Merge pull request #17 from brainelectronics/feature/parse-changelog-…

…to-json
brainelectronics · Oct 26, 2022 · 2bdec36 · 2bdec36
2 parents 2663266 + 0f3af96
commit 2bdec36
Show file tree

Hide file tree

Showing 8 changed files with 628 additions and 21 deletions.
diff --git a/README.md b/README.md
@@ -21,10 +21,15 @@ Create version info files based on the latest changelog entry.
     - [Available default template files](#available-default-template-files)
         - [C header file](#c-header-file)
         - [Python package file](#python-package-file)
+        - [JSON output](#json-output)
+            - [Console](#console)
+            - [File](#file)
 - [Advanced](#advanced)
     - [Custom regular expressions](#custom-regular-expressions)
     - [Custom template file](#custom-template-file)
     - [Additional version info content](#additional-version-info-content)
+- [Contributing](#contributing)
+    - [Unittests](#unittests)
 - [Credits](#credits)
 
 <!-- /MarkdownTOC -->
@@ -105,6 +110,36 @@ __version__ = '.'.join(__version_info__)
 
 ```
 
+#### JSON output
+
+The additional, optional argument `--pretty` will output the JSON data with an
+indentation of 4 in order to provide the data in an easy to read format.
+
+##### Console
+
+```bash
+changelog2version \
+    --changelog_file changelog.md \
+    --print \
+    --debug
+```
+
+```json
+{"info": {"version": "0.6.0"}, "releases": {"0.6.0": [{"upload_time": "2022-10-26"}], "0.5.0": [{"upload_time": "2022-10-20"}], "0.4.0": [{"upload_time": "2022-08-07"}], "0.3.0": [{"upload_time": "2022-08-05"}], "0.2.0": [{"upload_time": "2022-08-03"}], "0.1.1": [{"upload_time": "2022-07-31"}], "0.1.0": [{"upload_time": "2022-07-31"}]}}
+```
+
+##### File
+
+```bash
+changelog2version \
+    --changelog_file changelog.md \
+    --output changelog.json \
+    --pretty \
+    --debug
+```
+
+See [example JSON file][ref-example-json-file]
+
 ## Advanced
 
 ### Custom regular expressions
@@ -202,6 +237,33 @@ __version__ = '.'.join(__version_info__) + '-rc1234'
 
 ```
 
+## Contributing
+
+### Unittests
+
+Run the unittests locally with the following command after installing this
+package in a virtual environment or by using `tox` to create one on each run.
+
+```bash
+# install the package with all its development dependencies
+pip install .[dev]
+
+# run all tests
+nose2 --config tests/unittest.cfg
+
+# run only one specific tests
+nose2 tests.test_extract_version.TestExtractVersion.test_version_line_regex
+```
+
+Generate the coverage files with
+
+```bash
+python create_report_dirs.py
+coverage html
+```
+
+The coverage report is placed at `reports/coverage/html/index.html`
+
 ## Credits
 
 Based on the [PyPa sample project][ref-pypa-sample]. Also a big thank you to
@@ -211,6 +273,7 @@ documentation and [regex example][ref-semver-regex-example]
 <!-- Links -->
 [ref-package-version-file]: src/changelog2version/version.py
 [ref-templates-folder]: src/changelog2version/templates
+[ref-example-json-file]: examples/changelog.json
 [ref-py-core-metadata-spec]: https://packaging.python.org/specifications/core-metadat
 [ref-pep440]: https://peps.python.org/pep-0440/
 [ref-pypa-sample]: https://github.com/pypa/sampleproject

diff --git a/changelog.md b/changelog.md
@@ -17,6 +17,17 @@ r"^\#\# \[\d{1,}[.]\d{1,}[.]\d{1,}\] \- \d{4}\-\d{2}-\d{2}$"
 -->
 
 ## Released
+## [0.6.0] - 2022-10-26
+### Added
+- Changelog can be parsed as JSON file with a similar format as the pypi
+  package JSON info available at `https://pypi.org/pypi/PACKAGENAME/json`, see
+  [#16][ref-issue-16]
+  The new option `--output` writes the data in JSON valid format to a file specified file, `--print` flag sends the data in JSON valid format to
+  `sys.stdout`. In both cases `--pretty` flag can be set to output the data with an indentation of 4.
+
+### Changed
+- `--version_file` argument is optional in order to allow a JSON output only
+
 ## [0.5.0] - 2022-10-20
 ### Added
 - Support additional version info file content in python version files by
@@ -128,15 +139,17 @@ r"^\#\# \[\d{1,}[.]\d{1,}[.]\d{1,}\] \- \d{4}\-\d{2}-\d{2}$"
 - Data folder after fork
 
 <!-- Links -->
-[Unreleased]: https://github.com/brainelectronics/changelog2version/compare/0.5.0...develop
+[Unreleased]: https://github.com/brainelectronics/changelog2version/compare/0.6.0...develop
 
+[0.6.0]: https://github.com/brainelectronics/changelog2version/tree/0.6.0
 [0.5.0]: https://github.com/brainelectronics/changelog2version/tree/0.5.0
 [0.4.0]: https://github.com/brainelectronics/changelog2version/tree/0.4.0
 [0.3.0]: https://github.com/brainelectronics/changelog2version/tree/0.3.0
 [0.2.0]: https://github.com/brainelectronics/changelog2version/tree/0.2.0
 [0.1.1]: https://github.com/brainelectronics/changelog2version/tree/0.1.1
 [0.1.0]: https://github.com/brainelectronics/changelog2version/tree/0.1.0
 
