Skip to content

armgit / dify

Public Imported from GitHub
0

Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.

fix(api): correct dataset and document list has_more pagination (#41784)

Zhu Lei5355edefe366
diff --git a/api/controllers/service_api/dataset/dataset.py b/api/controllers/service_api/dataset/dataset.py
index 09a404b9eff..32035f539b1 100644
--- a/api/controllers/service_api/dataset/dataset.py
+++ b/api/controllers/service_api/dataset/dataset.py
@@ -430,10 +430,11 @@ class DatasetListApi(DatasetApiResource):
query_params["tag_ids"] = request.args.getlist("tag_ids")
query = DatasetListQuery.model_validate(query_params)
# provider = request.args.get("provider", default="vendor")
+ effective_limit = min(query.limit, 100)
datasets, total = DatasetService.get_datasets(
query.page,
- query.limit,
+ effective_limit,
session,
tenant_id,
current_user,
@@ -467,8 +468,8 @@ class DatasetListApi(DatasetApiResource):
item["embedding_available"] = True
response = {
"data": data,
- "has_more": len(datasets) == query.limit,
- "limit": query.limit,
+ "has_more": query.page * effective_limit < total,
+ "limit": effective_limit,
"total": total,
"page": query.page,
}
diff --git a/api/controllers/service_api/dataset/document.py b/api/controllers/service_api/dataset/document.py
index 5049c2a3782..a2288a99c43 100644
--- a/api/controllers/service_api/dataset/document.py
+++ b/api/controllers/service_api/dataset/document.py
@@ -1015,8 +1015,9 @@ class DocumentListApi(DatasetApiResource):
query = query.order_by(desc(Document.created_at), desc(Document.position))
+ effective_limit = min(query_params.limit, 100)
paginated_documents = paginate_query(
- query, session=session, page=query_params.page, per_page=query_params.limit, max_per_page=100
+ query, session=session, page=query_params.page, per_page=effective_limit, max_per_page=100
)
documents = paginated_documents.items
@@ -1029,8 +1030,8 @@ class DocumentListApi(DatasetApiResource):
response = {
"data": document_responses(documents, session=session),
- "has_more": len(documents) == query_params.limit,
- "limit": query_params.limit,
+ "has_more": query_params.page * effective_limit < paginated_documents.total,
+ "limit": effective_limit,
"total": paginated_documents.total,
"page": query_params.page,
}
diff --git a/api/tests/unit_tests/controllers/service_api/dataset/test_dataset_apis.py b/api/tests/unit_tests/controllers/service_api/dataset/test_dataset_apis.py
index 581f95d8ce5..8b3c7d42c79 100644
--- a/api/tests/unit_tests/controllers/service_api/dataset/test_dataset_apis.py
+++ b/api/tests/unit_tests/controllers/service_api/dataset/test_dataset_apis.py
@@ -257,6 +257,66 @@ class TestDatasetListApiGet:
False,
)
+ @patch("controllers.service_api.dataset.dataset.create_plugin_provider_manager")
+ @patch("controllers.service_api.dataset.dataset.DatasetService")
+ def test_list_datasets_has_more_false_on_last_page_exact_limit(
+ self,
+ mock_dataset_svc: MagicMock,
+ mock_provider_mgr: MagicMock,
+ app: Flask,
+ account: Account,
+ tenant: Tenant,
+ controller_session: Session,
+ ) -> None:
+ """A full last page must set has_more false instead of forcing another fetch."""
+ from controllers.service_api.dataset.dataset import DatasetListApi
+
+ page_size = 20
+ dataset = make_dataset(controller_session, tenant, account)
+ mock_dataset_svc.get_datasets.return_value = ([dataset] * page_size, page_size)
+ mock_provider_mgr.return_value.get_configurations.return_value.get_models.return_value = list[object]()
+
+ with app.test_request_context(f"/datasets?page=1&limit={page_size}", method="GET"):
+ api = DatasetListApi()
+ response, status = unwrap(api.get)(api, controller_session, tenant_id=tenant.id)
+
+ assert status == 200
+ assert response["has_more"] is False
+ assert response["limit"] == page_size
+ assert response["total"] == page_size
+ assert response["page"] == 1
+
+ @patch("controllers.service_api.dataset.dataset.create_plugin_provider_manager")
+ @patch("controllers.service_api.dataset.dataset.DatasetService")
+ def test_list_datasets_has_more_true_when_limit_exceeds_cap(
+ self,
+ mock_dataset_svc: MagicMock,
+ mock_provider_mgr: MagicMock,
