Refactor: Consolidation WEB API & HTTP API for document delete api (#14254)

### What problem does this PR solve? Before consolidation Web API: POST /v1/document/rm Http API - DELETE /api/v1/datasets/<dataset_id>/documents After consolidation, Restful API -- DELETE /api/v1/datasets/<dataset_id>/documents ### Type of change - [x] Refactoring
2026-06-29 15:31:05 +08:00 · 2026-04-22 10:49:52 +08:00
parent 6baf74afc1
commit 3d8a82c0aa
15 changed files with 178 additions and 278 deletions
--- a/api/apps/document_app.py
+++ b/api/apps/document_app.py
@@ -319,27 +319,6 @@ async def change_status():
    return get_json_result(data=result)


-@manager.route("/rm", methods=["POST"])  # noqa: F821
-@login_required
-@validate_request("doc_id")
-async def rm():
-    req = await get_request_json()
-    doc_ids = req["doc_id"]
-    if isinstance(doc_ids, str):
-        doc_ids = [doc_ids]
-
-    for doc_id in doc_ids:
-        if not DocumentService.accessible4deletion(doc_id, current_user.id):
-            return get_json_result(data=False, message="No authorization.", code=RetCode.AUTHENTICATION_ERROR)
-
-    errors = await thread_pool_exec(FileService.delete_docs, doc_ids, current_user.id)
-
-    if errors:
-        return get_json_result(data=False, message=errors, code=RetCode.SERVER_ERROR)
-
-    return get_json_result(data=True)
-
-
@manager.route("/run", methods=["POST"])  # noqa: F821
@login_required
@validate_request("doc_ids", "run")
--- a/api/apps/restful_apis/document_api.py
+++ b/api/apps/restful_apis/document_api.py
@@ -27,14 +27,17 @@ from api.constants import IMG_BASE64_PREFIX
 from api.db import VALID_FILE_TYPES
 from api.db.services.doc_metadata_service import DocMetadataService
 from api.db.services.document_service import DocumentService
+from api.db.services.file_service import FileService
 from api.db.services.knowledgebase_service import KnowledgebaseService
+from api.common.check_team_permission import check_kb_team_permission
 from api.utils.api_utils import get_data_error_result, get_error_data_result, get_result, get_json_result, \
-    server_error_response, add_tenant_id_to_kwargs, get_request_json
+    server_error_response, add_tenant_id_to_kwargs, get_request_json, get_error_argument_result, check_duplicate_ids
 from api.utils.validation_utils import (
-    UpdateDocumentReq, format_validation_error_message,
+    UpdateDocumentReq, format_validation_error_message, validate_and_parse_json_request, DeleteDocumentReq,
 )
 from common.constants import RetCode
 from common.metadata_utils import convert_conditions, meta_filter, turn2jsonschema
+from common.misc_utils import thread_pool_exec

@manager.route("/datasets/<dataset_id>/documents/<document_id>", methods=["PATCH"]) # noqa: F821
@login_required
@@ -260,9 +263,7 @@ async def upload_document(dataset_id, tenant_id):
                    description: Processing status.
    """
    from api.constants import FILE_NAME_LEN_LIMIT
-    from api.common.check_team_permission import check_kb_team_permission
    from api.db.services.file_service import FileService
-    from common.misc_utils import thread_pool_exec
    
    form = await request.form
    files = await request.files
@@ -674,6 +675,89 @@ def _parse_doc_id_filter_with_metadata(req, kb_id):
    return RetCode.SUCCESS, "", list(doc_ids_filter) if doc_ids_filter is not None else [], return_empty_metadata