+[ref-issue-16]: https://github.com/brainelectronics/changelog2version/issues/16
 [ref-issue-5]: https://github.com/brainelectronics/changelog2version/issues/5
 [ref-issue-8]: https://github.com/brainelectronics/changelog2version/issues/8
 [ref-issue-11]: https://github.com/brainelectronics/changelog2version/issues/11

diff --git a/examples/changelog.json b/examples/changelog.json
@@ -0,0 +1,42 @@
+{
+    "info": {
+        "version": "0.6.0"
+    },
+    "releases": {
+        "0.6.0": [
+            {
+                "upload_time": "2022-10-26"
+            }
+        ],
+        "0.5.0": [
+            {
+                "upload_time": "2022-10-20"
+            }
+        ],
+        "0.4.0": [
+            {
+                "upload_time": "2022-08-07"
+            }
+        ],
+        "0.3.0": [
+            {
+                "upload_time": "2022-08-05"
+            }
+        ],
+        "0.2.0": [
+            {
+                "upload_time": "2022-08-03"
+            }
+        ],
+        "0.1.1": [
+            {
+                "upload_time": "2022-07-31"
+            }
+        ],
+        "0.1.0": [
+            {
+                "upload_time": "2022-07-31"
+            }
+        ]
+    }
+}
diff --git a/src/changelog2version/extract_version.py b/src/changelog2version/extract_version.py
@@ -11,7 +11,7 @@
 import re
 from semver import VersionInfo
 from sys import stdout
-from typing import Optional
+from typing import List, Optional
 
 
 class ExtractVersionError(Exception):
@@ -61,6 +61,13 @@ def __init__(self, logger: Optional[logging.Logger] = None):
             r"(?P<timestamp>\d{2,}:\d{2,}:\d{2,}?))?"   # time as HH:MM:SS
         )
 
+        self._date_line_regex = (
+            r".*"    # anything
+            r"(?P<datetime>\d{4}\-\d{2}-\d{2})"     # datetime as YYYY-MM-DD
+            r"(([T ]{1})"   # seperation between date and time by "T" or space
+            r"(?P<timestamp>\d{2,}:\d{2,}:\d{2,}?))?"   # time as HH:MM:SS
+        )
+
     @property
     def version_line_regex(self) -> str:
         """
@@ -109,6 +116,30 @@ def semver_line_regex(self, value: str) -> None:
         except re.error:
             raise ExtractVersionError("Invalid regex pattern")
 
+    @property
+    def date_line_regex(self) -> str:
+        """
+        Get regex to extract the date part from the complete version line
+
+        :returns:   Regex to get the date part
+        :rtype:     str
+        """
+        return self._date_line_regex
+
+    @date_line_regex.setter
+    def date_line_regex(self, value: str) -> None:
+        """
+        Set regex to extract the date part from the complete version line
+
+        :param      value:  Regex to get the date part
+        :type       value:  str
+        """
+        try:
+            re.compile(value)
+            self._date_line_regex = value
+        except re.error:
+            raise ExtractVersionError("Invalid regex pattern")
+
     @property
     def semver_data(self) -> VersionInfo:
         """
@@ -173,17 +204,69 @@ def parse_changelog(self, changelog_file: Path) -> str:
         """
         release_version_line = ""
 
+        release_version_lines = self.parse_changelog_completely(
+            changelog_file=changelog_file,
+            first_line_only=True)
+
+        if len(release_version_lines) >= 1:
+            release_version_line = release_version_lines[0]
+
+        return release_version_line
+
+    def parse_changelog_completely(self,
+                                   changelog_file: Path,
+                                   first_line_only: bool = False) -> List[str]:
+        """
+        Parse the changelog for all matching version lines
+
+        :param      changelog_file:  The path to the changelog file
+        :type       changelog_file:  Path
+        :param      first_line_only: Flag to break after first match found
+        :type       first_line_only: bool
+
+        :returns:   List of extracted semantic version strings
+        :rtype:     List[str]
+        """
+        release_version_lines = []
+
         with open(changelog_file, "r") as f:
             for line in f:
                 match = re.search(self.version_line_regex, line)
                 if match:
-                    release_version_line = match.group()
-                    break
+                    release_version_lines.append(match.group())
+                    if first_line_only:
+                        break
 
-        self._logger.debug("First matching release version line: '{}'".
-                           format(release_version_line))
+        self._logger.debug("Matching release version lines: '{}'".
+                           format(release_version_lines))
 
-        return release_version_line
+        return release_version_lines
+
+    def parse_semver_line_date(self, release_version_line: str) -> str:
+        """
+        Parse a version line for a valid ISO8601 datetime
+
+        Examples of a valid ISO8601 datetime lines:
+        - "## [0.2.0] - 2022-05-19"
+        - "## [107.3.18] - 1900-01-01 12:34:56"
+        - "## [1.0.0-alpha-a.b-c-somethinglong+build.1-aef.1-its-okay] - 2012-01-02"    # noqa
+
+        :param      release_version_line:  The release version line
+        :type       release_version_line:  str
+
+        :returns:   ISO8601 datetime string, e.g. "1970-01-01"
+        :rtype:     str
+        """
+        date_string = "1970-01-01"
+
+        match = re.search(self.date_line_regex, release_version_line)
+        if match:
+            if len(match.groups()) >= 4 and match.group(2):
+                date_string = match.group(1) + match.group(2)
+            else:
+                date_string = match.group(1)
+
+        return date_string
 
     def parse_semver_line(self, release_version_line: str) -> str:
         """