[WIP] feat: Filter copr_build jobs based on paths and files changed

packit_service/events/github/push.py

@@ -1,11 +1,15 @@
 # Copyright Contributors to the Packit project.
 # SPDX-License-Identifier: MIT
 
-from packit_service.service.db_project_events import AddBranchPushEventToDb
+from packit_service.service.db_project_events import AddBranchPushEventToDb, CommitInfo
 
 from .abstract import GithubEvent
 
 
+class GithubCommitInfo(CommitInfo):
+    pass
+
+
 class Commit(AddBranchPushEventToDb, GithubEvent):
     def __init__(
         self,
@@ -15,6 +19,7 @@ def __init__(
         project_url: str,
         commit_sha: str,
         commit_sha_before: str,
+        commits: list[GithubCommitInfo],
     ):
         super().__init__(project_url=project_url)
         self.repo_namespace = repo_namespace
@@ -23,6 +28,7 @@ def __init__(
         self.commit_sha = commit_sha
         self.commit_sha_before = commit_sha_before
         self.identifier = git_ref
+        self.commits = commits
 
     @classmethod
     def event_type(cls) -> str:

packit_service/events/gitlab/push.py

@@ -1,13 +1,15 @@
 # Copyright Contributors to the Packit project.
 # SPDX-License-Identifier: MIT
 
-from packit_service.service.db_project_events import (
-    AddBranchPushEventToDb,
-)
+from packit_service.service.db_project_events import AddBranchPushEventToDb, CommitInfo
 
 from .abstract import GitlabEvent
 
 
+class GitlabCommitInfo(CommitInfo):
+    pass
+
+
 class Commit(AddBranchPushEventToDb, GitlabEvent):
     def __init__(
         self,
@@ -17,6 +19,7 @@ def __init__(
         project_url: str,
         commit_sha: str,
         commit_sha_before: str,
+        commits: list[GitlabCommitInfo],
     ):
         super().__init__(project_url=project_url)
         self.repo_namespace = repo_namespace
@@ -25,6 +28,7 @@ def __init__(
         self.commit_sha = commit_sha
         self.commit_sha_before = commit_sha_before
         self.identifier = git_ref
+        self.commits = commits
 
     @classmethod
     def event_type(cls) -> str:

packit_service/service/db_project_events.py

@@ -5,7 +5,7 @@
 This file contains helper classes for events.
 """
 
-from typing import Optional
+from typing import Optional, TypedDict
 
 from ogr.abstract import GitProject
 
@@ -61,12 +61,22 @@ def get_dict(self, default_dict: Optional[dict] = None) -> dict:
         return result
 
 
+class CommitInfo(TypedDict, total=False):
+    id: str
+    title: str
+    message: str
+    added: list[str]
+    modified: list[str]
+    removed: list[str]
+
+
 class AddBranchPushEventToDb:
     git_ref: str
     repo_namespace: str
     repo_name: str
     project_url: str
     commit_sha: str
+    commits: list[CommitInfo]
     _branch: GitBranchModel = None
     _event: ProjectEventModel = None
 

packit_service/worker/checker/copr.py

@@ -2,11 +2,14 @@
 # SPDX-License-Identifier: MIT
 
 import logging
+from pathlib import Path
+
+from packit.config import JobConfigTriggerType
 
