Add api_origin_metadata_search.

Reviewers: #reviewers, anlambert

Reviewed By: #reviewers, anlambert

Subscribers: anlambert, zack, swh-public-ci

Differential Revision: https://forge.softwareheritage.org/D751

swh/web/api/views/origin.py

@@ -6,6 +6,7 @@
 from distutils.util import strtobool
 
 from swh.web.common import service
+from swh.web.common.exc import BadInputExc
 from swh.web.common.utils import (
     reverse, get_origin_visits
 )
@@ -174,6 +175,56 @@ def api_origin_search(request, url_pattern):
     return result
 
 
+@api_route(r'/origin/metadata-search/',
+           'api-origin-metadata-search')
+@api_doc('/origin/metadata-search/', noargs=True)
+def api_origin_metadata_search(request):
+    """
+    .. http:get:: /api/1/origin/metadata-search/
+
+        Search for software origins whose metadata (expressed as a
+        JSON-LD/CodeMeta dictionary) match the provided criteria.
+        For now, only full-text search on this dictionary is supported.
+
+        :query str fulltext: a string that will be matched against origin metadata;
+            results are ranked and ordered starting with the best ones.
+        :query int limit: the maximum number of found origins to return
+            (bounded to 100)
+
+        :>jsonarr number origin_id: the origin unique identifier
+        :>jsonarr dict metadata: metadata of the origin (as a JSON-LD/CodeMeta dictionary)
+        :>jsonarr string from_revision: the revision used to extract these
+            metadata (the current HEAD or one of the former HEADs)
+        :>jsonarr dict tool: the tool used to extract these metadata
+
+        :reqheader Accept: the requested response content type,
+            either ``application/json`` (default) or ``application/yaml``
+        :resheader Content-Type: this depends on :http:header:`Accept` header of request
+
+        **Allowed HTTP Methods:** :http:method:`get`, :http:method:`head`, :http:method:`options`
+
+        :statuscode 200: no error
+
+        **Example:**
+
+        .. parsed-literal::
+
+            :swh_web_api:`origin/metadata-search/?limit=2&fulltext=Jane%20Doe`
+    """ # noqa
+    fulltext = request.query_params.get('fulltext', None)
+    limit = min(int(request.query_params.get('limit', '70')), 100)
+
+    if not fulltext:
+        content = '"fulltext" must be provided and non-empty.'
+        raise BadInputExc(content)
+
+    results = api_lookup(service.search_origin_metadata, fulltext, limit)
+
+    return {
+        'results': results,
+    }
+
+
 @api_route(r'/origin/(?P<origin_id>[0-9]+)/visits/', 'api-origin-visits')
 @api_doc('/origin/visits/')
 def api_origin_visits(request, origin_id):

swh/web/common/service.py

@@ -266,6 +266,25 @@ def search_origin(url_pattern, offset=0, limit=50, regexp=False,
     return map(converters.from_origin, origins)
 
 
+def search_origin_metadata(fulltext, limit=50):
+    """Search for origins whose metadata match a provided string pattern.
+
+    Args:
+        fulltext: the string pattern to search for in origin metadata
+        offset: number of found origins to skip before returning results
+        limit: the maximum number of found origins to return
+
+    Returns:
+        list of origin metadata as dict.
+
+    """
+    results = idx_storage.origin_intrinsic_metadata_search_fulltext(
+        conjunction=[fulltext], limit=limit)
+    for result in results:
+        result['from_revision'] = hashutil.hash_to_hex(result['from_revision'])
+    return results
+
+
 def lookup_person(person_id):
     """Return information about the person with id person_id.
 

