ansible-galaxy - set the cache file after getting all collection versions (#73557)

* Manage the in-memory cache in _call_galaxy but let the caller set the file cache after getting paginated results

* Add a test for caching successful and not caching unsuccessful paginated results

Co-authored-by: Sviatoslav Sydorenko <wk.cvs.github@sydorenko.org.ua>

3 files changed, 153 insertions, 4 deletions

diff --git a/changelogs/fragments/73557-ansible-galaxy-cache-paginated-response.yml b/changelogs/fragments/73557-ansible-galaxy-cache-paginated-response.yml
new file mode 100644
index 0000000000..6b95ab11dd
--- /dev/null
+++ b/changelogs/fragments/73557-ansible-galaxy-cache-paginated-response.yml
@@ -0,0 +1,3 @@
+bugfixes:
+  - ansible-galaxy - Cache the responses for available collection versions
+    after getting all pages. (https://github.com/ansible/ansible/issues/73071)
diff --git a/lib/ansible/galaxy/api.py b/lib/ansible/galaxy/api.py
index de5d6cc305..352082ded7 100644
--- a/lib/ansible/galaxy/api.py
+++ b/lib/ansible/galaxy/api.py
@@ -389,8 +389,6 @@ class GalaxyAPI:
             else:
                 path_cache['results'] = data
 
-            self._set_cache()
-
         return data
 
     def _add_auth_token(self, headers, url, token_type=None, required=False):
@@ -759,6 +757,7 @@ class GalaxyAPI:
         error_context_msg = 'Error when getting collection version metadata for %s.%s:%s from %s (%s)' \
                             % (namespace, name, version, self.name, self.api_server)
         data = self._call_galaxy(n_collection_url, error_context_msg=error_context_msg, cache=True)
+        self._set_cache()
 
         return CollectionVersionMetadata(data['namespace']['name'], data['collection']['name'], data['version'],
                                          data['download_url'], data['artifact']['sha256'],
@@ -843,5 +842,6 @@ class GalaxyAPI:
 
             data = self._call_galaxy(to_native(next_link, errors='surrogate_or_strict'),
                                      error_context_msg=error_context_msg, cache=True)
+        self._set_cache()
 
         return versions
diff --git a/test/units/galaxy/test_api.py b/test/units/galaxy/test_api.py
index 1e8af129d6..ebda706c58 100644
--- a/test/units/galaxy/test_api.py
+++ b/test/units/galaxy/test_api.py
@@ -63,10 +63,10 @@ def cache_dir(tmp_path_factory, monkeypatch):
     yield cache_dir
 
 
-def get_test_galaxy_api(url, version, token_ins=None, token_value=None):
+def get_test_galaxy_api(url, version, token_ins=None, token_value=None, no_cache=True):
     token_value = token_value or "my token"
     token_ins = token_ins or GalaxyToken(token_value)
-    api = GalaxyAPI(None, "test", url)
+    api = GalaxyAPI(None, "test", url, no_cache=no_cache)
     # Warning, this doesn't test g_connect() because _availabe_api_versions is set here.  That means
     # that urls for v2 servers have to append '/api/' themselves in the input data.
     api._available_api_versions = {version: '%s' % version}
@@ -75,6 +75,60 @@ def get_test_galaxy_api(url, version, token_ins=None, token_value=None):
     return api
 
 
+def get_collection_versions(namespace='namespace', name='collection'):
+    base_url = 'https://galaxy.server.com/api/v2/collections/{0}/{1}/'.format(namespace, name)
+    versions_url = base_url + 'versions/'
+
+    # Response for collection info
+    responses = [
+        {
+            "id": 1000,
+            "href": base_url,
+            "name": name,
+            "namespace": {
+                "id": 30000,
+                "href": "https://galaxy.ansible.com/api/v1/namespaces/30000/",
+                "name": namespace,
+            },
+            "versions_url": versions_url,
+            "latest_version": {
+                "version": "1.0.5",
+                "href": versions_url + "1.0.5/"
+            },
+            "deprecated": False,
+            "created": "2021-02-09T16:55:42.749915-05:00",
+            "modified": "2021-02-09T16:55:42.749915-05:00",
+        }
+    ]
+
+    # Paginated responses for versions
+    page_versions = (('1.0.0', '1.0.1',), ('1.0.2', '1.0.3',), ('1.0.4', '1.0.5'),)
+    last_page = None
+    for page in range(1, len(page_versions) + 1):
+        if page < len(page_versions):
+            next_page = versions_url + '?page={0}'.format(page + 1)
+        else:
+            next_page = None
+
+        version_results = []
+        for version in page_versions[int(page - 1)]:
+            version_results.append(
+                {'version': version, 'href': versions_url + '{0}/'.format(version)}
+            )
+
+        responses.append(
+            {
+                'count': 6,
+                'next': next_page,
+                'previous': last_page,
+                'results': version_results,
+            }
+        )
+        last_page = page
+
+    return responses
+
+
 def test_api_no_auth():
     api = GalaxyAPI(None, "test", "https://galaxy.ansible.com/api/")
     actual = {}
@@ -974,6 +1028,98 @@ def test_cache_invalid_cache_content(content, cache_dir):
     assert stat.S_IMODE(os.stat(cache_file).st_mode) == 0o664
 
