ascend NPU runtime libs (libascend*.so, libtorch*.so, etc.)
that must be provided by the runtime environment
4. Variant wheels: Generates hardware-variant wheels via variantlib for
hardware-specific distribution
5. OBS upload: Aggregates all variant wheels and a combined index JSON,
then uploads to Huawei OBS for hosting

### Does this PR introduce _any_ user-facing change?
Yes. Users will be able to install hardware-specific vllm-ascend wheels
from PyPI or the OBS variant index, eliminating the need to build from
source.

### How was this patch tested?
1. CI verification only — workflow syntax and job dependency logic
reviewed manually
2. Wheel build steps validated against existing Dockerfiles
(Dockerfile.buildwheel.a2/a3/310p)
3. auditwheel exclusion list verified against known Ascend runtime
shared libraries


- vLLM version: v0.17.0
- vLLM main:
4034c3d32e

---------

Signed-off-by: hfadzxy <starmoon_zhang@163.com>
Signed-off-by: YanZhicong <mryanzhicong@163.com>
Co-authored-by: YanZhicong <mryanzhicong@163.com>

.github/workflows/schedule_release_code_and_wheel.yml

@@ -33,6 +33,8 @@ on:
         type: choice
         options:
           - main
+          - v0.17.0rc1
+          - v0.16.0rc1
           - v0.15.0rc1
           - v0.14.0rc1
           - v0.13.0rc3
@@ -114,7 +116,7 @@ jobs:
       - name: Build wheel
         run: |
           ls
-          docker build -f ./.github/workflows/dockerfiles/Dockerfile.buildwheel \
+          docker build -f ./.github/workflows/dockerfiles/Dockerfile.buildwheel.a2 \
           --build-arg PY_VERSION=${{ matrix.python-version }} \
           -t wheel:v1 .
           docker run --rm \
@@ -167,14 +169,291 @@ jobs:
             auditwheel show "$wheel"
           done
 
+      - name: Generate variant wheels
+        env:
+          WHEEL_FILE: dist
+          PROJECT_TOML: .github/workflows/scripts/wheel/pyproject.toml
+          OUTPUT_DIR: dist/variants
+        run: |
+          pip install build git+https://github.com/wheelnext/variantlib.git --quiet
+          mkdir -p dist/variants
+          python3 .github/workflows/scripts/wheel/make_variant.py \
+            -c .github/workflows/scripts/wheel/config.json \
+            -l a2
+          echo "Generated variant wheels:"
+          ls dist/variants/
+
       - name: Archive wheel
         uses: actions/upload-artifact@v7
         with:
           name: vllm-ascend-${{ matrix.os }}-py${{ matrix.python-version }}-wheel
-          path: dist/*
+          path: dist/
 
