feat: add custom validation

commitizen/commands/check.py

@@ -78,25 +78,30 @@ def __call__(self) -> None:
         """Validate if commit messages follows the conventional pattern.
 
         Raises:
-            InvalidCommitMessageError: if the commit provided not follows the conventional pattern
+            InvalidCommitMessageError: if the commit provided does not follow the conventional pattern
         """
         commits = self._get_commits()
         if not commits:
             raise NoCommitsFoundError(f"No commit found with range: '{self.rev_range}'")
 
         pattern = re.compile(self.cz.schema_pattern())
-        invalid_msgs_content = "\n".join(
-            f'commit "{commit.rev}": "{commit.message}"'
+        invalid_commits = [
+            (commit, check.errors)
             for commit in commits
-            if not self._validate_commit_message(commit.message, pattern)
-        )
-        if invalid_msgs_content:
-            # TODO: capitalize the first letter of the error message for consistency in v5
+            if not (
+                check := self.cz.validate_commit_message(
+                    commit_msg=commit.message,
+                    pattern=pattern,
+                    allow_abort=self.allow_abort,
+                    allowed_prefixes=self.allowed_prefixes,
+                    max_msg_length=self.max_msg_length,
+                )
+            ).is_valid
+        ]
+
+        if invalid_commits:
             raise InvalidCommitMessageError(
-                "commit validation: failed!\n"
-                "please enter a commit message in the commitizen format.\n"
-                f"{invalid_msgs_content}\n"
-                f"pattern: {pattern.pattern}"
+                self.cz.format_exception_message(invalid_commits)
             )
         out.success("Commit validation: successful!")
 

commitizen/cz/base.py

@@ -1,8 +1,9 @@
 from __future__ import annotations
 
+import re
 from abc import ABCMeta, abstractmethod
 from collections.abc import Iterable, Mapping
-from typing import Any, Callable, Protocol
+from typing import Any, Callable, NamedTuple, Protocol
 
 from jinja2 import BaseLoader, PackageLoader
 from prompt_toolkit.styles import Style, merge_styles
@@ -24,6 +25,11 @@ def __call__(
     ) -> dict[str, Any]: ...
 
 
+class ValidationResult(NamedTuple):
+    is_valid: bool
+    errors: list
+
+
 class BaseCommitizen(metaclass=ABCMeta):
     bump_pattern: str | None = None
     bump_map: dict[str, str] | None = None
@@ -41,7 +47,7 @@ class BaseCommitizen(metaclass=ABCMeta):
         ("disabled", "fg:#858585 italic"),
     ]
 
-    # The whole subject will be parsed as message by default
+    # The whole subject will be parsed as a message by default
     # This allows supporting changelog for any rule system.
     # It can be modified per rule
     commit_parser: str | None = r"(?P<message>.*)"
@@ -96,6 +102,54 @@ def schema_pattern(self) -> str:
         """Regex matching the schema used for message validation."""
         raise NotImplementedError("Not Implemented yet")
 
+    def validate_commit_message(
+        self,
+        *,
+        commit_msg: str,
+        pattern: re.Pattern[str] | None,
+        allow_abort: bool,
+        allowed_prefixes: list[str],
+        max_msg_length: int,
+    ) -> ValidationResult:
+        """Validate commit message against the pattern."""
+        if not commit_msg:
+            return ValidationResult(allow_abort, [])
+
+        if pattern is None:
+            return ValidationResult(True, [])
+
+        if any(map(commit_msg.startswith, allowed_prefixes)):
+            return ValidationResult(True, [])
+
+        if max_msg_length:
+            msg_len = len(commit_msg.partition("\n")[0].strip())
+            if msg_len > max_msg_length:
+                return ValidationResult(
+                    False,
+                    [f"message is too long: {msg_len} > {max_msg_length}"],
+                )
+
+        return ValidationResult(
+            bool(pattern.match(commit_msg)),
+            [f"pattern: {pattern.pattern}"],
+        )
+
+    def format_exception_message(
+        self, invalid_commits: list[tuple[git.GitCommit, list]]
+    ) -> str:
+        """Format commit errors."""
+        displayed_msgs_content = "\n".join(
+            [
+                f'commit "{commit.rev}": "{commit.message}\n"' + "\n".join(errors)
+                for commit, errors in invalid_commits
+            ]
+        )
+        return (
+            "commit validation: failed!\n"
+            "please enter a commit message in the commitizen format.\n"
+            f"{displayed_msgs_content}"
+        )
+
     def info(self) -> str:
         """Information about the standardized commit message."""
         raise NotImplementedError("Not Implemented yet")

docs/customization.md

@@ -312,6 +312,73 @@ cz -n cz_strange bump
 
 [convcomms]: https://github.com/commitizen-tools/commitizen/blob/master/commitizen/cz/conventional_commits/conventional_commits.py
 
