feat: Introduce tag_regex option with smart default

commitizen/bump.py

@@ -217,42 +217,6 @@ def _version_to_regex(version: str) -> str:
     return version.replace(".", r"\.").replace("+", r"\+")
 
 
-def normalize_tag(
-    version: Union[VersionProtocol, str],
-    tag_format: Optional[str] = None,
-    version_type_cls: Optional[Type[VersionProtocol]] = None,
-) -> str:
-    """The tag and the software version might be different.
-
-    That's why this function exists.
-
-    Example:
-    | tag | version (PEP 0440) |
-    | --- | ------- |
-    | v0.9.0 | 0.9.0 |
-    | ver1.0.0 | 1.0.0 |
-    | ver1.0.0.a0 | 1.0.0a0 |
-    """
-    if version_type_cls is None:
-        version_type_cls = Version
-    if isinstance(version, str):
-        version = version_type_cls(version)
-
-    if not tag_format:
-        return str(version)
-
-    major, minor, patch = version.release
-    prerelease = ""
-    # version.pre is needed for mypy check
-    if version.is_prerelease and version.pre:
-        prerelease = f"{version.pre[0]}{version.pre[1]}"
-
-    t = Template(tag_format)
-    return t.safe_substitute(
-        version=version, major=major, minor=minor, patch=patch, prerelease=prerelease
-    )
-
-
 def create_commit_message(
     current_version: Union[Version, str],
     new_version: Union[Version, str],

commitizen/changelog.py

@@ -37,9 +37,9 @@
 from packaging.version import InvalidVersion, Version
 
 from commitizen import defaults
-from commitizen.bump import normalize_tag
 from commitizen.exceptions import InvalidConfigurationError, NoCommitsFoundError
 from commitizen.git import GitCommit, GitTag
+from commitizen.tags import tag_from_version
 
 if sys.version_info >= (3, 8):
     from commitizen.version_types import VersionProtocol
@@ -341,13 +341,13 @@ def get_oldest_and_newest_rev(
     except ValueError:
         newest = version
 
-    newest_tag = normalize_tag(
+    newest_tag = tag_from_version(
         newest, tag_format=tag_format, version_type_cls=version_type_cls
     )
 
     oldest_tag = None
     if oldest:
-        oldest_tag = normalize_tag(
+        oldest_tag = tag_from_version(
             oldest, tag_format=tag_format, version_type_cls=version_type_cls
         )
 

commitizen/cli.py

@@ -1,8 +1,8 @@
 import argparse
 import logging
 import sys
-from pathlib import Path
 from functools import partial
+from pathlib import Path
 from types import TracebackType
 from typing import List
 
@@ -274,6 +274,13 @@
                             "If not set, it will include prereleases in the changelog"
                         ),
                     },
+                    {
+                        "name": "--tag-regex",
+                        "help": (
+                            "regex match for tags represented "
+                            "within the changelog. default: '.*'"
+                        ),
+                    },
                 ],
             },
             {

commitizen/commands/bump.py

@@ -22,6 +22,7 @@
     NoVersionSpecifiedError,
 )
 from commitizen.providers import get_provider
+from commitizen.tags import tag_from_version
 
 logger = getLogger("commitizen")
 
@@ -161,7 +162,7 @@ def __call__(self):  # noqa: C901
                     f"--major-version-zero is meaningless for current version {current_version}"
                 )
 
-        current_tag_version: str = bump.normalize_tag(
+        current_tag_version: str = tag_from_version(
             current_version,
             tag_format=tag_format,
             version_type_cls=self.version_type,
@@ -223,7 +224,7 @@ def __call__(self):  # noqa: C901
                 version_type_cls=self.version_type,
             )
 
-        new_tag_version = bump.normalize_tag(
+        new_tag_version = tag_from_version(
             new_version,
             tag_format=tag_format,
             version_type_cls=self.version_type,

commitizen/commands/changelog.py

@@ -1,12 +1,14 @@
 import os.path
+import re
 from difflib import SequenceMatcher
 from operator import itemgetter
 from typing import Callable, Dict, List, Optional
 
 from packaging.version import parse
 
-from commitizen import bump, changelog, defaults, factory, git, out, version_types
+from commitizen import changelog, defaults, factory, git, out, version_types
 from commitizen.config import BaseConfig
+from commitizen.defaults import DEFAULT_SETTINGS
 from commitizen.exceptions import (
     DryRunExit,
     NoCommitsFoundError,
@@ -16,6 +18,7 @@
     NotAllowed,
 )
 from commitizen.git import GitTag, smart_open
+from commitizen.tags import make_tag_pattern, tag_from_version
 
 
 class Changelog:
@@ -55,8 +58,8 @@ def __init__(self, config: BaseConfig, args):
             or defaults.change_type_order
         )
         self.rev_range = args.get("rev_range")
-        self.tag_format = args.get("tag_format") or self.config.settings.get(
-            "tag_format"
+        self.tag_format: str = args.get("tag_format") or self.config.settings.get(
+            "tag_format", DEFAULT_SETTINGS["tag_format"]
         )
         self.merge_prerelease = args.get(
             "merge_prerelease"
@@ -65,6 +68,11 @@ def __init__(self, config: BaseConfig, args):
         version_type = self.config.settings.get("version_type")
         self.version_type = version_type and version_types.VERSION_TYPES[version_type]
 
