include total_count_per_object_type in search response

doc/release-notes/ 11065-extend-search-api-to-include-type-counts.md

@@ -0,0 +1 @@
+The JSON payload of the search endpoint has been extended to include total_count_per_object_type for types: dataverse, dataset, and files when the search parameter "&show_type_counts=true" is passed in.

doc/sphinx-guides/source/api/search.rst

@@ -21,9 +21,9 @@ Please note that in Dataverse Software 4.3 and older the "citation" field wrappe
 Parameters
 ----------
 
-===============  =======  ===========
+================ =======  ===========
 Name             Type     Description
-===============  =======  ===========
+================ =======  ===========
 q                string   The search term or terms. Using "title:data" will search only the "title" field. "*" can be used as a wildcard either alone or adjacent to a term (i.e. "bird*"). For example, https://demo.dataverse.org/api/search?q=title:data . For a list of fields to search, please see https://github.com/IQSS/dataverse/issues/2558 (for now).
 type             string   Can be either "dataverse", "dataset", or "file". Multiple "type" parameters can be used to include multiple types (i.e. ``type=dataset&type=file``). If omitted, all types will be returned.  For example, https://demo.dataverse.org/api/search?q=*&type=dataset
 subtree          string   The identifier of the Dataverse collection to which the search should be narrowed. The subtree of this Dataverse collection and all its children will be searched.  Multiple "subtree" parameters can be used to include multiple Dataverse collections. For example, https://demo.dataverse.org/api/search?q=data&subtree=birds&subtree=cats .
@@ -38,7 +38,8 @@ show_entity_ids  boolean  Whether or not to show the database IDs of the search
 geo_point        string	  Latitude and longitude in the form ``geo_point=42.3,-71.1``. You must supply ``geo_radius`` as well. See also :ref:`geospatial-search`.
 geo_radius       string	  Radial distance in kilometers from ``geo_point`` (which must be supplied as well) such as ``geo_radius=1.5``.
 metadata_fields  string	  Includes the requested fields for each dataset in the response. Multiple "metadata_fields" parameters can be used to include several fields. The value must be in the form "{metadata_block_name}:{field_name}" to include a specific field from a metadata block (see :ref:`example <dynamic-citation-some>`) or "{metadata_field_set_name}:\*" to include all the fields for a metadata block (see :ref:`example <dynamic-citation-all>`). "{field_name}" cannot be a subfield of a compound field. If "{field_name}" is a compound field, all subfields are included.
-===============  =======  ===========
+show_type_counts boolean  Whether or not to include total_count_per_object_type for types: Dataverse, Dataset, and Files.
+================ =======  ===========
 
 Basic Search Example
 --------------------
@@ -701,7 +702,11 @@ The above example ``metadata_fields=citation:dsDescription&metadata_fields=citat
                     "published_at": "2021-03-16T08:11:54Z"
                 }
             ],
-            "count_in_response": 4
+            "count_in_response": 4,
+            "total_count_per_object_type": {
+                "Datasets": 2,
+                "Dataverses": 2
+            }
         }
     }
 

src/main/java/edu/harvard/iq/dataverse/api/Search.java

@@ -73,6 +73,7 @@ public Response search(
             @QueryParam("metadata_fields") List<String> metadataFields,
             @QueryParam("geo_point") String geoPointRequested,
             @QueryParam("geo_radius") String geoRadiusRequested,
+            @QueryParam("show_type_counts") boolean showTypeCounts,
             @Context HttpServletResponse response
     ) {
 
@@ -210,6 +211,15 @@ public Response search(
             }
 
             value.add("count_in_response", solrSearchResults.size());
+            if (showTypeCounts && !solrQueryResponse.getTypeFacetCategories().isEmpty()) {
+                JsonObjectBuilder objectTypeCounts = Json.createObjectBuilder();
+                for (FacetCategory facetCategory : solrQueryResponse.getTypeFacetCategories()) {
+                    for (FacetLabel facetLabel : facetCategory.getFacetLabel()) {
+                        objectTypeCounts.add(facetLabel.getName(), facetLabel.getCount());
+                    }
+                }
+                value.add("total_count_per_object_type", objectTypeCounts);
+            }
             /**
              * @todo Returning the fq might be useful as a troubleshooting aid
              * but we don't want to expose the raw dataverse database ids in