+@manager.route("/datasets/<dataset_id>/documents", methods=["DELETE"])  # noqa: F821
+@login_required
+@add_tenant_id_to_kwargs
+async def delete_documents(tenant_id, dataset_id):
+    """
+    Delete documents from a dataset.
+    ---
+    tags:
+      - Documents
+    security:
+      - ApiKeyAuth: []
+    parameters:
+      - in: path
+        name: dataset_id
+        type: string
+        required: true
+        description: ID of the dataset containing the documents.
+      - in: header
+        name: Authorization
+        type: string
+        required: true
+        description: Bearer token for authentication.
+      - in: body
+        name: body
+        description: Document deletion parameters.
+        required: true
+        schema:
+          type: object
+          properties:
+            ids:
+              type: array or null
+              items:
+                type: string
+              description: |
+                Specifies the documents to delete:
+                - An array of IDs, only the specified documents will be deleted.
+            delete_all:
+              type: boolean
+              default: false
+              description: Whether to delete all documents in the dataset.
+    responses:
+      200:
+        description: Successful operation.
+        schema:
+          type: object
+    """
+    req, err = await validate_and_parse_json_request(request, DeleteDocumentReq)
+    if err is not None or req is None:
+        return get_error_argument_result(err)
+
+    try:
+        # Validate dataset exists and user has permission
+        if not KnowledgebaseService.accessible(kb_id=dataset_id, user_id=tenant_id):
+            return get_error_data_result(message=f"You don't own the dataset {dataset_id}. ")
+
+        # Get documents to delete
+        doc_ids = req.get("ids") or []
+        delete_all = req.get("delete_all", False)
+        if not delete_all and len(doc_ids) == 0:
+            return get_error_data_result(message=f"should either provide doc ids or set delete_all(true), dataset: {dataset_id}. ")
+
+        if len(doc_ids) > 0 and delete_all:
+            return get_error_data_result(message=f"should not provide both doc ids and delete_all(true), dataset: {dataset_id}. ")
+        if delete_all:
+            doc_ids = [doc.id for doc in DocumentService.query(kb_id=dataset_id)]
+
+        # make sure each id is unique
+        unique_doc_ids, duplicate_messages = check_duplicate_ids(doc_ids, "document")
+        if duplicate_messages:
+            logging.warning(f"duplicate_messages:{duplicate_messages}")
+        else:
+            doc_ids = unique_doc_ids
+
+        # Delete documents using existing FileService.delete_docs
+        errors = await thread_pool_exec(FileService.delete_docs, doc_ids, tenant_id)
+
+        if errors:
+            return get_error_data_result(message=str(errors))
+
+        return get_result(data={"deleted": len(doc_ids)})
+    except Exception as e:
+        logging.exception(e)
+        return get_error_data_result(message="Internal server error")
 def _aggregate_filters(docs):
    """Aggregate filter options from a list of documents.

--- a/api/apps/sdk/doc.py
+++ b/api/apps/sdk/doc.py
@@ -21,12 +21,11 @@ import xxhash
 from pydantic import BaseModel, Field, validator
 from quart import request, send_file

-from api.db.db_models import APIToken, Document, File, Task
+from api.db.db_models import APIToken, Document, Task
 from api.db.joint_services.tenant_model_service import get_model_config_by_id, get_model_config_by_type_and_name, get_tenant_default_model_by_type
 from api.db.services.doc_metadata_service import DocMetadataService
 from api.db.services.document_service import DocumentService
 from api.db.services.file2document_service import File2DocumentService
-from api.db.services.file_service import FileService
 from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.llm_service import LLMBundle
 from api.db.services.task_service import TaskService, cancel_all_task_of, queue_tasks
@@ -34,7 +33,7 @@ from api.db.services.tenant_llm_service import TenantLLMService
 from api.utils.api_utils import check_duplicate_ids, construct_json_result, get_error_data_result, get_request_json, get_result, server_error_response, token_required
 from api.utils.image_utils import store_chunk_image
 from common import settings
-from common.constants import FileSource, LLMType, ParserType, RetCode, TaskStatus
+from common.constants import LLMType, ParserType, RetCode, TaskStatus
 from common.metadata_utils import convert_conditions, meta_filter
 from common.misc_utils import thread_pool_exec
 from common.string_utils import is_content_empty, remove_redundant_spaces
@@ -209,120 +208,6 @@ async def metadata_batch_update(dataset_id, tenant_id):
    return get_result(data={"updated": updated, "matched_docs": len(target_doc_ids)})


