Feature: Allow page_size max value 100 (#15292)

Feature: Allow page_size max value 100

api/apps/restful_apis/agent_api.py

@@ -58,6 +58,7 @@ from api.utils.api_utils import (
     server_error_response,
     validate_request,
 )
+from api.utils.pagination_utils import validate_rest_api_page_size
 from common import settings
 from common.ssrf_guard import assert_host_is_safe
 from common.constants import RetCode
@@ -349,7 +350,7 @@ def list_agent_sessions(agent_id, tenant_id):
     session_id = request.args.get("id")
     user_id = request.args.get("user_id")
     page_number = int(request.args.get("page", 1))
-    items_per_page = int(request.args.get("page_size", 30))
+    items_per_page = validate_rest_api_page_size(int(request.args.get("page_size", 30)))
     keywords = request.args.get("keywords")
     from_date = request.args.get("from_date")
     to_date = request.args.get("to_date")
@@ -517,7 +518,7 @@ def list_agents(tenant_id):
     tags = [item for item in request.args.get("tags", "").strip().split(",") if item]
 
     page_number = int(request.args.get("page", 0))
-    items_per_page = int(request.args.get("page_size", 0))
+    items_per_page = validate_rest_api_page_size(int(request.args.get("page_size", 0)))
     order_by = request.args.get("orderby", "create_time")
     desc = str(request.args.get("desc", "true")).lower() != "false"
     tenants = TenantService.get_joined_tenants_by_user_id(tenant_id)

api/apps/restful_apis/chat_api.py

@@ -45,6 +45,7 @@ from api.utils.api_utils import (
     server_error_response,
     validate_request,
 )
+from api.utils.pagination_utils import validate_rest_api_page_size
 from api.utils.tenant_utils import ensure_tenant_model_id_for_params
 from common.constants import LLMType, RetCode, StatusEnum
 from common import settings
@@ -419,7 +420,7 @@ async def list_chats():
 
     try:
         page_number = int(request.args.get("page", 0))
-        items_per_page = int(request.args.get("page_size", 0))
+        items_per_page = validate_rest_api_page_size(int(request.args.get("page_size", 0)))
 
         tenants = TenantService.get_joined_tenants_by_user_id(current_user.id)
         authorized_owner_ids = {member["tenant_id"] for member in tenants}
@@ -769,7 +770,7 @@ async def list_sessions(chat_id):
                 code=RetCode.AUTHENTICATION_ERROR,
             )
         page_number = int(request.args.get("page", 1))
-        items_per_page = int(request.args.get("page_size", 30))
+        items_per_page = validate_rest_api_page_size(int(request.args.get("page_size", 30)))
         orderby = request.args.get("orderby", "create_time")
         desc = request.args.get("desc", "true").lower() != "false"
         session_id = request.args.get("id")

api/apps/restful_apis/chunk_api.py

@@ -46,6 +46,7 @@ from api.utils.api_utils import (
     server_error_response,
     token_required,
 )
+from api.utils.pagination_utils import validate_rest_api_page_size
 from api.utils.image_utils import store_chunk_image
 from api.utils.reference_metadata_utils import (
     enrich_chunks_with_document_metadata,
@@ -244,7 +245,7 @@ async def retrieval_test(tenant_id):
     if "question" not in req:
         return get_error_data_result("`question` is required.")
     page = int(req.get("page", 1))
-    size = int(req.get("page_size", 30))
+    size = validate_rest_api_page_size(int(req.get("page_size", 30)))
     question = req["question"].strip() if isinstance(req["question"], str) else req["question"]
     if not question:
         return get_result(data={"total": 0, "chunks": [], "doc_aggs": {}})
@@ -365,7 +366,7 @@ async def list_chunks(tenant_id, dataset_id, document_id):
     doc = doc[0]
     req = request.args
     page = int(req.get("page", 1))
-    size = int(req.get("page_size", 30))
+    size = validate_rest_api_page_size(int(req.get("page_size", 30)))
     question = req.get("keywords", "")
     query = {
         "doc_ids": [document_id],

api/apps/restful_apis/connector_api.py

@@ -27,6 +27,7 @@ from google_auth_oauthlib.flow import Flow
 from api.db import InputType
 from api.db.services.connector_service import ConnectorService, SyncLogsService
 from api.utils.api_utils import get_data_error_result, get_json_result, get_request_json, validate_request
+from api.utils.pagination_utils import validate_rest_api_page_size
 from common.constants import RetCode, TaskStatus
 from common.data_source.config import GOOGLE_DRIVE_WEB_OAUTH_REDIRECT_URI, GMAIL_WEB_OAUTH_REDIRECT_URI, BOX_WEB_OAUTH_REDIRECT_URI, DocumentSource
 from common.data_source.google_util.constant import WEB_OAUTH_POPUP_TEMPLATE, GOOGLE_SCOPES
@@ -140,7 +141,11 @@ def list_logs(connector_id):
         return _connector_auth_error(connector_id, current_user.id)
 
     req = request.args.to_dict(flat=True)
-    arr, total = SyncLogsService.list_sync_tasks(connector_id, int(req.get("page", 1)), int(req.get("page_size", 15)))
+    arr, total = SyncLogsService.list_sync_tasks(
+        connector_id,
+        int(req.get("page", 1)),
+        validate_rest_api_page_size(int(req.get("page_size", 15))),
+    )
     return get_json_result(data={"total": total, "logs": arr})
 