src/test/java/edu/harvard/iq/dataverse/api/SearchIT.java

@@ -1347,4 +1347,65 @@ public void testSearchFilesAndUrlImages() {
                 .body("data.items[0].url", CoreMatchers.containsString("/datafile/"))
                 .body("data.items[0]", CoreMatchers.not(CoreMatchers.hasItem("image_url")));
     }
+
+    @Test
+    public void testShowTypeCounts() {
+        //Create 1 user and 1 Dataverse/Collection
+        Response createUser = UtilIT.createRandomUser();
+        String username = UtilIT.getUsernameFromResponse(createUser);
+        String apiToken = UtilIT.getApiTokenFromResponse(createUser);
+        String affiliation = "testAffiliation";
+
+        // test total_count_per_object_type is not included because the results are empty
+        Response searchResp = UtilIT.search(username, apiToken, "&show_type_counts=true");
+        searchResp.then().assertThat()
+                .statusCode(OK.getStatusCode())
+                .body("data.total_count_per_object_type", CoreMatchers.equalTo(null));
+
+        Response createDataverseResponse = UtilIT.createRandomDataverse(apiToken, affiliation);
+        assertEquals(201, createDataverseResponse.getStatusCode());
+        String dataverseAlias = UtilIT.getAliasFromResponse(createDataverseResponse);
+
+        // create 3 Datasets, each with 2 Datafiles
+        for (int i = 0; i < 3; i++) {
+            Response createDatasetResponse = UtilIT.createRandomDatasetViaNativeApi(dataverseAlias, apiToken);
+            createDatasetResponse.then().assertThat()
+                    .statusCode(CREATED.getStatusCode());
+            String datasetId = UtilIT.getDatasetIdFromResponse(createDatasetResponse).toString();
+
+            // putting the dataverseAlias in the description of each file so the search q={dataverseAlias} will return dataverse, dataset, and files for this test only
+            String jsonAsString = "{\"description\":\"" + dataverseAlias + "\",\"directoryLabel\":\"data/subdir1\",\"categories\":[\"Data\"], \"restrict\":\"false\"  }";
+
+            String pathToFile = "src/main/webapp/resources/images/dataverseproject.png";
+            Response uploadImage = UtilIT.uploadFileViaNative(datasetId, pathToFile, jsonAsString, apiToken);
+            uploadImage.then().assertThat()
+                    .statusCode(200);
+            pathToFile = "src/main/webapp/resources/js/mydata.js";
+            Response uploadFile = UtilIT.uploadFileViaNative(datasetId, pathToFile, jsonAsString, apiToken);
+            uploadFile.then().assertThat()
+                    .statusCode(200);
+
+            // This call forces a wait for dataset indexing to finish and gives time for file uploads to complete
+            UtilIT.search("id:dataset_" + datasetId, apiToken);
+        }
+
+        // Test Search without show_type_counts
+        searchResp = UtilIT.search(dataverseAlias, apiToken);
+        searchResp.then().assertThat()
+                .statusCode(OK.getStatusCode())
+                .body("data.total_count_per_object_type", CoreMatchers.equalTo(null));
+        // Test Search with show_type_counts = FALSE
+        searchResp = UtilIT.search(dataverseAlias, apiToken, "&show_type_counts=false");
+        searchResp.then().assertThat()
+                .statusCode(OK.getStatusCode())
+                .body("data.total_count_per_object_type", CoreMatchers.equalTo(null));
+        // Test Search with show_type_counts = TRUE
+        searchResp = UtilIT.search(dataverseAlias, apiToken, "&show_type_counts=true");
+        searchResp.prettyPrint();
+        searchResp.then().assertThat()
+                .statusCode(OK.getStatusCode())
+                .body("data.total_count_per_object_type.Dataverses", CoreMatchers.is(1))
+                .body("data.total_count_per_object_type.Datasets", CoreMatchers.is(3))
+                .body("data.total_count_per_object_type.Files", CoreMatchers.is(6));
+    }
 }