-@manager.route("/datasets/<dataset_id>/documents", methods=["DELETE"])  # noqa: F821
-@token_required
-async def delete(tenant_id, dataset_id):
-    """
-    Delete documents from a dataset.
-    ---
-    tags:
-      - Documents
-    security:
-      - ApiKeyAuth: []
-    parameters:
-      - in: path
-        name: dataset_id
-        type: string
-        required: true
-        description: ID of the dataset.
-      - in: body
-        name: body
-        description: Document deletion parameters.
-        required: true
-        schema:
-          type: object
-          properties:
-            ids:
-              type: array
-              items:
-                type: string
-              description: |
-                List of document IDs to delete.
-                If omitted, `null`, or an empty array is provided, no documents will be deleted.
-      - in: header
-        name: Authorization
-        type: string
-        required: true
-        description: Bearer token for authentication.
-    responses:
-      200:
-        description: Documents deleted successfully.
-        schema:
-          type: object
-    """
-    if not KnowledgebaseService.accessible(kb_id=dataset_id, user_id=tenant_id):
-        return get_error_data_result(message=f"You don't own the dataset {dataset_id}. ")
-    req = await get_request_json()
-    if not req:
-        return get_result()
-
-    doc_ids = req.get("ids")
-    if not doc_ids:
-        if req.get("delete_all") is True:
-            doc_ids = [doc.id for doc in DocumentService.query(kb_id=dataset_id)]
-            if not doc_ids:
-                return get_result()
-        else:
-            return get_result()
-
-    doc_list = doc_ids
-
-    unique_doc_ids, duplicate_messages = check_duplicate_ids(doc_list, "document")
-    doc_list = unique_doc_ids
-
-    root_folder = FileService.get_root_folder(tenant_id)
-    pf_id = root_folder["id"]
-    FileService.init_knowledgebase_docs(pf_id, tenant_id)
-    errors = ""
-    not_found = []
-    success_count = 0
-    for doc_id in doc_list:
-        try:
-            e, doc = DocumentService.get_by_id(doc_id)
-            if not e:
-                not_found.append(doc_id)
-                continue
-            tenant_id = DocumentService.get_tenant_id(doc_id)
-            if not tenant_id:
-                return get_error_data_result(message="Tenant not found!")
-
-            b, n = File2DocumentService.get_storage_address(doc_id=doc_id)
-
-            if not DocumentService.remove_document(doc, tenant_id):
-                return get_error_data_result(message="Database error (Document removal)!")
-
-            f2d = File2DocumentService.get_by_document_id(doc_id)
-            FileService.filter_delete(
-                [
-                    File.source_type == FileSource.KNOWLEDGEBASE,
-                    File.id == f2d[0].file_id,
-                ]
-            )
-            File2DocumentService.delete_by_document_id(doc_id)
-
-            settings.STORAGE_IMPL.rm(b, n)
-            success_count += 1
-        except Exception as e:
-            errors += str(e)
-
-    if not_found:
-        return get_result(message=f"Documents not found: {not_found}", code=RetCode.DATA_ERROR)
-
-    if errors:
-        return get_result(message=errors, code=RetCode.SERVER_ERROR)
-
-    if duplicate_messages:
-        if success_count > 0:
-            return get_result(
-                message=f"Partially deleted {success_count} datasets with {len(duplicate_messages)} errors",
-                data={"success_count": success_count, "errors": duplicate_messages},
-            )
-        else:
-            return get_error_data_result(message=";".join(duplicate_messages))
-
-    return get_result()
-
-
 DOC_STOP_PARSING_INVALID_STATE_MESSAGE = "Can't stop parsing document that has not started or already completed"
 DOC_STOP_PARSING_INVALID_STATE_ERROR_CODE = "DOC_STOP_PARSING_INVALID_STATE"

--- a/api/utils/validation_utils.py
+++ b/api/utils/validation_utils.py
@@ -818,6 +818,9 @@ class DeleteReq(Base):
 class DeleteDatasetReq(DeleteReq): ...


