feat: Introduce tag_regex option with smart default

robertschweizer · robertschweizer · commit 8eea9ea61454 · 2023-03-23T16:15:56.000+01:00
Closes #519 CLI flag name: --tag-regex Heavily inspired by #537, but extends it with a smart default value to exclude non-release tags. This was suggested in #519 (comment)
diff --git a/commitizen/changelog.py b/commitizen/changelog.py
@@ -80,8 +80,6 @@ def generate_tree_from_commits(
                 "date": current_tag_date,
                 "changes": changes,
             }
-            # TODO: Check if tag matches the version pattern, otherwise skip it.
-            # This in order to prevent tags that are not versions.
             current_tag_name = commit_tag.name
             current_tag_date = commit_tag.date
             changes = defaultdict(list)
diff --git a/commitizen/cli.py b/commitizen/cli.py
@@ -251,6 +251,13 @@
                             "If not set, it will generate changelog from the start"
                         ),
                     },
+                    {
+                        "name": "--tag-regex",
+                        "help": (
+                            "regex match for tags represented "
+                            "within the changelog. default: '.*'"
+                        ),
+                    },
                 ],
             },
             {
diff --git a/commitizen/commands/changelog.py b/commitizen/commands/changelog.py
@@ -1,4 +1,5 @@
 import os.path
+import re
 from difflib import SequenceMatcher
 from operator import itemgetter
 from typing import Callable, Dict, List, Optional
@@ -15,7 +16,7 @@
     NotAllowed,
 )
 from commitizen.git import GitTag, smart_open
-from commitizen.tags import tag_from_version
+from commitizen.tags import make_tag_pattern, tag_from_version
 
 
 class Changelog:
@@ -51,6 +52,10 @@ def __init__(self, config: BaseConfig, args):
         self.tag_format: str = args.get("tag_format") or self.config.settings.get(
             "tag_format", DEFAULT_SETTINGS["tag_format"]
         )
+        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'.
@@ -124,7 +129,7 @@ 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 = []
 
diff --git a/commitizen/git.py b/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:
diff --git a/commitizen/tags.py b/commitizen/tags.py
@@ -1,7 +1,8 @@
+import re
 from string import Template
 from typing import Union
 
-from packaging.version import Version
+from packaging.version import VERSION_PATTERN, Version
 