+ app: Flask,
+ account: Account,
+ tenant: Tenant,
+ controller_session: Session,
+ ) -> None:
+ """limit>100 still reports remaining rows after the server cap of 100."""
+ from controllers.service_api.dataset.dataset import DatasetListApi
+
+ returned_count = 100
+ total = 150
+ dataset = make_dataset(controller_session, tenant, account)
+ mock_dataset_svc.get_datasets.return_value = ([dataset] * returned_count, total)
+ mock_provider_mgr.return_value.get_configurations.return_value.get_models.return_value = list[object]()
+
+ with app.test_request_context("/datasets?page=1&limit=200", method="GET"):
+ api = DatasetListApi()
+ response, status = unwrap(api.get)(api, controller_session, tenant_id=tenant.id)
+
+ assert status == 200
+ assert response["has_more"] is True
+ assert response["limit"] == 100
+ assert response["total"] == total
+ assert response["page"] == 1
+ assert mock_dataset_svc.get_datasets.call_args.args[1] == 100
+
class TestDatasetListApiPost:
"""Test suite for DatasetListApi.post() endpoint."""
diff --git a/api/tests/unit_tests/controllers/service_api/dataset/test_document.py b/api/tests/unit_tests/controllers/service_api/dataset/test_document.py
index a2e8d760feb..c297b95987e 100644
--- a/api/tests/unit_tests/controllers/service_api/dataset/test_document.py
+++ b/api/tests/unit_tests/controllers/service_api/dataset/test_document.py
@@ -1189,6 +1189,76 @@ class TestDocumentListApi(SQLiteControllerTest):
assert "data_source_info_dict" not in response["data"][0]
assert "doc_metadata_details" not in response["data"][0]
+ @patch("controllers.service_api.dataset.document.paginate_query")
+ @patch("controllers.service_api.dataset.document.DocumentService")
+ def test_list_documents_has_more_false_on_last_page_exact_limit(
+ self, mock_doc_svc, mock_paginate, app: Flask, mock_tenant, mock_dataset
+ ):
+ """A full last page must set has_more false instead of forcing another fetch."""
+ self._persist_dataset(mock_dataset)
+ page_size = 20
+ documents = [
+ make_serializable_document(
+ id=f"doc-{index}",
+ name=f"Document {index}",
+ tenant_id=mock_tenant,
+ dataset_id=mock_dataset.id,
+ )
+ for index in range(page_size)
+ ]
+ mock_paginate.return_value = _PaginationRecord(items=documents, total=page_size)
+ mock_doc_svc.enrich_documents_with_summary_index_status.return_value = None
+
+ with app.test_request_context(
+ f"/datasets/{mock_dataset.id}/documents?page=1&limit={page_size}",
+ method="GET",
+ ):
+ api = DocumentListApi()
+ response = inspect.unwrap(type(api).get)(
+ api, self.session, tenant_id=mock_tenant, dataset_id=mock_dataset.id
+ )
+
+ assert response["has_more"] is False
+ assert response["limit"] == page_size
+ assert response["total"] == page_size
+ assert response["page"] == 1
+
+ @patch("controllers.service_api.dataset.document.paginate_query")
+ @patch("controllers.service_api.dataset.document.DocumentService")
+ def test_list_documents_has_more_true_when_limit_exceeds_cap(
+ self, mock_doc_svc, mock_paginate, app: Flask, mock_tenant, mock_dataset
+ ):
+ """limit>100 still reports remaining rows after the server cap of 100."""
+ self._persist_dataset(mock_dataset)
+ returned_count = 100
+ total = 150
+ documents = [
+ make_serializable_document(
+ id=f"doc-{index}",
+ name=f"Document {index}",
+ tenant_id=mock_tenant,
+ dataset_id=mock_dataset.id,
+ )
+ for index in range(returned_count)
+ ]
+ mock_paginate.return_value = _PaginationRecord(items=documents, total=total)
+ mock_doc_svc.enrich_documents_with_summary_index_status.return_value = None
+
+ with app.test_request_context(
+ f"/datasets/{mock_dataset.id}/documents?page=1&limit=200",
+ method="GET",
+ ):
+ api = DocumentListApi()
+ response = inspect.unwrap(type(api).get)(
+ api, self.session, tenant_id=mock_tenant, dataset_id=mock_dataset.id
+ )
+
+ assert response["has_more"] is True
+ assert response["limit"] == 100
+ assert response["total"] == total
+ assert response["page"] == 1
+ assert mock_paginate.call_args.kwargs["per_page"] == 100
+
def test_list_documents_dataset_not_found(self, app: Flask, mock_tenant, mock_dataset):
"""Test 404 when dataset not found."""
# Arrange