+class DeleteDocumentReq(DeleteReq): ...
+
+
 class BaseListReq(BaseModel):
    model_config = ConfigDict(extra="forbid")

--- a/test/testcases/test_http_api/test_file_management_within_dataset/test_delete_documents.py
+++ b/test/testcases/test_http_api/test_file_management_within_dataset/test_delete_documents.py
@@ -26,11 +26,11 @@ class TestAuthorization:
    @pytest.mark.parametrize(
        "invalid_auth, expected_code, expected_message",
        [
-            (None, 0, "`Authorization` can't be empty"),
+            (None, 401, "<Unauthorized '401: Unauthorized'>"),
            (
                RAGFlowHttpApiAuth(INVALID_API_TOKEN),
-                109,
-                "Authentication error: API key is invalid!",
+                401,
+                "<Unauthorized '401: Unauthorized'>",
            ),
        ],
    )
@@ -45,19 +45,19 @@ class TestDocumentsDeletion:
    @pytest.mark.parametrize(
        "payload, expected_code, expected_message, remaining",
        [
-            (None, 0, "", 3),
-            ({"ids": []}, 0, "", 3),
-            ({"ids": ["invalid_id"]}, 102, "Documents not found: ['invalid_id']", 3),
+            ({}, 102, "should either provide doc ids or set delete_all(true), dataset", 3),
+            ({"ids": []}, 102, "should either provide doc ids or set delete_all(true), dataset", 3),
+            ({"ids": ["invalid_id"]}, 101, "Field: <ids> - Message: <Invalid UUID1 format> - Value: <['invalid_id']>", 3),
            (
                {"ids": ["\n!?。；！？\"'"]},
-                102,
-                """Documents not found: [\'\\n!?。；！？"\\\'\']""",
+                101,
+                "Field: <ids> - Message: <Invalid UUID1 format> - Value:",
                3,
            ),
            (
                "not json",
-                100,
-                "AttributeError(\"'str' object has no attribute 'get'\")",
+                101,
+                "Invalid request payload: expected object, got str",
                3,
            ),
            (lambda r: {"ids": r[:1]}, 0, "", 2),
@@ -79,7 +79,7 @@ class TestDocumentsDeletion:
        res = delete_documents(HttpApiAuth, dataset_id, payload)
        assert res["code"] == expected_code
        if res["code"] != 0:
-            assert res["message"] == expected_message
+            assert expected_message in res["message"]

        res = list_documents(HttpApiAuth, dataset_id)
        assert len(res["data"]["docs"]) == remaining
@@ -117,12 +117,12 @@ class TestDocumentsDeletion:
        if callable(payload):
            payload = payload(document_ids)
        res = delete_documents(HttpApiAuth, dataset_id, payload)
-        assert res["code"] == 102
-        assert res["message"] == "Documents not found: ['invalid_id']"
+        assert res["code"] == 101
+        assert "Field: <ids> - Message: <Invalid UUID1 format> - Value" in res["message"]

        res = list_documents(HttpApiAuth, dataset_id)
-        assert len(res["data"]["docs"]) == 0
-        assert res["data"]["total"] == 0
+        assert len(res["data"]["docs"]) == 3
+        assert res["data"]["total"] == 3

    @pytest.mark.p2
    def test_repeated_deletion(self, HttpApiAuth, add_documents_func):
@@ -132,19 +132,18 @@ class TestDocumentsDeletion:

        res = delete_documents(HttpApiAuth, dataset_id, {"ids": document_ids})
        assert res["code"] == 102
-        assert "Documents not found" in res["message"]
+        assert "Document not found" in res["message"]

    @pytest.mark.p2
    def test_duplicate_deletion(self, HttpApiAuth, add_documents_func):
        dataset_id, document_ids = add_documents_func
        res = delete_documents(HttpApiAuth, dataset_id, {"ids": document_ids + document_ids})
-        assert res["code"] == 0
-        assert "Duplicate document ids" in res["data"]["errors"][0]
-        assert res["data"]["success_count"] == 3
+        assert res["code"] == 101, res
+        assert "Field: <ids> - Message: <Duplicate ids:" in res["message"]

        res = list_documents(HttpApiAuth, dataset_id)
-        assert len(res["data"]["docs"]) == 0
-        assert res["data"]["total"] == 0
+        assert len(res["data"]["docs"]) == 3
+        assert res["data"]["total"] == 3


