Fix queries over paged collections always being empty

This was a regression of #1180 due to the v3 release refactoring.

This commit adds a test to ensure this one does not break again.

Closes #1180

engines/query-sparql/test/QuerySparql-test.ts

@@ -731,6 +731,18 @@ SELECT * WHERE {
         });
         expect(called).toBe(0);
       });
+
+      it('with two triple patterns over a paged collection', async() => {
+        const bindingsStream = await engine.queryBindings(`
+SELECT *
+WHERE {
+  <https://opendata.picturae.com/dataset/dre_a2a_webservice> <http://purl.org/dc/terms/identifier> ?i.
+  <https://opendata.picturae.com/dataset/dre_a2a_webservice> <http://purl.org/dc/terms/issued> ?issued.
+}`, {
+          sources: [ 'https://opendata.picturae.com/catalog.ttl?page=1' ],
+        });
+        expect((await bindingsStream.toArray()).length > 0).toBeTruthy();
+      });
     });
 
     describe('property paths', () => {

packages/actor-query-source-identify-hypermedia/lib/LinkedRdfSourcesAsyncRdfIterator.ts

@@ -71,8 +71,19 @@ export abstract class LinkedRdfSourcesAsyncRdfIterator extends BufferedIterator<
               const bindingsStream = sourceState.source.queryBindings(this.operation, this.context);
               bindingsStream.getProperty('metadata', (metadata: MetadataBindings) => {
                 metadata.state = new MetadataValidationState();
-                resolve(metadata);
                 bindingsStream.destroy();
+                this.accumulateMetadata(sourceState.metadata, metadata)
+                  .then((accumulatedMetadata) => {
+                    // Also merge fields that were not explicitly accumulated
+                    const returnMetadata = { ...sourceState.metadata, ...metadata, ...accumulatedMetadata };
+                    resolve(returnMetadata);
+                  })
+                  .catch(() => {
+                    resolve({
+                      ...sourceState.metadata,
+                      state: new MetadataValidationState(),
+                    });
+                  });
               });
             })
             .catch(reject);

packages/actor-query-source-identify-hypermedia/test/ActorQuerySourceIdentifyHypermedia-test.ts

@@ -161,6 +161,11 @@ describe('ActorQuerySourceIdentifyHypermedia', () => {
           const bindings = querySource.source.queryBindings(operation, context);
           await expect(new Promise(resolve => bindings.getProperty('metadata', resolve))).resolves
             .toEqual({
+              a: 1,
+              cardinality: {
+                type: 'estimate',
+                value: Number.POSITIVE_INFINITY,
+              },
               state: expect.any(MetadataValidationState),
               firstMeta: true,
             });
@@ -217,7 +222,14 @@ describe('ActorQuerySourceIdentifyHypermedia', () => {
           const { querySource } = await actor.run({ context, querySourceUnidentified });
           const bindings = querySource.source.queryBindings(operation, context);
           await expect(new Promise(resolve => bindings.getProperty('metadata', resolve))).resolves
-            .toEqual({ state: expect.any(MetadataValidationState), firstMeta: true });
+            .toEqual({
+              cardinality: {
+                type: 'estimate',
+                value: Number.POSITIVE_INFINITY,
+              },
+              state: expect.any(MetadataValidationState),
+              firstMeta: true,
+            });
 
           expect(spy).toHaveBeenCalledWith({
             map: expect.anything(),
@@ -343,6 +355,11 @@ describe('ActorQuerySourceIdentifyHypermedia', () => {
           const bindings = querySource.source.queryBindings(operation, context);
           await expect(new Promise(resolve => bindings.getProperty('metadata', resolve))).resolves
             .toEqual({
+              a: 1,
+              cardinality: {
+                type: 'estimate',
+                value: Number.POSITIVE_INFINITY,
+              },
               state: expect.any(MetadataValidationState),
               firstMeta: true,
             });
@@ -425,6 +442,11 @@ describe('ActorQuerySourceIdentifyHypermedia', () => {
           const bindings = querySource.source.queryBindings(operation, context);
           await expect(new Promise(resolve => bindings.getProperty('metadata', resolve))).resolves
             .toEqual({
+              a: 1,
+              cardinality: {
+                type: 'estimate',
+                value: Number.POSITIVE_INFINITY,
+              },
               state: expect.any(MetadataValidationState),
               firstMeta: true,
             });

packages/actor-query-source-identify-hypermedia/test/LinkedRdfSourcesAsyncRdfIterator-test.ts

