Merge pull request ClickHouse#61725 from ClickHouse/ci_modify_ci_from_pr_body

CI: modify CI from PR body

Author: maxknv

.github/PULL_REQUEST_TEMPLATE.md

@@ -40,3 +40,33 @@ At a minimum, the following information should be added (but add more as needed)
 
 
 > Information about CI checks: https://clickhouse.com/docs/en/development/continuous-integration/
+
+---
+### Modify your CI run:
+##### NOTE:
+- if your merge the PR with modified CI you **MUST** know what you are doing.
+- modifiers can be applied only if set before CI starts
+- remove `!` to apply
+- return all `!` to restore defaults
+```
+!#ci_set_<SET_NAME> - to run only preconfigured set of tests, e.g.:
+!#ci_set_arm - to run only integration tests on ARM
+!#ci_set_integration - to run only integration tests on AMD
+!#ci_set_analyzer - to run only tests for analyzer
+NOTE: you can configure your own ci set
+```
+```
+!#job_<JOB NAME> - to run only specified job, e.g.:
+!#job_stateless_tests_release
+!#job_package_debug
+!#job_style_check
+!#job_integration_tests_asan
+```
+```
+!#batch_2 - to run only 2nd batch for all multi-batch jobs
+!#btach_1_2_3 - to run only 1, 2, 3rd batch for all multi-batch jobs
+```
+```
+!#no_merge_commit - to disable merge commit (no merge from master)
+!#do_not_test - to disable whole CI (except style check)
+```

.gitmessage

@@ -26,4 +26,4 @@
 
 ## To run only specified batches for multi-batch job(s)
 #batch_2
-#btach_1_2_3
+#batch_1_2_3

tests/ci/ci.py

@@ -1407,15 +1407,25 @@ def _concurrent_create_status(job: str, batch: int, num_batches: int) -> None:
     print("... CI report update - done")
 
 
-def _fetch_commit_tokens(message: str) -> List[str]:
-    pattern = r"#[\w-]+"
-    matches = [match[1:] for match in re.findall(pattern, message)]
+def _fetch_commit_tokens(message: str, pr_info: PRInfo) -> List[str]:
+    pattern = r"([^!]|^)#(\w+)"
+    matches = [match[-1] for match in re.findall(pattern, message)]
     res = [
         match
         for match in matches
         if match in Labels or match.startswith("job_") or match.startswith("batch_")
     ]
-    return res
+    print(f"CI modifyers from commit message: [{res}]")
+    res_2 = []
+    if pr_info.is_pr():
+        matches = [match[-1] for match in re.findall(pattern, pr_info.body)]
+        res_2 = [
+            match
+            for match in matches
+            if match in Labels or match.startswith("job_") or match.startswith("batch_")
+        ]
+        print(f"CI modifyers from PR body: [{res_2}]")
+    return list(set(res + res_2))
 
 
 def _upload_build_artifacts(
@@ -1701,8 +1711,7 @@ def main() -> int:
             message = args.commit_message or git_runner.run(
                 f"{GIT_PREFIX} log {pr_info.sha} --format=%B -n 1"
             )
-            tokens = _fetch_commit_tokens(message)
-            print(f"Commit message tokens: [{tokens}]")
+            tokens = _fetch_commit_tokens(message, pr_info)
             if Labels.NO_MERGE_COMMIT in tokens and CI:
                 git_runner.run(f"{GIT_PREFIX} checkout {pr_info.sha}")
                 git_ref = git_runner.run(f"{GIT_PREFIX} rev-parse HEAD")

tests/ci/pr_info.py

@@ -316,6 +316,9 @@ def is_release(self) -> bool:
     def is_release_branch(self) -> bool:
         return self.number == 0
 
+    def is_pr(self):
+        return self.event_type == EventType.PULL_REQUEST
+
     def is_scheduled(self):
         return self.event_type == EventType.SCHEDULE
 

tests/ci/style_check.py

@@ -8,9 +8,8 @@
 import sys
 from concurrent.futures import ProcessPoolExecutor
 from pathlib import Path
-from typing import List, Tuple, Union
+from typing import List, Tuple
 
-import magic
 from docker_images_helper import get_docker_image, pull_image
 from env_helper import CI, REPO_COPY, TEMP_PATH
 from git_helper import GIT_PREFIX, git_runner
@@ -96,32 +95,34 @@ def commit_push_staged(pr_info: PRInfo) -> None:
     git_runner(push_cmd)
 
 
-def is_python(file: Union[Path, str]) -> bool:
+def is_python(file: str) -> bool:
     """returns if the changed file in the repository is python script"""
     # WARNING: python-magic v2:0.4.24-2 is used in ubuntu 22.04,
     # and `Support os.PathLike values in magic.from_file` is only from 0.4.25
-    try:
-        return bool(
-            magic.from_file(os.path.join(REPO_COPY, file), mime=True)
-            == "text/x-script.python"
-        )
-    except IsADirectoryError:
-        # Process submodules w/o errors
-        return False
-
-
-def is_shell(file: Union[Path, str]) -> bool:
+    # try:
+    #     return bool(
+    #         magic.from_file(os.path.join(REPO_COPY, file), mime=True)
+    #         == "text/x-script.python"
+    #     )
+    # except IsADirectoryError:
+    #     # Process submodules w/o errors
+    #     return False
+    return file.endswith(".py")
+
+
+def is_shell(file: str) -> bool:
     """returns if the changed file in the repository is shell script"""
     # WARNING: python-magic v2:0.4.24-2 is used in ubuntu 22.04,
     # and `Support os.PathLike values in magic.from_file` is only from 0.4.25
-    try:
-        return bool(
-            magic.from_file(os.path.join(REPO_COPY, file), mime=True)
-            == "text/x-shellscript"
-        )
-    except IsADirectoryError:
-        # Process submodules w/o errors
-        return False
+    # try:
+    #     return bool(
+    #         magic.from_file(os.path.join(REPO_COPY, file), mime=True)
+    #         == "text/x-shellscript"
+    #     )
+    # except IsADirectoryError:
+    #     # Process submodules w/o errors
+    #     return False
+    return file.endswith(".sh")
 
 
 def main():