@pytest.mark.p3
--- a/test/testcases/test_http_api/test_file_management_within_dataset/test_doc_sdk_routes_unit.py
+++ b/test/testcases/test_http_api/test_file_management_within_dataset/test_doc_sdk_routes_unit.py
@@ -478,46 +478,6 @@ class TestDocRoutesUnit:
        assert res["data"]["matched_docs"] == 1


-    def test_delete_branches(self, monkeypatch):
-        module = _load_doc_module(monkeypatch)
-        monkeypatch.setattr(module.KnowledgebaseService, "accessible", lambda **_kwargs: False)
-        res = _run(module.delete.__wrapped__("tenant-1", "ds-1"))
-        assert "don't own the dataset" in res["message"]
-
-        monkeypatch.setattr(module.KnowledgebaseService, "accessible", lambda **_kwargs: True)
-        monkeypatch.setattr(module, "get_request_json", lambda: _AwaitableValue({}))
-        res = _run(module.delete.__wrapped__("tenant-1", "ds-1"))
-        assert res["code"] == module.RetCode.SUCCESS
-
-        monkeypatch.setattr(module, "get_request_json", lambda: _AwaitableValue({"ids": ["doc-1"]}))
-        monkeypatch.setattr(module, "check_duplicate_ids", lambda ids, _kind: (ids, []))
-        monkeypatch.setattr(module.FileService, "get_root_folder", lambda _tenant: {"id": "pf-1"})
-        monkeypatch.setattr(module.FileService, "init_knowledgebase_docs", lambda *_args, **_kwargs: None)
-        monkeypatch.setattr(module.DocumentService, "get_by_id", lambda _id: (True, _DummyDoc()))
-        monkeypatch.setattr(module.DocumentService, "get_tenant_id", lambda _id: None)
-        res = _run(module.delete.__wrapped__("tenant-1", "ds-1"))
-        assert res["message"] == "Tenant not found!"
-
-        monkeypatch.setattr(module.DocumentService, "get_tenant_id", lambda _id: "tenant-1")
-        monkeypatch.setattr(module.File2DocumentService, "get_storage_address", lambda **_kwargs: ("b", "n"))
-        monkeypatch.setattr(module.DocumentService, "remove_document", lambda *_args, **_kwargs: False)
-        res = _run(module.delete.__wrapped__("tenant-1", "ds-1"))
-        assert "Document removal" in res["message"]
-
-        def _raise_get_by_id(_id):
-            raise RuntimeError("boom")
-
-        monkeypatch.setattr(module.DocumentService, "get_by_id", _raise_get_by_id)
-        res = _run(module.delete.__wrapped__("tenant-1", "ds-1"))
-        assert res["code"] == module.RetCode.SERVER_ERROR
-        assert "boom" in res["message"]
-
-        monkeypatch.setattr(module, "check_duplicate_ids", lambda _ids, _kind: ([], ["Duplicate document ids: doc-1"]))
-        monkeypatch.setattr(module.DocumentService, "get_by_id", lambda _id: (False, None))
-        res = _run(module.delete.__wrapped__("tenant-1", "ds-1"))
-        assert res["code"] == module.RetCode.DATA_ERROR
-        assert "Duplicate document ids" in res["message"]
-
    def test_parse_branches(self, monkeypatch):
        module = _load_doc_module(monkeypatch)
        monkeypatch.setattr(module.KnowledgebaseService, "accessible", lambda **_kwargs: False)
--- a/test/testcases/test_sdk_api/test_file_management_within_dataset/test_delete_documents.py
+++ b/test/testcases/test_sdk_api/test_file_management_within_dataset/test_delete_documents.py
@@ -24,10 +24,10 @@ class TestDocumentsDeletion:
    @pytest.mark.parametrize(
        "payload, expected_message, remaining",
        [
-            ({"ids": None}, "", 3),
-            ({"ids": []}, "", 3),
-            ({"ids": ["invalid_id"]}, "Documents not found: ['invalid_id']", 3),
-            ({"ids": ["\n!?。；！？\"'"]}, "Documents not found: ['\\n!?。；！？\"\\'']", 3),
+            ({"ids": None}, "should either provide doc ids or set delete_all(true), dataset:", 3),
+            ({"ids": []}, "should either provide doc ids or set delete_all(true), dataset:", 3),
+            ({"ids": ["invalid_id"]}, "Field: <ids> - Message: <Invalid UUID1 format> - Value: <['invalid_id']>", 3),
+            ({"ids": ["\n!?。；！？\"'"]}, "Field: <ids> - Message: <Invalid UUID1 format> - Value:", 3),
            ("not json", "must be a mapping", 3),
            (lambda r: {"ids": r[:1]}, "", 2),
            (lambda r: {"ids": r}, "", 0),
@@ -69,10 +69,10 @@ class TestDocumentsDeletion:

