feat: handle multiple tags filter on datasets

CHANGELOG.md

@@ -2,15 +2,15 @@
 
 ## Current (in progress)
 
-- Nothing yet
+- feat: handle multiple tags filter on datasets [#50](https://github.com/opendatateam/udata-search-service/pull/50)
 
 ## 2.2.0 (2024-11-07)
 
 - Add badge organization filter on dataset & reuse [#49](https://github.com/opendatateam/udata-search-service/pull/49)
 - Add dataservices search [#48](https://github.com/opendatateam/udata-search-service/pull/48)
 
 :warning: To use these new features, you will need to:
-    - init ES indices with `udata-search-service init-es` 
+    - init ES indices with `udata-search-service init-es`
     - index datasets and dataservices on udata side with `udata search index dataset` and `udata search index dataservice`
 
 ## 2.1.0 (2024-10-07)

pyproject.toml

@@ -22,6 +22,8 @@ dependencies = [
     "pytest-flask==1.2.0",
     "markdown==3.3.3",
     "beautifulsoup4==4.10.0",
+    # pinned to a known working version, since 3.1.3 will fail while running tests
+    "werkzeug==3.0.4",
 ]
 
 [project.urls]

tests/test_api.py

@@ -678,7 +678,7 @@ def test_api_dataset_search_with_temporal_filter(app, client, search_client, fak
 
 def test_api_search_with_tag_filter(app, client, search_client, faker):
     for i in range(4):
-        search_client.index_dataset(DatasetFactory(tags=['test-tag'] if i % 2 else ['not-test-tag']))
+        search_client.index_dataset(DatasetFactory(tags=['test-tag', 'test-tag-2'] if i % 2 else ['not-test-tag']))
         search_client.index_reuse(ReuseFactory(tags=['test-tag'] if i % 2 else ['not-test-tag']))
 
     # Without this, ElasticSearch does not seem to have the time to index.
@@ -688,7 +688,10 @@ def test_api_search_with_tag_filter(app, client, search_client, faker):
     resp = client.get(url_for('api.dataset_search'))
     assert resp.json['total'] == 4
 
-    resp = client.get(url_for('api.dataset_search', tag='test-tag'))
+    resp = client.get(url_for('api.dataset_search', tag=['test-tag', 'test-tag-2']))
+    assert resp.json['total'] == 2
+
+    resp = client.get(url_for('api.dataset_search', tag=['not-test-tag']))
     assert resp.json['total'] == 2
 
     resp = client.get(url_for('api.reuse_search'))

tests/test_search_client.py

@@ -190,7 +190,7 @@ def test_search_datasets_with_temporal_filters(app, client, search_client, faker
 def test_search_with_tag_filter(app, client, search_client, faker):
     for i in range(4):
         search_client.index_dataset(DatasetFactory(
-            tags=['test-tag'] if i % 2 else ['not-test-tag']
+            tags=['test-tag', f'test-tag-{i}'] if i % 2 else ['not-test-tag']
         ))
 
         search_client.index_reuse(ReuseFactory(
@@ -206,7 +206,9 @@ def test_search_with_tag_filter(app, client, search_client, faker):
 
     results_number, res = search_client.query_datasets(None, 0, 20, {})
     assert results_number == 4
-    results_number, res = search_client.query_datasets(None, 0, 20, {'tags': 'test-tag'})
+    results_number, res = search_client.query_datasets(None, 0, 20, {'tags': ['test-tag', 'test-tag-1']})
+    assert results_number == 1
+    results_number, res = search_client.query_datasets(None, 0, 20, {'tags': ['not-test-tag']})
     assert results_number == 2
     results_number, res = search_client.query_reuses(None, 0, 20, {})
     assert results_number == 4

udata_search_service/infrastructure/search_clients.py

@@ -258,6 +258,12 @@ def query_datasets(self, query_text: str, offset: int, page_size: int, filters:
                 search = search.filter('range', **{'temporal_coverage_start': {'lte': value}})
             elif key == 'temporal_coverage_end':
                 search = search.filter('range', **{'temporal_coverage_end': {'gte': value}})
+            elif key == 'tags':
+                # build an AND filter from tags list
+                tag_filters = [query.Q('term', tags=tag) for tag in value]
+                search = search.filter(
+                    query.Bool(must=tag_filters)
+                )
             else:
                 search = search.filter('term', **{key: value})
 

udata_search_service/presentation/api.py

@@ -10,6 +10,7 @@
 from udata_search_service.infrastructure.search_clients import ElasticClient
 from udata_search_service.infrastructure.consumers import DatasetConsumer, ReuseConsumer, OrganizationConsumer, DataserviceConsumer
 from udata_search_service.infrastructure.migrate import set_alias as set_alias_func
+from udata_search_service.presentation.utils import is_list_type
 
 
 bp = Blueprint('api', __name__, url_prefix='/api/1')
@@ -38,7 +39,7 @@ class DatasetArgs(BaseModel):
     page: Optional[int] = 1
     page_size: Optional[int] = 20
     sort: Optional[str] = None
-    tag: Optional[str] = None
+    tag: Optional[list[str]] = None
     badge: Optional[str] = None
     organization: Optional[str] = None
     organization_badge: Optional[str] = None
@@ -69,6 +70,22 @@ def sort_validate(cls, value):
             raise ValueError('Temporal coverage does not match the right pattern.')
         return value
 
+    @classmethod
+    def from_request_args(cls, request_args) -> 'DatasetArgs':
+        def get_list_args() -> dict:
+            return {
+                key: value
+                for key, value in request_args.to_dict(flat=False).items()
+                if key in cls.__fields__
+                and is_list_type(cls.__fields__[key].outer_type_)
+            }
+
+        return cls(
+            **{
+                **request_args.to_dict(),
+                **get_list_args(),
+            }
+        )
 
 class ReuseArgs(BaseModel):
     q: Optional[str] = None
@@ -192,7 +209,7 @@ def dataset_unindex(dataset_id: str, dataset_service: DatasetService = Provide[C
 @inject
 def datasets_search(dataset_service: DatasetService = Provide[Container.dataset_service]):
     try:
-        request_args = DatasetArgs(**request.args)
+        request_args = DatasetArgs.from_request_args(request.args)
     except ValidationError as e:
         abort(400, e)
 

udata_search_service/presentation/utils.py

@@ -0,0 +1,10 @@
+from typing import Type, get_origin, Union, get_args
+
+
+def is_list_type(type_: Type) -> bool:
+    origin = get_origin(type_)
+    if origin is list:
+        return True
+    if origin is Union:
+        return any(is_list_type(t) for t in get_args(type_))
+    return False