swh/web/tests/api/views/test_origin.py

@@ -244,3 +244,109 @@ class OriginApiTestCase(SWHWebTestCase, APITestCase):
         })
 
         mock_service.lookup_origin.assert_called_with({'id': '4321'})
+
+    @patch('swh.web.common.service.idx_storage')
+    def test_api_origin_metadata_search(self, mock_idx_storage):
+        # given
+        mock_idx_storage.origin_intrinsic_metadata_search_fulltext \
+            .return_value = [{
+                'from_revision':
+                b'p&\xb7\xc1\xa2\xafVR\x1e\x95\x1c\x01\xed \xf2U\xfa\x05B8',
+                'metadata': {'author': 'Jane Doe'},
+                'origin_id': 54974445,
+                'tool': {
+                    'configuration': {
+                        'context': ['NpmMapping', 'CodemetaMapping'],
+                        'type': 'local'
+                    },
+                    'id': 3,
+                    'name': 'swh-metadata-detector',
+                    'version': '0.0.1'
+                }
+            }]
+
+        # when
+        rv = self.client.get(
+            '/api/1/origin/metadata-search/?fulltext=Jane%20Doe')
+
+        # then
+        self.assertEqual(rv.status_code, 200, rv.content)
+        self.assertEqual(rv['Content-Type'], 'application/json')
+        expected_data = [{
+            'origin_id': 54974445,
+            'metadata': {'author': 'Jane Doe'},
+            'from_revision': '7026b7c1a2af56521e951c01ed20f255fa054238',
+            'tool': {
+                'configuration': {
+                    'context': ['NpmMapping', 'CodemetaMapping'],
+                    'type': 'local'
+                },
+                'id': 3,
+                'name': 'swh-metadata-detector',
+                'version': '0.0.1',
+            }
+        }]
+        self.assertEqual(rv.data, expected_data)
+        mock_idx_storage.origin_intrinsic_metadata_search_fulltext \
+            .assert_called_with(conjunction=['Jane Doe'], limit=70)
+
+    @patch('swh.web.common.service.idx_storage')
+    def test_api_origin_metadata_search_limit(self, mock_idx_storage):
+        # given
+        mock_idx_storage.origin_intrinsic_metadata_search_fulltext \
+            .return_value = [{
+                'from_revision':
+                b'p&\xb7\xc1\xa2\xafVR\x1e\x95\x1c\x01\xed \xf2U\xfa\x05B8',
+                'metadata': {'author': 'Jane Doe'},
+                'origin_id': 54974445,
+                'tool': {
+                    'configuration': {
+                        'context': ['NpmMapping', 'CodemetaMapping'],
+                        'type': 'local'
+                    },
+                    'id': 3,
+                    'name': 'swh-metadata-detector',
+                    'version': '0.0.1'
+                }
+            }]
+
+        # when
+        rv = self.client.get(
+            '/api/1/origin/metadata-search/?fulltext=Jane%20Doe')
+
+        # then
+        self.assertEqual(rv.status_code, 200, rv.content)
+        self.assertEqual(rv['Content-Type'], 'application/json')
+        self.assertEqual(len(rv.data), 1)
+        mock_idx_storage.origin_intrinsic_metadata_search_fulltext \
+            .assert_called_with(conjunction=['Jane Doe'], limit=70)
+
+        # when
+        rv = self.client.get(
+            '/api/1/origin/metadata-search/?fulltext=Jane%20Doe&limit=10')
+
+        # then
+        self.assertEqual(rv.status_code, 200, rv.content)
+        self.assertEqual(rv['Content-Type'], 'application/json')
+        self.assertEqual(len(rv.data), 1)
+        mock_idx_storage.origin_intrinsic_metadata_search_fulltext \
+            .assert_called_with(conjunction=['Jane Doe'], limit=10)
+
+        # when
+        rv = self.client.get(
+            '/api/1/origin/metadata-search/?fulltext=Jane%20Doe&limit=987')
+
+        # then
+        self.assertEqual(rv.status_code, 200, rv.content)
+        self.assertEqual(rv['Content-Type'], 'application/json')
+        self.assertEqual(len(rv.data), 1)
+        mock_idx_storage.origin_intrinsic_metadata_search_fulltext \
+            .assert_called_with(conjunction=['Jane Doe'], limit=100)
+
+    @patch('swh.web.common.service.idx_storage')
+    def test_api_origin_metadata_search_invalid(self, mock_idx_storage):
+        rv = self.client.get('/api/1/origin/metadata-search/')
+
+        # then
+        self.assertEqual(rv.status_code, 400, rv.content)
+        mock_idx_storage.assert_not_called()

[Phabricator Migration user]

Some references in the commit message have been migrated:

• D751 is now !61 (closed)