+### Custom commit validation and error message
+
+The commit message validation can be customized by overriding the `validate_commit_message` and `format_error_message`
+methods from `BaseCommitizen`. This allows for a more detailed feedback to the user where the error originates from.
+
+```python
+import re
+
+from commitizen.cz.base import BaseCommitizen
+from commitizen import git
+
+
+class CustomValidationCz(BaseCommitizen):
+    def validate_commit_message(
+        self,
+        *,
+        commit_msg: str,
+        pattern: str | None,
+        allow_abort: bool,
+        allowed_prefixes: list[str],
+        max_msg_length: int,
+    ) -> tuple[bool, list]:
+        """Validate commit message against the pattern."""
+        if not commit_msg:
+            return allow_abort, [] if allow_abort else [f"commit message is empty"]
+
+        if pattern is None:
+            return True, []
+
+        if any(map(commit_msg.startswith, allowed_prefixes)):
+            return True, []
+        if max_msg_length:
+            msg_len = len(commit_msg.partition("\n")[0].strip())
+            if msg_len > max_msg_length:
+                return False, [
+                    f"commit message is too long. Max length is {max_msg_length}"
+                ]
+        pattern_match = re.match(pattern, commit_msg)
+        if pattern_match:
+            return True, []
+        else:
+            # Perform additional validation of the commit message format
+            # and add custom error messages as needed
+            return False, ["commit message does not match the pattern"]
+
+    def format_exception_message(
+        self, ill_formated_commits: list[tuple[git.GitCommit, list]]
+    ) -> str:
+        """Format commit errors."""
+        displayed_msgs_content = "\n".join(
+            [
+                (
+                    f'commit "{commit.rev}": "{commit.message}"'
+                    f"errors:\n"
+                    "\n".join((f"- {error}" for error in errors))
+                )
+                for commit, errors in ill_formated_commits
+            ]
+        )
+        return (
+            "commit validation: failed!\n"
+            "please enter a commit message in the commitizen format.\n"
+            f"{displayed_msgs_content}\n"
+            f"pattern: {self.schema_pattern()}"
+        )
+```
+
 ### Custom changelog generator
 
 The changelog generator should just work in a very basic manner without touching anything.

tests/commands/test_check_command.py

@@ -452,3 +452,44 @@ def test_check_command_with_message_length_limit_exceeded(config, mocker: MockFi
     with pytest.raises(InvalidCommitMessageError):
         check_cmd()
         error_mock.assert_called_once()
+
+
+@pytest.mark.usefixtures("use_cz_custom_validator")
+def test_check_command_with_custom_validator_succeed(mocker: MockFixture, capsys):
+    testargs = [
+        "cz",
+        "--name",
+        "cz_custom_validator",
+        "check",
+        "--commit-msg-file",
+        "some_file",
+    ]
+    mocker.patch.object(sys, "argv", testargs)
+    mocker.patch(
+        "commitizen.commands.check.open",
+        mocker.mock_open(read_data="ABC-123: add commitizen pre-commit hook"),
+    )
+    cli.main()
+    out, _ = capsys.readouterr()
+    assert "Commit validation: successful!" in out
+
+
+@pytest.mark.usefixtures("use_cz_custom_validator")
+def test_check_command_with_custom_validator_failed(mocker: MockFixture):
+    testargs = [
+        "cz",
+        "--name",
+        "cz_custom_validator",
+        "check",
+        "--commit-msg-file",
+        "some_file",
+    ]
+    mocker.patch.object(sys, "argv", testargs)
+    mocker.patch(
+        "commitizen.commands.check.open",
+        mocker.mock_open(read_data="ABC-123 add commitizen pre-commit hook"),
+    )
+    with pytest.raises(InvalidCommitMessageError) as excinfo:
+        cli.main()
+    assert "commit validation: failed!" in str(excinfo.value)
+    assert "pattern: " in str(excinfo.value)