 

api/apps/restful_apis/dataset_api.py

@@ -20,6 +20,7 @@ from quart import request
 from common.constants import RetCode
 from api.apps import login_required, current_user
 from api.utils.api_utils import get_error_argument_result, get_error_data_result, get_json_result, get_result, add_tenant_id_to_kwargs
+from api.utils.pagination_utils import validate_rest_api_page_size
 from api.utils.validation_utils import (
     CreateDatasetReq,
     DeleteDatasetReq,
@@ -665,7 +666,7 @@ async def check_embedding(tenant_id, dataset_id):
 def list_ingestion_logs(tenant_id, dataset_id):
     try:
         page = int(request.args.get("page", 0))
-        page_size = int(request.args.get("page_size", 0))
+        page_size = validate_rest_api_page_size(int(request.args.get("page_size", 0)))
         orderby = request.args.get("orderby", "create_time")
         desc = request.args.get("desc", "true").lower() != "false"
         operation_status = request.args.getlist("operation_status")

api/apps/restful_apis/document_api.py

@@ -41,6 +41,7 @@ from api.common.check_team_permission import check_kb_team_permission
 from api.db.services.task_service import TaskService, cancel_all_task_of
 from api.utils.api_utils import construct_json_result, get_data_error_result, get_error_data_result, get_result, get_json_result, \
     server_error_response, add_tenant_id_to_kwargs, get_request_json, get_error_argument_result, check_duplicate_ids
+from api.utils.pagination_utils import validate_rest_api_page_size
 from api.utils.validation_utils import (
     UpdateDocumentReq, format_validation_error_message, validate_and_parse_json_request, DeleteDocumentReq,
 )
@@ -795,7 +796,7 @@ def _get_docs_with_request(req, dataset_id:str):
     q = req.args
 
     page = int(q.get("page", 1))
-    page_size = int(q.get("page_size", 30))
+    page_size = validate_rest_api_page_size(int(q.get("page_size", 30)))
 
     orderby = q.get("orderby", "create_time")
     desc = str(q.get("desc", "true")).strip().lower() != "false"

api/apps/restful_apis/mcp_api.py

@@ -21,6 +21,7 @@ from api.db.db_models import MCPServer
 from api.db.services.mcp_server_service import MCPServerService
 from api.db.services.user_service import TenantService
 from api.utils.api_utils import get_data_error_result, get_json_result, get_mcp_tools, get_request_json, server_error_response, validate_request
+from api.utils.pagination_utils import validate_rest_api_page_size
 from api.utils.web_utils import get_float, safe_json_parse
 from common.constants import VALID_MCP_SERVER_TYPES
 from common.mcp_tool_call_conn import MCPToolCallSession, close_multiple_mcp_toolcall_sessions
@@ -71,7 +72,7 @@ def _assert_mcp_url_is_safe(url, invalid_message: str = "Invalid url.") -> tuple
 async def list_mcp() -> Response:
     keywords = request.args.get("keywords", "")
     page_number = int(request.args.get("page", 0))
-    items_per_page = int(request.args.get("page_size", 0))
+    items_per_page = validate_rest_api_page_size(int(request.args.get("page_size", 0)))
     orderby = request.args.get("orderby", "create_time")
     if request.args.get("desc", "true").lower() == "false":
         desc = False

api/apps/restful_apis/memory_api.py

@@ -24,6 +24,7 @@ from api.apps import login_required, current_user
 from api.utils.api_utils import validate_request, get_request_json, get_error_argument_result, get_json_result
 from api.apps.services import memory_api_service
 from api.utils.tenant_utils import ensure_tenant_model_id_for_params
+from api.utils.pagination_utils import validate_rest_api_page_size
 
 
 @manager.route("/memories", methods=["POST"])  # noqa: F821
@@ -134,7 +135,7 @@ async def list_memory():
     }
     keywords = request.args.get("keywords")
     page = int(request.args.get("page", 1))
-    page_size = int(request.args.get("page_size", 50))
+    page_size = validate_rest_api_page_size(int(request.args.get("page_size", 50)))
     try:
         res = await memory_api_service.list_memory(filter_params, keywords, page, page_size)
         return get_json_result(message=True, data=res)
@@ -167,7 +168,7 @@ async def get_memory_messages(memory_id):
     keywords = args.get("keywords", "")
     keywords = keywords.strip()
     page = int(args.get("page", 1))
-    page_size = int(args.get("page_size", 50))
+    page_size = validate_rest_api_page_size(int(args.get("page_size", 50)))
     try:
         res = await memory_api_service.get_memory_messages(
             memory_id, agent_ids, keywords, page, page_size

api/apps/restful_apis/search_api.py

@@ -30,6 +30,7 @@ from api.db.services.user_service import TenantService, UserTenantService
 from common.misc_utils import get_uuid
 from common.constants import RetCode, StatusEnum
 from api.utils.api_utils import get_data_error_result, get_json_result, get_request_json, server_error_response, validate_request
+from api.utils.pagination_utils import validate_rest_api_page_size
 
 
 def _full_text_weight(vector_similarity_weight):
@@ -77,7 +78,7 @@ async def create():
 def list_searches():
     keywords = request.args.get("keywords", "")
     page_number = int(request.args.get("page", 0))
-    items_per_page = int(request.args.get("page_size", 0))
+    items_per_page = validate_rest_api_page_size(int(request.args.get("page_size", 0)))
     orderby = request.args.get("orderby", "create_time")
     desc = request.args.get("desc", "true").lower() != "false"
     owner_ids = request.args.getlist("owner_ids")