[NPU] Adapt pr-gate for pr-test workflow & workflows refresh (#13567)

.github/labeler.yml

@@ -66,6 +66,13 @@ amd:
       - '**/*amd*'
       - '**/*rocm*'
 
+# NPU specific
+npu:
+  - changed-files:
+    - any-glob-to-any-file:
+      - '**/*npu*'
+      - '**/*ascend*'
+
 # DeepSeek specific
 deepseek:
   - changed-files:

.github/workflows/pr-test-npu.yml

@@ -1,22 +1,10 @@
-name: PR Test (Ascend NPU)
+name: PR Test (NPU)
 
 on:
   push:
     branches: [ main ]
-    paths:
-      - "python/**"
-      - "!python/sglang/multimodal_gen/**"
-      - "scripts/ci/**"
-      - "test/**"
-      - ".github/workflows/pr-test-npu.yml"
   pull_request:
     branches: [ main ]
-    paths:
-      - "python/**"
-      - "!python/sglang/multimodal_gen/**"
-      - "scripts/ci/**"
-      - "test/**"
-      - ".github/workflows/pr-test-npu.yml"
   workflow_dispatch:
 
 concurrency:
@@ -24,8 +12,39 @@ concurrency:
   cancel-in-progress: true
 
 jobs:
-  per-commit-1-ascend-npu:
+
-    if: github.event_name != 'pull_request' || contains(github.event.pull_request.labels.*.name, 'run-ci')
+  # ==================== PR Gate ==================== #
+  pr-gate:
+    uses: ./.github/workflows/pr-gate.yml
+    secrets: inherit
+  # ================================================= #
+
+  # ==================== Check Changes ==================== #
+  check-changes:
+    needs: [pr-gate]
+    runs-on: ubuntu-latest
+    outputs:
+      main_package: ${{ steps.filter.outputs.main_package }}
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Detect file changes
+        id: filter
+        uses: dorny/paths-filter@v3
+        with:
+          filters: |
+            main_package:
+              - "python/sglang/!(multimodal_gen)/**"
+              - "python/*.toml"
+              - "scripts/ci/npu_ci_install_dependency.sh"
+              - "test/srt/ascend/**"
+              - ".github/workflows/pr-test-npu.yml"
+  # ======================================================= #
+
+  per-commit-1-npu-a2:
+    needs: [check-changes]
+    if: needs.check-changes.outputs.main_package == 'true'
     runs-on: linux-arm64-npu-1
     container:
       image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.2.rc1-910b-ubuntu22.04-py3.11
@@ -60,10 +79,11 @@ jobs:
         run: |
           export PATH="/usr/local/Ascend/8.3.RC1/compiler/bishengir/bin:${PATH}"
           cd test/srt
-          python3 run_suite.py --suite per-commit-1-ascend-npu
+          python3 run_suite.py --suite per-commit-1-npu-a2
 
-  per-commit-2-ascend-npu:
+  per-commit-2-npu-a2:
-    if: github.event_name != 'pull_request' || contains(github.event.pull_request.labels.*.name, 'run-ci')
+    needs: [check-changes]
+    if: needs.check-changes.outputs.main_package == 'true'
     runs-on: linux-arm64-npu-2
     strategy:
       fail-fast: true
@@ -102,10 +122,11 @@ jobs:
         run: |
           export PATH="/usr/local/Ascend/8.3.RC1/compiler/bishengir/bin:${PATH}"
           cd test/srt
-          python3 run_suite.py --suite per-commit-2-ascend-npu --auto-partition-id ${{ matrix.part }} --auto-partition-size 3
+          python3 run_suite.py --suite per-commit-2-npu-a2 --auto-partition-id ${{ matrix.part }} --auto-partition-size 3
 
-  per-commit-4-ascend-npu:
+  per-commit-4-npu-a2:
-    if: github.event_name != 'pull_request' || contains(github.event.pull_request.labels.*.name, 'run-ci')
+    needs: [check-changes]
+    if: needs.check-changes.outputs.main_package == 'true'
     runs-on: linux-arm64-npu-4
     container:
       image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.2.rc1-910b-ubuntu22.04-py3.11
@@ -140,10 +161,11 @@ jobs:
         run: |
           export PATH="/usr/local/Ascend/8.3.RC1/compiler/bishengir/bin:${PATH}"
           cd test/srt
-          python3 run_suite.py --suite per-commit-4-ascend-npu --timeout-per-file 3600
+          python3 run_suite.py --suite per-commit-4-npu-a2 --timeout-per-file 3600
 
-  per-commit-16-ascend-a3:
+  per-commit-16-npu-a3:
-    if: github.event_name != 'pull_request' || contains(github.event.pull_request.labels.*.name, 'run-ci')
+    needs: [check-changes]
+    if: needs.check-changes.outputs.main_package == 'true'
     runs-on: linux-aarch64-a3-16
     strategy:
       fail-fast: true
@@ -182,4 +204,4 @@ jobs:
         run: |
           export PATH="/usr/local/Ascend/8.3.RC1/compiler/bishengir/bin:${PATH}"
           cd test/srt
-          python3 run_suite.py --suite per-commit-16-ascend-a3 --timeout-per-file 3600 --auto-partition-id ${{ matrix.part }} --auto-partition-size 2
+          python3 run_suite.py --suite per-commit-16-npu-a3 --timeout-per-file 3600 --auto-partition-id ${{ matrix.part }} --auto-partition-size 2