       - name: Release
         if: ${{ github.event_name == 'push' || github.event_name == 'workflow_dispatch' }}
         run: |
           python3 -m pip install twine
-          python3 -m twine upload --verbose dist/* -u __token__ -p ${{ secrets.PYPI_TOKEN }}
+          python3 -m twine upload --verbose dist/*.whl -u __token__ -p ${{ secrets.PYPI_TOKEN }}
+
+
+  build_and_release_wheel_a3:
+    name: build and release wheel (A3)
+    strategy:
+      matrix:
+        os: [ubuntu-24.04, ubuntu-24.04-arm]
+        python-version: ["3.10", "3.11"]
+
+    runs-on: ${{ matrix.os }}
+    steps:
+      - name: checkout vllm-ascend
+        if: ${{ github.event_name != 'workflow_dispatch' }}
+        uses: actions/checkout@v6
+
+      - name: checkout vllm-ascend ${{ inputs.tag }}
+        if: ${{ github.event_name == 'workflow_dispatch' }}
+        uses: actions/checkout@v6
+        with:
+          ref: ${{ inputs.tag }}
+
+      - name: Free up disk space
+        uses: jlumbroso/free-disk-space@54081f138730dfa15788a46383842cd2f914a1be # v1.3.1
+        with:
+          tool-cache: true
+          docker-images: false
+
+      - name: Build wheel
+        run: |
+          ls
+          docker build -f ./.github/workflows/dockerfiles/Dockerfile.buildwheel.a3 \
+          --build-arg PY_VERSION=${{ matrix.python-version }} \
+          -t wheel-a3:v1 .
+          docker run --rm \
+          -u "$(id -u):$(id -g)" \
+          -v "$(pwd):/outpwd" \
+          wheel-a3:v1 \
+          bash -c "cp -r /workspace/vllm-ascend/dist /outpwd"
+          ls dist
+
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Repair wheels with auditwheel
+        run: |
+          python3 -m pip install auditwheel
+          python3 -m pip install patchelf
+          mkdir -p dist/repaired
+          for whl in dist/*.whl; do
+            auditwheel repair "$whl" -w dist/repaired/ \
+            --exclude libplatform.so \
+            --exclude libregister.so \
+            --exclude libge_common_base.so \
+            --exclude libc10.so \
+            --exclude libc_sec.so \
+            --exclude libnnopbase.so \
+            --exclude libprofapi.so \
+            --exclude libgraph_base.so \
+            --exclude libgraph.so \
+            --exclude libexe_graph.so \
+            --exclude "libascend*.so" \
+            --exclude "libtorch*.so" \
+            --exclude "libopapi.so" \
+            --exclude "liberror_manager.so" \
+            --exclude "libruntime.so" \
+            --exclude "libmmpa.so"
+
+          done
+          rm -f dist/*.whl
+          mv dist/repaired/*.whl dist/
+          rmdir dist/repaired
+          ls dist
+
+      - name: Verify automatic platform tags
+        run: |
+          cd dist
+          for wheel in *.whl; do
+            echo "verification file: $wheel"
+            auditwheel show "$wheel"
+          done
+
+      - name: Generate variant wheels
+        env:
+          WHEEL_FILE: dist
+          PROJECT_TOML: .github/workflows/scripts/wheel/pyproject.toml
+          OUTPUT_DIR: dist/variants
+        run: |
+          pip install build git+https://github.com/wheelnext/variantlib.git --quiet
+          mkdir -p dist/variants
+          python3 .github/workflows/scripts/wheel/make_variant.py \
+            -c .github/workflows/scripts/wheel/config.json \
+            -l a3
+          echo "Generated variant wheels:"
+          ls dist/variants/
+
+      - name: Archive wheel
+        uses: actions/upload-artifact@v7
+        with:
+          name: vllm-ascend-a3-${{ matrix.os }}-py${{ matrix.python-version }}-wheel
+          path: dist/
+
+
+  build_and_release_wheel_310p:
+    name: build and release wheel (310P)
+    strategy:
+      matrix:
+        os: [ubuntu-24.04, ubuntu-24.04-arm]
+        python-version: ["3.10", "3.11"]
+
+    runs-on: ${{ matrix.os }}
+    steps:
+      - name: checkout vllm-ascend
+        if: ${{ github.event_name != 'workflow_dispatch' }}
+        uses: actions/checkout@v6
+
+      - name: checkout vllm-ascend ${{ inputs.tag }}
+        if: ${{ github.event_name == 'workflow_dispatch' }}
+        uses: actions/checkout@v6
+        with:
+          ref: ${{ inputs.tag }}
+
+      - name: Free up disk space
+        uses: jlumbroso/free-disk-space@54081f138730dfa15788a46383842cd2f914a1be # v1.3.1
+        with:
+          tool-cache: true
+          docker-images: false
+
+      - name: Build wheel
+        run: |
+          ls
+          docker build -f ./.github/workflows/dockerfiles/Dockerfile.buildwheel.310p \
+          --build-arg PY_VERSION=${{ matrix.python-version }} \
+          -t wheel-310p:v1 .
