added method to retrieve release info

CHANGELOG.md

@@ -1,5 +1,11 @@
 # Changes
 
+## [Unreleased]
+
+- Add Python 3.12 to supported versions.
+- Added API method to retrieve release information from GitHub/Zenodo.
+
+
 ## [2.0.0] - 2023-10-20
 
 Zenodo's upgrade from Oct. 13, 2023 brought a couple of breaking changes for the Zenodo API.

setup.cfg

@@ -24,6 +24,7 @@ classifiers =
     Programming Language :: Python :: 3.9
     Programming Language :: Python :: 3.10
     Programming Language :: Python :: 3.11
+    Programming Language :: Python :: 3.12
     Programming Language :: Python :: Implementation :: CPython
     Programming Language :: Python :: Implementation :: PyPy
     License :: OSI Approved :: Apache Software License
@@ -92,7 +93,7 @@ show_missing = true
 skip_covered = true
 
 [tox:tox]
-envlist = py38, py39, py310, py311
+envlist = py38, py39, py310, py311, py312
 isolated_build = true
 skip_missing_interpreter = true
 

src/cldfzenodo/api.py

@@ -1,11 +1,15 @@
 import re
 import json
+import shlex
 import typing
+import subprocess
 import urllib.error
 import urllib.parse
 import urllib.request
 
+from pycldf import Source
 from pycldf.ext.discovery import DatasetResolver
+from clldutils.path import ensure_cmd
 
 from cldfzenodo.record import Record, ZENODO_DOI_FORMAT, ZENODO_DOI_PATTERN, get_doi
 
@@ -189,6 +193,69 @@ def iter_records(self,
         if res.error:
             return res  # pragma: no cover
 
+    def get_github_release_info(
+            self,
+            repos: str,
+            tag: typing.Optional[str] = None,
+            bibid: typing.Optional[str] = None
+    ) -> typing.Tuple[
+        typing.Union[str, None],
+        typing.Union[str, None],
+        typing.Union[str, None],
+        typing.Union[Source, None]
+    ]:
+        """
+        Get information about a GitHub repository release that has been archived with Zenodo.
+
+        We assume, that the DOI of the upload on Zenodo has been added to the release description
+        on GitHub.
+
+        :param repos: GitHub repository specifier in the form `org/repo`.
+        :param tag: Optional version tag.
+        :param bibid: Optional identifier to use for the BibTeX record.
+        :return: A quadruple (tag, DOI, APA citation, BibTeX record)
+        """
+        def gh(args):
+            return subprocess.check_output([ensure_cmd('gh')] + shlex.split(args)).decode('utf8')
+
+        version_pattern = re.compile(r'(?P<tag>v[0-9]+\.[0-9]+(\.[0-9]+)?)')
+        doi_pattern = re.compile(r'https://doi\.org/(?P<doi>10\.5281/zenodo\.[0-9]+)')
+
+        if not tag:
+            latest = gh('release list -L 1 -R {}'.format(repos))
+            if not latest:
+                return None, None, None, None  # pragma: no cover
+            match = version_pattern.search(latest)
+            assert match, latest
+            tag = match.group('tag')
+
+        info = gh('release view {} -R {}'.format(tag, repos))
+        match = None
+        for match in doi_pattern.finditer(info):
+            pass
+        if match:
+            doi = match.group('doi')
+            if doi:
+                cit = self.records(
+                    id_=ZENODO_DOI_PATTERN.match(doi).group('recid'),
+                    params=dict(style='apa'),
+                    headers=dict(Accept='text/x-bibliography')).strip()
+                bib = self.records(
+                    id_=ZENODO_DOI_PATTERN.match(doi).group('recid'),
+                    params=dict(style='bibtex'),
+                    headers=dict(Accept='text/x-bibliography')).strip()
+                bib = Source.from_bibtex(re.sub(
+                    r', (?P<field>[a-zA-Z]+)=',
+                    lambda m: ',\n  {} = '.format(m.group('field')),
+                    bib), _check_id=False)
+                bib.id = bibid or repos.replace('/', '_')
+                del bib['abstractNote']
+                del bib['month']
+                bib['edition'] = tag
+                bib['type'] = 'Data set'
+                return tag, doi, cit, bib
+        return tag, None, None, None  # pragma: no cover
+
 
 # Since an Api instance does some caching, we provide one instance as module global.
 API = Api()

tests/test_api.py

@@ -53,3 +53,52 @@ def test_Api_get_record(urlopen, mocker):
     assert Api().get_record(conceptdoi='10.5281/zenodo.5173799', version='v1.1') is not None
 
     assert list(Api().iter_versions('10.5281/zenodo.5173799'))
+
+
+def test_Api_get_github_release_info(mocker):
+    class Subprocess:
+        @staticmethod
+        def check_output(cmd):
+            if 'list' in cmd:
+                return b"""TITLE                   TYPE    TAG NAME  PUBLISHED         
+Glottolog Database 4.8  Latest  v4.8      about 4 months ago"""
+            return """v4.8
+xrotwang released this about 4 months ago
+
+  Hammarström, Harald & Forkel, Robert & Haspelmath, Martin & Bank, Sebastian. 2023. Glottolog 4.8. Leipzig: Max      
+  Planck Institute for Evolutionary Anthropology. (Available online at https://glottolog.org/)                        
+                                                                                                                      
+  DOI https://doi.org/10.5281/zenodo.8131084                                                                          
+
+
+View on GitHub: https://github.com/glottolog/glottolog/releases/tag/v4.8""".encode('utf8')
+
+    class Response:
+        def __init__(self, what):
+            self.what = what
+
+        def read(self):
+            if self.what == 'apa':
+                return b'The Citation'
+            return (' @misc{harald hammarström_robert forkel_martin haspelmath_sebastian bank_2023, '
+                    'title={glottolog/glottolog: Glottolog database 4.8}, '
+                    'DOI={10.5281/zenodo.8131084}, '
+                    'abstractNote={Hammarström, Harald & Forkel, Robert & Haspelmath, Martin & Bank, Sebastian. 2023. Glottolog 4.8. Leipzig: Max Planck Institute for Evolutionary Anthropology. (Available online at https://glottolog.org)}, '
+                    'publisher={Zenodo}, '
+                    'author={Harald Hammarström and Robert Forkel and Martin Haspelmath and Sebastian Bank}, '
+                    'year={2023}, '
+                    'month={Jul} }').encode('utf8')
+
+    def urlopen(req):
+        if 'style=apa' in req.full_url:
+            return Response('apa')
+        return Response('bibtex')
+
+    mocker.patch('cldfzenodo.api.subprocess', Subprocess())
+    mocker.patch('cldfzenodo.api.urllib.request.urlopen', urlopen)
+
+    tag, doi, cit, bib = Api().get_github_release_info('glottolog/glottolog', bibid='x')
+    assert tag == 'v4.8'
+    assert doi == '10.5281/zenodo.8131084'
+    assert cit == 'The Citation'
+    assert bib.id == 'x'