.github/workflows/release-docker-npu-nightly.yml

@@ -1,4 +1,4 @@
-name: Release Docker Images Nightly (Ascend NPU)
+name: Release Docker Images Nightly (NPU)
 on:
   pull_request:
     branches:

.github/workflows/release-docker-npu.yml

@@ -1,18 +1,16 @@
-name: Release Docker Images (Ascend NPU)
+name: Release Docker Images (NPU)
 on:
   push:
-    tags:
-      - "*" # Trigger on all tags and filterred by pep440 later
     tags-ignore:
       - "gateway-*" # Exclude gateway/router tags
       - "router-*" # Exclude router tags
-  workflow_dispatch:
   pull_request:
     branches:
       - main
     paths:
       - ".github/workflows/release-docker-npu.yml"
       - "docker/npu.Dockerfile"
+  workflow_dispatch:
 
 jobs:
   build:

docker/npu.Dockerfile

@@ -45,7 +45,6 @@ RUN apt-get update -y && apt upgrade -y && apt-get install -y \
     libssl-dev \
     pkg-config \
     ca-certificates \
-    protobuf-compiler \
     && rm -rf /var/cache/apt/* \
     && rm -rf /var/lib/apt/lists/* \
     && update-ca-certificates \
@@ -54,18 +53,11 @@ RUN apt-get update -y && apt upgrade -y && apt-get install -y \
 ENV LANG=en_US.UTF-8
 ENV LANGUAGE=en_US:en
 ENV LC_ALL=en_US.UTF-8
-ENV PATH="/root/.cargo/bin:${PATH}"
 
 # Install dependencies
 # TODO: install from pypi released memfabric
 RUN pip install $MEMFABRIC_URL --no-cache-dir
 
-RUN pip install setuptools-rust wheel build --no-cache-dir
-
-# install rustup from rustup.rs
-RUN curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y \
-    && rustc --version && cargo --version && protoc --version
-
 # Install vLLM
 RUN git clone --depth 1 https://github.com/vllm-project/vllm.git --branch $VLLM_TAG && \
     (cd vllm && VLLM_TARGET_DEVICE="empty" pip install -v . --no-cache-dir) && rm -rf vllm
@@ -79,12 +71,14 @@ RUN pip install torch==$PYTORCH_VERSION torchvision==$TORCHVISION_VERSION --inde
 # Install SGLang
 RUN git clone https://github.com/sgl-project/sglang --branch $SGLANG_TAG && \
     (cd sglang/python && rm -rf pyproject.toml && mv pyproject_other.toml pyproject.toml && pip install -v .[srt_npu] --no-cache-dir) && \
-    (cd sglang/sgl-router && python -m build && pip install --force-reinstall dist/*.whl) && \
     rm -rf sglang
 
+# Install SGLang Model Gateway
+RUN pip install sglang-router --no-cache-dir
+
 # Install Deep-ep
 # pin wheel to 0.45.1 ref: https://github.com/pypa/wheel/issues/662
-RUN pip install wheel==0.45.1 && git clone  --branch $SGLANG_KERNEL_NPU_TAG https://github.com/sgl-project/sgl-kernel-npu.git \
+RUN pip install wheel==0.45.1 && git clone --branch $SGLANG_KERNEL_NPU_TAG https://github.com/sgl-project/sgl-kernel-npu.git \
     && export LD_LIBRARY_PATH=${ASCEND_CANN_PATH}/latest/runtime/lib64/stub:$LD_LIBRARY_PATH && \
     source ${ASCEND_CANN_PATH}/set_env.sh && \
     cd sgl-kernel-npu && \

test/srt/run_suite.py

@@ -514,25 +514,26 @@ suite_xpu = {
 }
 
 # Add Ascend NPU tests
+# TODO: Set accurate estimate time
 # NOTE: please sort the test cases alphabetically by the test file name
 suite_ascend = {
-    "per-commit-1-ascend-npu": [
+    "per-commit-1-npu-a2": [
         TestFile("ascend/test_ascend_graph_tp1_bf16.py", 400),
-        TestFile("ascend/test_ascend_tp1_bf16.py", 400),
         TestFile("ascend/test_ascend_hicache_mha.py", 400),
         TestFile("ascend/test_ascend_sampling_backend.py", 400),
+        TestFile("ascend/test_ascend_tp1_bf16.py", 400),
     ],
-    "per-commit-2-ascend-npu": [
+    "per-commit-2-npu-a2": [
         TestFile("ascend/test_ascend_graph_tp2_bf16.py", 400),
         TestFile("ascend/test_ascend_mla_fia_w8a8int8.py", 400),
         TestFile("ascend/test_ascend_tp2_bf16.py", 400),
         TestFile("ascend/test_ascend_tp2_fia_bf16.py", 400),
     ],
-    "per-commit-4-ascend-npu": [
+    "per-commit-4-npu-a2": [
         TestFile("ascend/test_ascend_mla_w8a8int8.py", 400),
         TestFile("ascend/test_ascend_tp4_bf16.py", 400),
     ],
-    "per-commit-16-ascend-a3": [
+    "per-commit-16-npu-a3": [
         TestFile("ascend/test_ascend_deepep.py", 400),
         TestFile("ascend/test_ascend_deepseek_mtp.py", 400),
     ],