Merge pull request #61125 from ClickHouse/ci_fast_style

CI: make style check fast

docker/test/style/Dockerfile

@@ -60,5 +60,4 @@ RUN arch=${TARGETARCH:-amd64} \
 
 
 COPY run.sh /
-COPY process_style_check_result.py /
 CMD ["/bin/bash", "/run.sh"]

tests/ci/style_check.py

@@ -1,5 +1,6 @@
 #!/usr/bin/env python3
 import argparse
+from concurrent.futures import ProcessPoolExecutor
 import csv
 import logging
 import os
@@ -119,7 +120,7 @@ def checkout_last_ref(pr_info: PRInfo) -> None:
 def main():
     logging.basicConfig(level=logging.INFO)
     logging.getLogger("git_helper").setLevel(logging.DEBUG)
-    args = parse_args()
+    # args = parse_args()
 
     stopwatch = Stopwatch()
 
@@ -127,28 +128,46 @@ def main():
     temp_path = Path(TEMP_PATH)
     temp_path.mkdir(parents=True, exist_ok=True)
 
-    pr_info = PRInfo()
+    # pr_info = PRInfo()
 
     IMAGE_NAME = "clickhouse/style-test"
     image = pull_image(get_docker_image(IMAGE_NAME))
-    cmd = (
+    cmd_1 = (
         f"docker run -u $(id -u ${{USER}}):$(id -g ${{USER}}) --cap-add=SYS_PTRACE "
         f"--volume={repo_path}:/ClickHouse --volume={temp_path}:/test_output "
-        f"{image}"
+        f"--entrypoint= -w/ClickHouse/utils/check-style "
+        f"{image} ./check_cpp_docs.sh"
     )
+    cmd_2 = (
+        f"docker run -u $(id -u ${{USER}}):$(id -g ${{USER}}) --cap-add=SYS_PTRACE "
+        f"--volume={repo_path}:/ClickHouse --volume={temp_path}:/test_output "
+        f"--entrypoint= -w/ClickHouse/utils/check-style "
+        f"{image} ./check_py.sh"
+    )
+    logging.info("Is going to run the command: %s", cmd_1)
+    logging.info("Is going to run the command: %s", cmd_2)
 
-    if args.push:
-        checkout_head(pr_info)
+    with ProcessPoolExecutor(max_workers=2) as executor:
+        # Submit commands for execution in parallel
+        future1 = executor.submit(subprocess.run, cmd_1, shell=True)
+        future2 = executor.submit(subprocess.run, cmd_2, shell=True)
+        # Wait for both commands to complete
+        _ = future1.result()
+        _ = future2.result()
+
+    # if args.push:
+    #     checkout_head(pr_info)
 
-    logging.info("Is going to run the command: %s", cmd)
     subprocess.check_call(
-        cmd,
+        f"python3 ../../utils/check-style/process_style_check_result.py --in-results-dir {temp_path} "
+        f"--out-results-file {temp_path}/test_results.tsv --out-status-file {temp_path}/check_status.tsv || "
+        f'echo -e "failure\tCannot parse results" > {temp_path}/check_status.tsv',
         shell=True,
     )
 
-    if args.push:
-        commit_push_staged(pr_info)
-        checkout_last_ref(pr_info)
+    # if args.push:
+    #     commit_push_staged(pr_info)
+    #     checkout_last_ref(pr_info)
 
     state, description, test_results, additional_files = process_result(temp_path)
 

utils/check-style/check-pylint

@@ -0,0 +1,15 @@
+#!/usr/bin/env bash
+
+
+LC_ALL="en_US.UTF-8"
+ROOT_PATH=$(git rev-parse --show-toplevel)
+
+function xargs-pylint {
+  # $1 is number maximum arguments per pylint process
+  sort | awk '$2=="text/x-script.python" {print $1}' | \
+  xargs -P "$(nproc)" -n "$1" pylint --rcfile="$ROOT_PATH/pyproject.toml" --persistent=no --score=n
+}
+
+find "$ROOT_PATH/tests" -maxdepth 2 -type f -exec file -F' ' --mime-type {} + | xargs-pylint 50
+# Beware, there lambdas are checked. All of them contain `app`, and it causes brain-cucumber-zalgo
+find "$ROOT_PATH/tests/ci" -mindepth 2 -type f -exec file -F' ' --mime-type {} + | xargs-pylint 1

utils/check-style/check-style

@@ -152,16 +152,6 @@ find $ROOT_PATH/{src,base,programs,utils} -name '*.xml' |
     grep -vP $EXCLUDE_DIRS |
     xargs xmllint --noout --nonet
 