 from packit_service.constants import (
     INTERNAL_TF_BUILDS_AND_TESTS_NOT_ALLOWED,
 )
-from packit_service.events import gitlab
+from packit_service.events import github, gitlab
 from packit_service.worker.checker.abstract import (
     ActorChecker,
     Checker,
@@ -138,3 +141,53 @@ def _pre_check(self) -> bool:
                 )
                 return False
         return True
+
+
+class AreFilesChanged(Checker, GetCoprBuildJobHelperForIdMixin, ConfigFromEventMixin):
+    """
+    Check if any files under the current package's `paths` field is changed.
+    If not, then just skip the current copr build job.
+    """
+
+    def get_files_changed(self) -> list[Path]:
+        """
+        Get the list of files changed in the current commit or the current pullrequest
+        """
+        # Get the changes object
+        if self.job_config.trigger == JobConfigTriggerType.pull_request:
+            pr_event = self.data.to_event()
+            if not isinstance(pr_event, (github.pr.Action, gitlab.mr.Action)):
+                # TODO: What about comments?
+                raise NotImplementedError()
+            # TODO: How to handle PRs?
+            raise NotImplementedError()
+        if self.job_config.trigger == JobConfigTriggerType.commit:
+            push_event = self.data.to_event()
+            if not isinstance(push_event, (github.push.Commit, gitlab.push.Commit)):
+                raise NotImplementedError()
+            files = set()
+            for commit in push_event.commits:
+                files |= set(commit["modified"])
+                files |= set(commit["added"])
+            # TODO: Check what the path is relative to
+            return [Path(file) for file in files]
+        raise NotImplementedError(f"Trigger not supported: {self.job_config.trigger}")
+
+    def pre_check(self) -> bool:
+        if self.job_config.trigger == JobConfigTriggerType.release:
+            # For releases we don't do any checks
+            return True
+        # FIXME: This is probably unnecessary
+        package_config = self.package_config.get_package_config_for(self.job_config)
+        # The paths that we need to check for files changed
+        paths = package_config["paths"]
+        # Early check if the git root was included, in which case we don't need to
+        # check the files changed
+        # TODO: refine this check when gitignore-like patterns are supported
+        if "./" in paths:
+            return True
+        for changed_file in self.get_files_changed():
+            # Check if any of the files changed are under the paths that are being tracked
+            if any(changed_file.is_relative_to(p) for p in paths):
+                return True
+        return False

packit_service/worker/handlers/copr.py

@@ -41,6 +41,7 @@
 )
 from packit_service.worker.checker.abstract import Checker
 from packit_service.worker.checker.copr import (
+    AreFilesChanged,
     AreOwnerAndProjectMatchingJob,
     BuildNotAlreadyStarted,
     CanActorRunTestsJob,
@@ -116,6 +117,7 @@ def get_checkers() -> tuple[type[Checker], ...]:
             IsJobConfigTriggerMatching,
             IsGitForgeProjectAndEventOk,
             CanActorRunTestsJob,
+            AreFilesChanged,
         )
 
     def run(self) -> TaskResults:

packit_service/worker/parser.py

@@ -9,7 +9,7 @@
 from dataclasses import dataclass
 from datetime import datetime, timezone
 from os import getenv
-from typing import Any, Callable, ClassVar, Optional, Union
+from typing import Any, Callable, ClassVar, Optional, Union, cast
 
 from ogr.parsing import RepoUrl, parse_git_repo
 from packit.config import JobConfigTriggerType
@@ -64,8 +64,9 @@ class _GitlabCommonData:
     project_url: str
     parsed_url: Optional[RepoUrl]
     ref: str
-    head_commit: dict
+    head_commit: gitlab.push.GitlabCommitInfo
     commit_sha_before: str
+    commits: list[gitlab.push.GitlabCommitInfo]
 
     @property
     def commit_sha(self) -> str:
@@ -420,7 +421,9 @@ def get_gitlab_push_common_data(event) -> _GitlabCommonData:
         before = event.get("before")
         checkout_sha = event.get("checkout_sha")
         actor = event.get("user_username")
-        commits = event.get("commits", [])
+        commits = [
+            cast(gitlab.push.GitlabCommitInfo, commit) for commit in event.get("commits", [])
+        ]
         number_of_commits = event.get("total_commits_count")
 
         if not Parser.is_gitlab_push_a_create_event(event):
@@ -455,6 +458,7 @@ def get_gitlab_push_common_data(event) -> _GitlabCommonData:
             ref=ref,
             head_commit=head_commit,
             commit_sha_before=before,
+            commits=commits,
         )
 
     @staticmethod
@@ -515,6 +519,7 @@ def parse_gitlab_push_event(event) -> Optional[gitlab.push.Commit]:
             project_url=data.project_url,
             commit_sha=data.commit_sha,
             commit_sha_before=data.commit_sha_before,
+            commits=data.commits,
         )
 
     @staticmethod
@@ -561,13 +566,16 @@ def parse_github_push_event(event) -> Optional[github.push.Commit]:
 
         repo_url = nested_get(event, "repository", "html_url")
 
+        commits = [cast(github.push.GithubCommitInfo, commit) for commit in event.get("commits")]
+
         return github.push.Commit(
             repo_namespace=repo_namespace,
             repo_name=repo_name,
             git_ref=ref,
             project_url=repo_url,
             commit_sha=head_commit,
             commit_sha_before=before,
+            commits=commits,
         )
 
     @staticmethod