 
+def test_cache_complete_pagination(cache_dir, monkeypatch):
+
+    responses = get_collection_versions()
+    cache_file = os.path.join(cache_dir, 'api.json')
+
+    api = get_test_galaxy_api('https://galaxy.server.com/api/', 'v2', no_cache=False)
+
+    mock_open = MagicMock(
+        side_effect=[
+            StringIO(to_text(json.dumps(r)))
+            for r in responses
+        ]
+    )
+    monkeypatch.setattr(galaxy_api, 'open_url', mock_open)
+
+    actual_versions = api.get_collection_versions('namespace', 'collection')
+    assert actual_versions == [u'1.0.0', u'1.0.1', u'1.0.2', u'1.0.3', u'1.0.4', u'1.0.5']
+
+    with open(cache_file) as fd:
+        final_cache = json.loads(fd.read())
+
+    cached_server = final_cache['galaxy.server.com:']
+    cached_collection = cached_server['/api/v2/collections/namespace/collection/versions/']
+    cached_versions = [r['version'] for r in cached_collection['results']]
+
+    assert final_cache == api._cache
+    assert cached_versions == actual_versions
+
+
+def test_cache_flaky_pagination(cache_dir, monkeypatch):
+
+    responses = get_collection_versions()
+    cache_file = os.path.join(cache_dir, 'api.json')
+
+    api = get_test_galaxy_api('https://galaxy.server.com/api/', 'v2', no_cache=False)
+
+    # First attempt, fail midway through
+    mock_open = MagicMock(
+        side_effect=[
+            StringIO(to_text(json.dumps(responses[0]))),
+            StringIO(to_text(json.dumps(responses[1]))),
+            urllib_error.HTTPError(responses[1]['next'], 500, 'Error', {}, StringIO()),
+            StringIO(to_text(json.dumps(responses[3]))),
+        ]
+    )
+    monkeypatch.setattr(galaxy_api, 'open_url', mock_open)
+
+    expected = (
+        r'Error when getting available collection versions for namespace\.collection '
+        r'from test \(https://galaxy\.server\.com/api/\) '
+        r'\(HTTP Code: 500, Message: Error Code: Unknown\)'
+    )
+    with pytest.raises(GalaxyError, match=expected):
+        api.get_collection_versions('namespace', 'collection')
+
+    with open(cache_file) as fd:
+        final_cache = json.loads(fd.read())
+
+    assert final_cache == {
+        'version': 1,
+        'galaxy.server.com:': {
+            'modified': {
+                'namespace.collection': responses[0]['modified']
+            }
+        }
+    }
+
+    # Reset API
+    api = get_test_galaxy_api('https://galaxy.server.com/api/', 'v2', no_cache=False)
+
+    # Second attempt is successful so cache should be populated
+    mock_open = MagicMock(
+        side_effect=[
+            StringIO(to_text(json.dumps(r)))
+            for r in responses
+        ]
+    )
+    monkeypatch.setattr(galaxy_api, 'open_url', mock_open)
+
+    actual_versions = api.get_collection_versions('namespace', 'collection')
+    assert actual_versions == [u'1.0.0', u'1.0.1', u'1.0.2', u'1.0.3', u'1.0.4', u'1.0.5']
+
+    with open(cache_file) as fd:
+        final_cache = json.loads(fd.read())
+
+    cached_server = final_cache['galaxy.server.com:']
+    cached_collection = cached_server['/api/v2/collections/namespace/collection/versions/']
+    cached_versions = [r['version'] for r in cached_collection['results']]
+
+    assert cached_versions == actual_versions
+
+
 def test_world_writable_cache(cache_dir, monkeypatch):
     mock_warning = MagicMock()
     monkeypatch.setattr(Display, 'warning', mock_warning)