tests/conftest.py

@@ -3,20 +3,21 @@
 import os
 import re
 import tempfile
-from collections.abc import Iterator, Mapping
+from collections.abc import Iterable, Iterator, Mapping
 from pathlib import Path
+from typing import Any
 
 import pytest
 from pytest_mock import MockerFixture
 
-from commitizen import cmd, defaults
+from commitizen import cmd, defaults, git
 from commitizen.changelog_formats import (
     ChangelogFormat,
     get_changelog_format,
 )
 from commitizen.config import BaseConfig
 from commitizen.cz import registry
-from commitizen.cz.base import BaseCommitizen
+from commitizen.cz.base import BaseCommitizen, ValidationResult
 from commitizen.question import CzQuestion
 from tests.utils import create_file_and_commit
 
@@ -223,7 +224,7 @@ def use_cz_semver(mocker):
 
 
 class MockPlugin(BaseCommitizen):
-    def questions(self) -> list[CzQuestion]:
+    def questions(self) -> Iterable[CzQuestion]:
         return []
 
     def message(self, answers: Mapping) -> str:
@@ -237,6 +238,82 @@ def mock_plugin(mocker: MockerFixture, config: BaseConfig) -> BaseCommitizen:
     return mock
 
 
+class ValidationCz(BaseCommitizen):
+    def questions(self) -> Iterable[CzQuestion]:
+        return [
+            {"type": "input", "name": "commit", "message": "Initial commit:\n"},
+            {"type": "input", "name": "issue_nb", "message": "ABC-123"},
+        ]
+
+    def message(self, answers: Mapping[str, Any]) -> str:
+        return f"{answers['issue_nb']}: {answers['commit']}"
+
+    def schema(self) -> str:
+        return "<issue_nb>: <commit>"
+
+    def schema_pattern(self) -> str:
+        return r"^(?P<issue_nb>[A-Z]{3}-\d+): (?P<commit>.*)$"
+
+    def validate_commit_message(
+        self,
+        *,
+        commit_msg: str,
+        pattern: re.Pattern[str] | None,
+        allow_abort: bool,
+        allowed_prefixes: list[str],
+        max_msg_length: int,
+    ) -> ValidationResult:
+        """Validate commit message against the pattern."""
+        if not commit_msg:
+            return ValidationResult(
+                allow_abort, [] if allow_abort else ["commit message is empty"]
+            )
+
+        if pattern is None:
+            return ValidationResult(True, [])
+
+        if any(map(commit_msg.startswith, allowed_prefixes)):
+            return ValidationResult(True, [])
+
+        if max_msg_length:
+            msg_len = len(commit_msg.partition("\n")[0].strip())
+            if msg_len > max_msg_length:
+                return ValidationResult(
+                    False,
+                    [f"message is too long: {msg_len} > {max_msg_length}"],
+                )
+
+        return ValidationResult(
+            bool(pattern.match(commit_msg)), [f"pattern: {pattern.pattern}"]
+        )
+
+    def format_exception_message(
+        self, invalid_commits: list[tuple[git.GitCommit, list]]
+    ) -> str:
+        """Format commit errors."""
+        displayed_msgs_content = "\n".join(
+            [
+                (
+                    f'commit "{commit.rev}": "{commit.message}"\nerrors:\n\n'.join(
+                        f"- {error}" for error in errors
+                    )
+                )
+                for (commit, errors) in invalid_commits
+            ]
+        )
+        return (
+            "commit validation: failed!\n"
+            "please enter a commit message in the commitizen format.\n"
+            f"{displayed_msgs_content}"
+        )
+
+
+@pytest.fixture
+def use_cz_custom_validator(mocker):
+    new_cz = {**registry, "cz_custom_validator": ValidationCz}
+    mocker.patch.dict("commitizen.cz.registry", new_cz)
+
+
 SUPPORTED_FORMATS = ("markdown", "textile", "asciidoc", "restructuredtext")