@@ -182,15 +182,17 @@ describe('LinkedRdfSourcesAsyncRdfIterator', () => {
       jest.spyOn(<any> it, 'startIteratorsForNextUrls');
 
       await expect(new Promise(resolve => it.getProperty('metadata', resolve))).resolves.toEqual({
+        firstPageToken: true,
         state: expect.any(MetadataValidationState),
         next: 'P1',
+        requestedPage: 0,
         subseq: true,
       });
 
       await expect(it).toEqualBindingsStream(data.flat());
     });
 
-    it('handles metadata for a single page before consuming data and handle errors', async() => {
+    it('handles metadata for a single page before consuming data and handle source getter errors', async() => {
       const sourceStateGetterOld = sourceStateGetter;
       let called = false;
       sourceStateGetter = async(link: ILink): Promise<ISourceState> => {
@@ -217,6 +219,24 @@ describe('LinkedRdfSourcesAsyncRdfIterator', () => {
       await expect(it).toEqualBindingsStream(data.flat());
     });
 
+    it('handles metadata for a single page before consuming data and handle accumulateMetadata errors', async() => {
+      data = toBindings([[
+        [ 'a', 'b', 'c' ],
+        [ 'd', 'e', 'f' ],
+        [ 'g', 'h', 'i' ],
+      ]]);
+      const it =
+        new DummyIteratorErrorAccumulate(operation, queryBindingsOptions, context, 'first', sourceStateGetter);
+      jest.spyOn(<any> it, 'startIteratorsForNextUrls');
+
+      await expect(new Promise(resolve => it.getProperty('metadata', resolve))).resolves.toEqual({
+        firstPageToken: true,
+        state: expect.any(MetadataValidationState),
+        next: 'P1',
+        requestedPage: 0,
+      });
+    });
+
     it('handles metadata for a single page after consuming data', async() => {
       data = toBindings([[
         [ 'a', 'b', 'c' ],
@@ -250,8 +270,10 @@ describe('LinkedRdfSourcesAsyncRdfIterator', () => {
       jest.spyOn(<any> it, 'startIteratorsForNextUrls');
 
       await expect(new Promise(resolve => it.getProperty('metadata', resolve))).resolves.toEqual({
+        firstPageToken: true,
         state: expect.any(MetadataValidationState),
         next: 'P1',
+        requestedPage: 0,
         subseq: true,
       });
 

packages/actor-query-source-identify-hypermedia/test/QuerySourceHypermedia-test.ts

@@ -167,6 +167,11 @@ describe('QuerySourceHypermedia', () => {
           }),
         ]);
         await expect(metaPromise).resolves.toEqual({
+          a: 1,
+          cardinality: {
+            type: 'estimate',
+            value: Number.POSITIVE_INFINITY,
+          },
           state: expect.any(MetadataValidationState),
           firstMeta: true,
         });
@@ -291,8 +296,10 @@ describe('QuerySourceHypermedia', () => {
           }),
         ]);
         await expect(metaPromise).resolves.toEqual({
+          a: 2,
           state: expect.any(MetadataValidationState),
-          cardinality: { type: 'exact', value: 1 },
+          cardinality: { type: 'exact', value: 2 },
+          canContainUndefs: false,
           firstMeta: true,
         });
       });
@@ -731,6 +738,7 @@ describe('QuerySourceHypermedia', () => {
 
         expect(it1Meta).toHaveBeenCalledTimes(4);
         expect(it1Meta).toHaveBeenNthCalledWith(1, {
+          a: 1,
           state: expect.any(MetadataValidationState),
           firstMeta: true,
           cardinality: { type: 'exact', value: 2 },