+        tag_regex = args.get("tag_regex") or self.config.settings.get("tag_regex")
+        if not tag_regex:
+            tag_regex = make_tag_pattern(self.tag_format)
+        self.tag_pattern = re.compile(str(tag_regex), re.VERBOSE | re.IGNORECASE)
+
     def _find_incremental_rev(self, latest_version: str, tags: List[GitTag]) -> str:
         """Try to find the 'start_rev'.
 
@@ -138,7 +146,7 @@ def __call__(self):
         # Don't continue if no `file_name` specified.
         assert self.file_name
 
-        tags = git.get_tags()
+        tags = git.get_tags(pattern=self.tag_pattern)
         if not tags:
             tags = []
 
@@ -148,7 +156,7 @@ def __call__(self):
             changelog_meta = changelog.get_metadata(self.file_name)
             latest_version = changelog_meta.get("latest_version")
             if latest_version:
-                latest_tag_version: str = bump.normalize_tag(
+                latest_tag_version: str = tag_from_version(
                     latest_version,
                     tag_format=self.tag_format,
                     version_type_cls=self.version_type,

commitizen/commands/init.py

@@ -10,7 +10,7 @@
 from commitizen.__version__ import __version__
 from commitizen.config import BaseConfig, JsonConfig, TomlConfig, YAMLConfig
 from commitizen.cz import registry
-from commitizen.defaults import config_files
+from commitizen.defaults import DEFAULT_SETTINGS, config_files
 from commitizen.exceptions import InitFailedError, NoAnswersError
 from commitizen.git import get_latest_tag_name, get_tag_names, smart_open
 from commitizen.version_types import VERSION_TYPES
@@ -203,14 +203,15 @@ def _ask_tag_format(self, latest_tag) -> str:
                 f'Is "{tag_format}" the correct tag format?', style=self.cz.style
             ).unsafe_ask()
 
+        default_format = DEFAULT_SETTINGS["tag_format"]
         if not is_correct_format:
             tag_format = questionary.text(
-                'Please enter the correct version format: (default: "$version")',
+                f'Please enter the correct version format: (default: "{default_format}")',
                 style=self.cz.style,
             ).unsafe_ask()
 
             if not tag_format:
-                tag_format = "$version"
+                tag_format = default_format
         return tag_format
 
     def _ask_version_provider(self) -> str:

commitizen/defaults.py

@@ -35,7 +35,7 @@ class Settings(TypedDict, total=False):
     version: Optional[str]
     version_files: List[str]
     version_provider: Optional[str]
-    tag_format: Optional[str]
+    tag_format: str
     bump_message: Optional[str]
     allow_abort: bool
     changelog_file: str
@@ -68,7 +68,7 @@ class Settings(TypedDict, total=False):
     "version": None,
     "version_files": [],
     "version_provider": "commitizen",
-    "tag_format": None,  # example v$version
+    "tag_format": "$version",  # example v$version
     "bump_message": None,  # bumped v$current_version to $new_version
     "allow_abort": False,
     "changelog_file": "CHANGELOG.md",

commitizen/git.py

@@ -1,4 +1,5 @@
 import os
+import re
 from enum import Enum
 from os import linesep
 from pathlib import Path
@@ -140,7 +141,7 @@ def get_filenames_in_commit(git_reference: str = ""):
         raise GitCommandError(c.err)
 
 
-def get_tags(dateformat: str = "%Y-%m-%d") -> List[GitTag]:
+def get_tags(dateformat: str = "%Y-%m-%d", *, pattern: re.Pattern) -> List[GitTag]:
     inner_delimiter = "---inner_delimiter---"
     formatter = (
         f'"%(refname:lstrip=2){inner_delimiter}'
@@ -163,7 +164,9 @@ def get_tags(dateformat: str = "%Y-%m-%d") -> List[GitTag]:
         for line in c.out.split("\n")[:-1]
     ]
 