        with pytest.raises(Exception) as exception_info:
            dataset.delete_documents(**payload)
-        assert "Documents not found: ['invalid_id']" in str(exception_info.value), str(exception_info.value)
+        assert "Field: <ids> - Message: <Invalid UUID1 format> - Value: <" in str(exception_info.value), str(exception_info.value)

        documents = dataset.list_documents()
-        assert len(documents) == 0, str(documents)
+        assert len(documents) == 3, str(documents)

    @pytest.mark.p2
    def test_repeated_deletion(self, add_documents_func):
@@ -81,14 +81,16 @@ class TestDocumentsDeletion:
        dataset.delete_documents(ids=document_ids)
        with pytest.raises(Exception) as exception_info:
            dataset.delete_documents(ids=document_ids)
-        assert "Documents not found" in str(exception_info.value), str(exception_info.value)
+        assert "Document not found" in str(exception_info.value), str(exception_info.value)

    @pytest.mark.p2
    def test_duplicate_deletion(self, add_documents_func):
        dataset, documents = add_documents_func
        document_ids = [document.id for document in documents]
-        dataset.delete_documents(ids=document_ids + document_ids)
-        assert len(dataset.list_documents()) == 0, str(dataset.list_documents())
+        with pytest.raises(Exception) as exception_info:
+            dataset.delete_documents(ids=document_ids + document_ids)
+        assert "Field: <ids> - Message: <Duplicate ids:" in str(exception_info.value), str(exception_info.value)
+        assert len(dataset.list_documents()) == 3, str(dataset.list_documents())


@pytest.mark.p3
--- a/test/testcases/test_web_api/test_chunk_app/test_create_chunk.py
+++ b/test/testcases/test_web_api/test_chunk_app/test_create_chunk.py
@@ -218,8 +218,8 @@ class TestAddChunk:

    @pytest.mark.p2
    def test_add_chunk_to_deleted_document(self, WebApiAuth, add_document):
-        _, doc_id = add_document
-        delete_document(WebApiAuth, {"doc_id": doc_id})
+        kb_id, doc_id = add_document
+        delete_document(WebApiAuth, kb_id, {"ids": [doc_id]})
        res = add_chunk(WebApiAuth, {"doc_id": doc_id, "content_with_weight": "chunk test"})
        assert res["code"] == 102, res
        assert res["message"] == "Document not found!", res
--- a/test/testcases/test_web_api/test_chunk_app/test_update_chunk.py
+++ b/test/testcases/test_web_api/test_chunk_app/test_update_chunk.py
@@ -251,8 +251,8 @@ class TestUpdateChunk:

    @pytest.mark.p3
    def test_update_chunk_to_deleted_document(self, WebApiAuth, add_chunks):
-        _, doc_id, chunk_ids = add_chunks
-        delete_document(WebApiAuth, {"doc_id": doc_id})
+        kb_id, doc_id, chunk_ids = add_chunks
+        delete_document(WebApiAuth, kb_id, {"ids": [doc_id]})
        payload = {"doc_id": doc_id, "chunk_id": chunk_ids[0], "content_with_weight": "test content"}
        res = update_chunk(WebApiAuth, payload)
        assert res["code"] == 102, res
--- a/test/testcases/test_web_api/test_common.py
+++ b/test/testcases/test_web_api/test_common.py
@@ -382,8 +382,10 @@ def list_documents(auth, params=None, payload=None, *, headers=HEADERS, data=Non
    return res.json()