 
 def tag_from_version(version: Union[Version, str], tag_format: str) -> str:
@@ -29,3 +30,23 @@ def tag_from_version(version: Union[Version, str], tag_format: str) -> str:
     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
diff --git a/docs/changelog.md b/docs/changelog.md
@@ -161,6 +161,28 @@ cz changelog --start-rev="v0.2.0"
 changelog_start_rev = "v0.2.0"
 ```
 
+### `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:
diff --git a/docs/config.md b/docs/config.md
@@ -8,6 +8,7 @@
 | `version`                  | `str`  | `None`                      | Current version. Example: "0.1.2"                                                                                                                                                                                                                    |
 | `version_files`            | `list` | `[ ]`                       | Files were the version will be updated. A pattern to match a line, can also be specified, separated by `:` [See more][version_files]                                                                                                                 |
 | `tag_format`               | `str`  | `$version`                  | Format for the git tag, useful for old projects, that use a convention like `"v1.2.1"`. [See more][tag_format]                                                                                                                                       |
+| `tag_regex`                | `str`  | Based on `tag_format`       | Tags must match this to be included in the changelog (e.g. `"([0-9.])*"` to exclude pre-releases). [See more][tag_regex]                                                                                                                             |
 | `update_changelog_on_bump` | `bool` | `false`                     | Create changelog when running `cz bump`                                                                                                                                                                                                              |
 | `gpg_sign`                 | `bool` | `false`                     | Use gpg signed tags instead of lightweight tags.                                                                                                                                                                                                     |
 | `annotated_tag`            | `bool` | `false`                     | Use annotated tags instead of lightweight tags. [See difference][annotated-tags-vs-lightweight]                                                                                                                                                      |
@@ -114,6 +115,7 @@ commitizen:
 
 [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
diff --git a/tests/commands/test_bump_command.py b/tests/commands/test_bump_command.py
@@ -531,6 +531,24 @@ def test_bump_with_changelog_config(mocker: MockFixture, changelog_path, config_
     assert "0.2.0" in out
 
 
+@pytest.mark.usefixtures("tmp_commitizen_project")
+def test_bump_with_changelog_excludes_custom_tags(mocker: MockFixture, changelog_path):
+    create_file_and_commit("feat(user): new file")
+    git.tag("custom-tag")
+    create_file_and_commit("feat(user): Another new file")
+    testargs = ["cz", "bump", "--yes", "--changelog"]
+    mocker.patch.object(sys, "argv", testargs)
+    cli.main()
+    tag_exists = git.tag_exist("0.2.0")
+    assert tag_exists is True
+
+    with open(changelog_path, "r") as f:
+        out = f.read()
+    assert out.startswith("#")
+    assert "## 0.2.0" in out
+    assert "custom-tag" not in out
+
+
 def test_prevent_prerelease_when_no_increment_detected(
     mocker: MockFixture, capsys, tmp_commitizen_project
 ):
diff --git a/tests/commands/test_changelog_command.py b/tests/commands/test_changelog_command.py
@@ -1,5 +1,7 @@
 import sys
 from datetime import datetime
+from typing import List
+from unittest.mock import patch
 
 import pytest
 from pytest_mock import MockFixture
@@ -968,3 +970,53 @@ def test_empty_commit_list(mocker):
     mocker.patch.object(sys, "argv", testargs)
     with pytest.raises(NoCommitsFoundError):
         cli.main()
+
+
+@pytest.mark.parametrize(
+    "config_file, expected_versions",
+    [
+        pytest.param("", ["Unreleased"], id="v-prefix-not-configured"),
+        pytest.param(
+            'tag_format = "v$version"',
+            ["v1.1.0", "v1.1.0-beta", "v1.0.0"],
+            id="v-prefix-configured-as-tag-format",
+        ),
+        pytest.param(
+            'tag_format = "v$version"\n' + 'tag_regex = ".*"',
+            ["v1.1.0", "custom-tag", "v1.1.0-beta", "v1.0.0"],
+            id="tag-regex-matches-all-tags",
+        ),