-    return git_tags
+    filtered_git_tags = [t for t in git_tags if pattern.fullmatch(t.name)]
+
+    return filtered_git_tags
 
 
 def tag_exist(tag: str) -> bool:

commitizen/tags.py

@@ -0,0 +1,65 @@
+import re
+import sys
+from string import Template
+from typing import Any, Optional, Type, Union
+
+from packaging.version import VERSION_PATTERN, Version
+
+if sys.version_info >= (3, 8):
+    from commitizen.version_types import VersionProtocol
+else:
+    # workaround mypy issue for 3.7 python
+    VersionProtocol = Any
+
+
+def tag_from_version(
+    version: Union[VersionProtocol, str],
+    tag_format: str,
+    version_type_cls: Optional[Type[VersionProtocol]] = None,
+) -> str:
+    """The tag and the software version might be different.
+
+    That's why this function exists.
+
+    Example:
+    | tag | version (PEP 0440) |
+    | --- | ------- |
+    | v0.9.0 | 0.9.0 |
+    | ver1.0.0 | 1.0.0 |
+    | ver1.0.0.a0 | 1.0.0a0 |
+    """
+    if version_type_cls is None:
+        version_type_cls = Version
+    if isinstance(version, str):
+        version = version_type_cls(version)
+
+    major, minor, patch = version.release
+    prerelease = ""
+    # version.pre is needed for mypy check
+    if version.is_prerelease and version.pre:
+        prerelease = f"{version.pre[0]}{version.pre[1]}"
+
+    t = Template(tag_format)
+    return t.safe_substitute(
+        version=version, major=major, minor=minor, patch=patch, prerelease=prerelease
+    )
+
+
+def make_tag_pattern(tag_format: str) -> str:
+    """Make regex pattern to match all tags created by tag_format."""
+    escaped_format = re.escape(tag_format)
+    escaped_format = re.sub(
+        r"\\\$(version|major|minor|patch|prerelease)", r"$\1", escaped_format
+    )
+    # pre-release part of VERSION_PATTERN
+    pre_release_pattern = r"([-_\.]?(a|b|c|rc|alpha|beta|pre|preview)([-_\.]?[0-9]+)?)?"
+    filter_regex = Template(escaped_format).safe_substitute(
+        # VERSION_PATTERN allows the v prefix, but we'd rather have users configure it
+        # explicitly.
+        version=VERSION_PATTERN.lstrip("\n v?"),
+        major="[0-9]+",
+        minor="[0-9]+",
+        patch="[0-9]+",
+        prerelease=pre_release_pattern,
+    )
+    return filter_regex

docs/bump.md

@@ -327,7 +327,7 @@ In your `pyproject.toml` or `.cz.toml`
 tag_format = "v$major.$minor.$patch$prerelease"
 ```
 
-The variables must be preceded by a `$` sign.
+The variables must be preceded by a `$` sign. Default is `$version`.
 
 Supported variables:
 

docs/changelog.md

@@ -186,6 +186,28 @@ cz changelog --merge-prerelease
 changelog_merge_prerelease = true
 ```
 
+### `tag-regex`
+
+This value can be set in the `toml` file with the key `tag_regex` under `tools.commitizen`.
+
+`tag_regex` is the regex pattern that selects tags to include in the changelog.
+By default, the changelog will capture all git tags matching the `tag_format`, including pre-releases.
+
+Example use-cases:
+
+- Exclude pre-releases from the changelog
+- Include existing tags that do not follow `tag_format` in the changelog
+
+```bash
+cz changelog --tag-regex="[0-9]*\\.[0-9]*\\.[0-9]"
+```
+
+```toml
+[tools.commitizen]
+# ...
+tag_regex = "[0-9]*\\.[0-9]*\\.[0-9]"
+```
+
 ## Hooks
 
 Supported hook methods:

docs/config.md

@@ -38,10 +38,18 @@ Version provider used to read and write version [Read more](#version-providers)
 
 Type: `str`
 
-Default: `None`
+Default: `$version`
 
 Format for the git tag, useful for old projects, that use a convention like `"v1.2.1"`. [Read more][tag_format]
 
+### `tag_regex`
+
+Type: `str`
+
+Default: Based on `tag_format`
+
+Tags must match this to be included in the changelog (e.g. `"([0-9.])*"` to exclude pre-releases). [Read more][tag_regex]
+
 ### `update_changelog_on_bump`
 
 Type: `bool`
@@ -339,6 +347,7 @@ setup(
 
 [version_files]: bump.md#version_files
 [tag_format]: bump.md#tag_format
+[tag_regex]: changelog.md#tag_regex
 [bump_message]: bump.md#bump_message
 [major-version-zero]: bump.md#-major-version-zero
 [prerelease-offset]: bump.md#-prerelease_offset

poetry.toml

@@ -0,0 +1,2 @@
+[virtualenvs]
+in-project = true