+          docker run --rm \
+          -u "$(id -u):$(id -g)" \
+          -v "$(pwd):/outpwd" \
+          wheel-310p:v1 \
+          bash -c "cp -r /workspace/vllm-ascend/dist /outpwd"
+          ls dist
+
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Repair wheels with auditwheel
+        run: |
+          python3 -m pip install auditwheel
+          python3 -m pip install patchelf
+          mkdir -p dist/repaired
+          for whl in dist/*.whl; do
+            auditwheel repair "$whl" -w dist/repaired/ \
+            --exclude libplatform.so \
+            --exclude libregister.so \
+            --exclude libge_common_base.so \
+            --exclude libc10.so \
+            --exclude libc_sec.so \
+            --exclude libnnopbase.so \
+            --exclude libprofapi.so \
+            --exclude libgraph_base.so \
+            --exclude libgraph.so \
+            --exclude libexe_graph.so \
+            --exclude "libascend*.so" \
+            --exclude "libtorch*.so" \
+            --exclude "libopapi.so" \
+            --exclude "liberror_manager.so" \
+            --exclude "libruntime.so" \
+            --exclude "libmmpa.so"
+
+          done
+          rm -f dist/*.whl
+          mv dist/repaired/*.whl dist/
+          rmdir dist/repaired
+          ls dist
+
+      - name: Verify automatic platform tags
+        run: |
+          cd dist
+          for wheel in *.whl; do
+            echo "verification file: $wheel"
+            auditwheel show "$wheel"
+          done
+
+      - name: Generate variant wheels
+        env:
+          WHEEL_FILE: dist
+          PROJECT_TOML: .github/workflows/scripts/wheel/pyproject.toml
+          OUTPUT_DIR: dist/variants
+        run: |
+          pip install build git+https://github.com/wheelnext/variantlib.git --quiet
+          mkdir -p dist/variants
+          python3 .github/workflows/scripts/wheel/make_variant.py \
+            -c .github/workflows/scripts/wheel/config.json \
+            -l 310p
+          echo "Generated variant wheels:"
+          ls dist/variants/
+
+      - name: Archive wheel
+        uses: actions/upload-artifact@v7
+        with:
+          name: vllm-ascend-310p-${{ matrix.os }}-py${{ matrix.python-version }}-wheel
+          path: dist/
+
+
+  generate_and_upload_variant_index:
+    name: generate and upload variant index
+    needs: [build_and_release_wheel, build_and_release_wheel_a3, build_and_release_wheel_310p]
+    if: ${{ github.event_name == 'push' || github.event_name == 'workflow_dispatch' }}
+    runs-on: ubuntu-24.04
+    steps:
+      - name: Download all variant wheels
+        uses: actions/download-artifact@v4
+        with:
+          pattern: '*-wheel'
+          path: all-wheels/
+          merge-multiple: true
+
+      - name: Collect variant wheels
+        run: |
+          mkdir -p combined-variants
+          find all-wheels/ -path '*/variants/*.whl' -exec cp {} combined-variants/ \;
+          echo "Combined variant wheels:"
+          ls combined-variants/
+
+      - name: Set up Python
+        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
+        with:
+          python-version: "3.11"
+
+      - name: Generate combined variant index
+        run: |
+          pip install git+https://github.com/wheelnext/variantlib.git --quiet
+          variantlib generate-index-json -d combined-variants/
+          echo "Generated index files:"
+          ls combined-variants/
+
+      - name: Upload wheels and variant index to OBS
+        env:
+          OBS_ACCESS_KEY: ${{ secrets.OBS_ACCESS_KEY_ID }}
+          OBS_SECRET_KEY: ${{ secrets.OBS_SECRET_ACCESS_KEY }}
+        run: |
+          pip install esdk-obs-python --quiet
+          python3 - <<'EOF'
+          import os, glob
+          from obs import ObsClient
+          OBS_BUCKET = 'ascend-artifcat-packages'
+          OBS_PATH = 'pypi/packages/ascend/repos/pypi/variant/vllm-ascend'
+          client = ObsClient(
+              access_key_id=os.environ['OBS_ACCESS_KEY'],
+              secret_access_key=os.environ['OBS_SECRET_KEY'],
+              server='https://obs.cn-north-4.myhuaweicloud.com'
+          )
+          files = glob.glob('combined-variants/*.whl') + glob.glob('combined-variants/*.json')
+          for file in files:
+              filename = os.path.basename(file)
+              resp = client.putFile(OBS_BUCKET, f'{OBS_PATH}/{filename}', file)
+              if resp.status < 300:
+                  print(f'Uploaded: {filename}')
+              else:
+                  raise Exception(f'Failed to upload {filename}: {resp.errorMessage}')
+          EOF