+        pytest.param(
+            'tag_format = "v$version"\n' + r'tag_regex = "v[0-9\\.]*"',
+            ["v1.1.0", "v1.0.0"],
+            id="tag-regex-excludes-pre-releases",
+        ),
+    ],
+)
+def test_changelog_tag_regex(
+    config_path, changelog_path, config_file: str, expected_versions: List[str]
+):
+    with open(config_path, "a") as f:
+        f.write(config_file)
+
+    # Create 4 tags with one valid feature each
+    create_file_and_commit("feat: initial")
+    git.tag("v1.0.0")
+    create_file_and_commit("feat: add 1")
+    git.tag("v1.1.0-beta")
+    create_file_and_commit("feat: add 2")
+    git.tag("custom-tag")
+    create_file_and_commit("feat: add 3")
+    git.tag("v1.1.0")
+
+    # call CLI
+    with patch.object(sys, "argv", ["cz", "changelog"]):
+        cli.main()
+
+    # open CLI output
+    with open(changelog_path, "r") as f:
+        out = f.read()
+
+    headings = [line for line in out.splitlines() if line.startswith("## ")]
+    changelog_versions = [heading[3:].split()[0] for heading in headings]
+    assert changelog_versions == expected_versions
diff --git a/tests/test_git.py b/tests/test_git.py
@@ -1,12 +1,14 @@
 import inspect
 import os
+import re
 import shutil
 from typing import List, Optional
 
 import pytest
 from pytest_mock import MockFixture
 
 from commitizen import cmd, exceptions, git
+from commitizen.tags import make_tag_pattern
 from tests.utils import FakeCommand, create_file_and_commit
 
 
@@ -28,7 +30,7 @@ def test_get_tags(mocker: MockFixture):
     )
     mocker.patch("commitizen.cmd.run", return_value=FakeCommand(out=tag_str))
 
-    git_tags = git.get_tags()
+    git_tags = git.get_tags(pattern=re.compile(r"v[0-9\.]+"))
     latest_git_tag = git_tags[0]
     assert latest_git_tag.rev == "333"
     assert latest_git_tag.name == "v1.0.0"
@@ -37,7 +39,60 @@ def test_get_tags(mocker: MockFixture):
     mocker.patch(
         "commitizen.cmd.run", return_value=FakeCommand(out="", err="No tag available")
     )
-    assert git.get_tags() == []
+    assert git.get_tags(pattern=re.compile(r"v[0-9\.]+")) == []
+
+
+@pytest.mark.parametrize(
+    "pattern, expected_tags",
+    [
+        pytest.param(
+            make_tag_pattern(tag_format="$version"),
+            [],  # No versions with normal 1.2.3 pattern
+            id="default-tag-format",
+        ),
+        pytest.param(
+            make_tag_pattern(tag_format="$major-$minor-$patch$prerelease"),
+            ["1-0-0", "1-0-0alpha2"],
+            id="tag-format-with-hyphens",
+        ),
+        pytest.param(
+            r"[0-9]+\-[0-9]+\-[0-9]+",
+            ["1-0-0"],
+            id="tag-regex-with-hyphens-that-excludes-alpha",
+        ),
+        pytest.param(
+            make_tag_pattern(tag_format="v$version"),
+            ["v0.5.0", "v0.0.1-pre"],
+            id="tag-format-with-v-prefix",
+        ),
+        pytest.param(
+            make_tag_pattern(tag_format="custom-prefix-$version"),
+            ["custom-prefix-0.0.1"],
+            id="tag-format-with-custom-prefix",
+        ),
+        pytest.param(
+            ".*",
+            ["1-0-0", "1-0-0alpha2", "v0.5.0", "v0.0.1-pre", "custom-prefix-0.0.1"],
+            id="custom-tag-regex-to-include-all-tags",
+        ),
+    ],
+)
+def test_get_tags_filtering(
+    mocker: MockFixture, pattern: str, expected_tags: List[str]
+):
+    tag_str = (
+        "1-0-0---inner_delimiter---333---inner_delimiter---2020-01-20---inner_delimiter---\n"
+        "1-0-0alpha2---inner_delimiter---333---inner_delimiter---2020-01-20---inner_delimiter---\n"
+        "v0.5.0---inner_delimiter---222---inner_delimiter---2020-01-17---inner_delimiter---\n"
+        "v0.0.1-pre---inner_delimiter---111---inner_delimiter---2020-01-17---inner_delimiter---\n"
+        "custom-prefix-0.0.1---inner_delimiter---111---inner_delimiter---2020-01-17---inner_delimiter---\n"
+        "custom-non-release-tag"
+    )
+    mocker.patch("commitizen.cmd.run", return_value=FakeCommand(out=tag_str))
+
+    git_tags = git.get_tags(pattern=re.compile(pattern, flags=re.VERBOSE))
+    actual_name_list = [t.name for t in git_tags]
+    assert actual_name_list == expected_tags
 
 
 def test_get_tag_names(mocker: MockFixture):
diff --git a/tests/test_tags.py b/tests/test_tags.py

Original file line number	Diff line number	Diff line change
`@@ -80,8 +80,6 @@ def generate_tree_from_commits(`
`80`	`80`	`"date": current_tag_date,`
`81`	`81`	`"changes": changes,`
`82`	`82`	`}`
`83`		`- # TODO: Check if tag matches the version pattern, otherwise skip it.`
`84`		`- # This in order to prevent tags that are not versions.`
`85`	`83`	`current_tag_name = commit_tag.name`
`86`	`84`	`current_tag_date = commit_tag.date`
`87`	`85`	`changes = defaultdict(list)`