[CI] optimize lint term (#5986)

### What this PR does / why we need it?
This patch purpose to optimize the lint check term. The main idea is to
reduce unnecessary installation time.
1. The installation of vllm is not must, only append the path of vllm
src to the `PATHONPATH` is effective
2. This installation of `requirements-dev.txt` is not must, we have a
pre-built image `quay.io/ascend-ci/vllm-ascend:lint` with all the
requirements installed in advance.
**NOTE**: the conditions for triggering image builds are: 1).Daily
scheduled build; 2) Build when requirements are modified; 3) Manual
build. This ensures that the dependencies in our image are up-to-date to
the greatest extent possible.
3. The `mypy` was separated from the `pre-commit` hook for performance
reasons; we found that integrating `mypy` into the `pre-commit` hook
resulted in poor performance.
4. Reduce the CPU core consumption from 16 -> 8

### Does this PR introduce _any_ user-facing change?
The end-to-end lint time was optimized from 20min/per PR to 8min/per PR
### How was this patch tested?

- vLLM version: v0.13.0
- vLLM main:
2c24bc6996

---------

Signed-off-by: wangli <wangli858794774@gmail.com>
This commit is contained in:
Li Wang
2026-01-22 15:46:59 +08:00
committed by GitHub
parent 9bba0a2a68
commit 484e7c59dc
10 changed files with 196 additions and 77 deletions

View File

@@ -12,19 +12,16 @@ permissions:
jobs:
pre-commit:
runs-on: linux-amd64-cpu-16-hk
runs-on: linux-amd64-cpu-8-hk
container:
# Build it from https://github.com/nv-action/vllm-benchmarks/blob/main/Dockerfile
image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/vllm-ascend:nightly-cpu
env:
GOPROXY: https://goproxy.io,direct
GITHUB_WORKSPACE: /__w/vllm-ascend/vllm-ascend
image: quay.io/ascend-ci/vllm-ascend:lint
steps:
- name: Checkout vllm-project/vllm-ascend repo
uses: actions/checkout@v6
# With problem matchers in a container, the output of $GITHUB_WORKSPACE and ${{ github.workspace }} are different.
# So we will just copy it into the path ${{ github.workspace }}. see https://github.com/actions/runner/issues/2058
# So we will just copy it into a temp path. see https://github.com/actions/runner/issues/2058
- name: cp problem matchers
run: |
cp .github/workflows/matchers/actionlint.json "$RUNNER_TEMP/actionlint.json"
@@ -41,14 +38,21 @@ jobs:
repository: vllm-project/vllm
path: ./vllm-empty
ref: ${{ inputs.vllm }}
- name: Install vllm
working-directory: vllm-empty
run: |
VLLM_TARGET_DEVICE=empty python3 -m pip install . --extra-index https://download.pytorch.org/whl/cpu/
- name: Install vllm-ascend dev
- uses: dorny/paths-filter@v3
id: filter
with:
filters: |
lint_tracker:
- 'requirements.txt'
- 'requirements-dev.txt'
- 'requirements-lint.txt'
- name: Install vllm-ascend dev (conditional)
if: steps.filter.outputs.lint_tracker == 'true'
run: |
git config --global --add safe.directory /__w/vllm-ascend/vllm-ascend
pip install -r requirements-dev.txt --extra-index-url https://download.pytorch.org/whl/cpu
- name: Run pre-commit
env:
PRE_COMMIT_COLOR: always
@@ -56,4 +60,17 @@ jobs:
TERM: xterm-256color
SHELLCHECK_OPTS: "--exclude=SC2046,SC2006,SC2086" # Exclude SC2046, SC2006, SC2086 for actionlint
run: |
git config --global --add safe.directory /__w/vllm-ascend/vllm-ascend
pre-commit run --all-files --hook-stage manual --show-diff-on-failure
- name: Run mypy
run: |
PYTHONPATH="$PYTHONPATH:$(pwd)/vllm-empty"
export PYTHONPATH
git config --global --add safe.directory /__w/vllm-ascend/vllm-ascend
# Run mypy for Python 3.10, 3.11, 3.12 manually
# Note: We are now separating mypy from pre-commit hooks for performance reasons.
for python_version in "3.10" "3.11" "3.12"; do
echo "============================"
tools/mypy.sh 1 "$python_version"
echo "============================"
done