-function xargs-pylint {
-  # $1 is number maximum arguments per pylint process
-  sort | awk '$2=="text/x-script.python" {print $1}' | \
-  xargs -P "$(nproc)" -n "$1" pylint --rcfile="$ROOT_PATH/pyproject.toml" --persistent=no --score=n
-}
-
-find "$ROOT_PATH/tests" -maxdepth 2 -type f -exec file -F' ' --mime-type {} + | xargs-pylint 50
-# Beware, there lambdas are checked. All of them contain `app`, and it causes brain-cucumber-zalgo
-find "$ROOT_PATH/tests/ci" -mindepth 2 -type f -exec file -F' ' --mime-type {} + | xargs-pylint 1
-
 find $ROOT_PATH -not -path $ROOT_PATH'/contrib*' \( -name '*.yaml' -or -name '*.yml' \) -type f |
     grep -vP $EXCLUDE_DIRS |
     xargs yamllint --config-file=$ROOT_PATH/.yamllint

utils/check-style/check_cpp_docs.sh

@@ -0,0 +1,34 @@
+#!/bin/bash
+
+# yaml check is not the best one
+
+cd /ClickHouse/utils/check-style || echo -e "failure\tRepo not found" > /test_output/check_status.tsv
+
+# FIXME: 30 sec to wait
+# echo "Check duplicates" | ts
+# ./check-duplicate-includes.sh |& tee /test_output/duplicate_includes_output.txt
+
+echo "Check style" | ts
+./check-style -n              |& tee /test_output/style_output.txt
+echo "Check typos" | ts
+./check-typos                 |& tee /test_output/typos_output.txt
+echo "Check docs spelling" | ts
+./check-doc-aspell            |& tee /test_output/docs_spelling_output.txt
+echo "Check whitespaces" | ts
+./check-whitespaces -n        |& tee /test_output/whitespaces_output.txt
+echo "Check workflows" | ts
+./check-workflows             |& tee /test_output/workflows_output.txt
+echo "Check submodules" | ts
+./check-submodules            |& tee /test_output/submodules_output.txt
+echo "Check style. Done" | ts
+
+# FIXME: 6 min to wait
+# echo "Check shell scripts with shellcheck" | ts
+# ./shellcheck-run.sh           |& tee /test_output/shellcheck_output.txt
+
+
+# FIXME: move out
+# /process_style_check_result.py || echo -e "failure\tCannot parse results" > /test_output/check_status.tsv
+# echo "Check help for changelog generator works" | ts
+# cd ../changelog || exit 1
+# ./changelog.py -h 2>/dev/null 1>&2

utils/check-style/check_py.sh

@@ -0,0 +1,17 @@
+#!/bin/bash
+
+# yaml check is not the best one
+
+cd /ClickHouse/utils/check-style || echo -e "failure\tRepo not found" > /test_output/check_status.tsv
+
+# FIXME: 1 min to wait + head checkout
+# echo "Check python formatting with black" | ts
+# ./check-black -n              |& tee /test_output/black_output.txt
+
+echo "Check pylint" | ts
+./check-pylint -n               |& tee /test_output/pylint_output.txt
+echo "Check pylint. Done" | ts
+
+echo "Check python type hinting with mypy" | ts
+./check-mypy -n               |& tee /test_output/mypy_output.txt
+echo "Check python type hinting with mypy. Done" | ts

utils/check-style/process_style_check_result.py

@@ -13,10 +13,11 @@ def process_result(result_folder):
     description = ""
     test_results = []
     checks = (
-        "duplicate includes",
-        "shellcheck",
+        #"duplicate includes",
+        #"shellcheck",
         "style",
-        "black",
+        "pylint",
+        #"black",
         "mypy",
         "typos",
         "whitespaces",