-def delete_document(auth, payload=None, *, headers=HEADERS, data=None):
-    res = requests.post(url=f"{HOST_ADDRESS}{DOCUMENT_APP_URL}/rm", headers=headers, auth=auth, json=payload, data=data)
+def delete_document(auth, dataset_id, payload=None, *, headers=HEADERS, data=None):
+    # New API: DELETE /api/v1/datasets/<dataset_id>/documents
+    url = f"{HOST_ADDRESS}{DATASETS_URL}/{dataset_id}/documents"
+    res = requests.delete(url=url, headers=headers, auth=auth, json=payload, data=data)
    return res.json()


--- a/test/testcases/test_web_api/test_document_app/conftest.py
+++ b/test/testcases/test_web_api/test_document_app/conftest.py
@@ -36,7 +36,7 @@ def add_document_func(request, WebApiAuth, add_dataset, ragflow_tmp_dir):
    def cleanup():
        res = list_documents(WebApiAuth, {"kb_id": dataset_id})
        for doc in res["data"]["docs"]:
-            delete_document(WebApiAuth, {"doc_id": doc["id"]})
+            delete_document(WebApiAuth, dataset_id, {"ids": [doc["id"]]})

    request.addfinalizer(cleanup)

@@ -49,7 +49,7 @@ def add_documents(request, WebApiAuth, add_dataset, ragflow_tmp_dir):
    def cleanup():
        res = list_documents(WebApiAuth, {"kb_id": dataset_id})
        for doc in res["data"]["docs"]:
-            delete_document(WebApiAuth, {"doc_id": doc["id"]})
+            delete_document(WebApiAuth, dataset_id, {"ids": [doc["id"]]})

    request.addfinalizer(cleanup)

@@ -62,7 +62,7 @@ def add_documents_func(request, WebApiAuth, add_dataset_func, ragflow_tmp_dir):
    def cleanup():
        res = list_documents(WebApiAuth, {"kb_id": dataset_id})
        for doc in res["data"]["docs"]:
-            delete_document(WebApiAuth, {"doc_id": doc["id"]})
+            delete_document(WebApiAuth, dataset_id, {"ids": [doc["id"]]})

    request.addfinalizer(cleanup)

--- a/test/testcases/test_web_api/test_document_app/test_rm_documents.py
+++ b/test/testcases/test_web_api/test_document_app/test_rm_documents.py
@@ -36,7 +36,7 @@ class TestAuthorization:
        ],
    )
    def test_invalid_auth(self, invalid_auth, expected_code, expected_message):
-        res = delete_document(invalid_auth)
+        res = delete_document(invalid_auth, "kb_id")
        assert res["code"] == expected_code, res
        assert res["message"] == expected_message, res

@@ -46,22 +46,23 @@ class TestDocumentsDeletion:
    @pytest.mark.parametrize(
        "payload, expected_code, expected_message, remaining",
        [
-            (None, 101, "required argument are missing: doc_id; ", 3),
-            ({"doc_id": ""}, 109, "No authorization.", 3),
-            ({"doc_id": "invalid_id"}, 109, "No authorization.", 3),
-            ({"doc_id": "\n!?。；！？\"'"}, 109, "No authorization.", 3),
-            ("not json", 101, "required argument are missing: doc_id; ", 3),
-            (lambda r: {"doc_id": r[0]}, 0, "", 2),
+            ({}, 102, "should either provide doc ids or set delete_all(true), dataset:", 3),
+            ({"invalid_key":[]}, 101, "Field: <invalid_key> - Message: <Extra inputs are not permitted> - Value: <[]>", 3),
+            ({"ids": ""}, 101, "Field: <ids> - Message: <Input should be a valid list> - Value: <>", 3),
+            ({"ids": ["invalid_id"]}, 101, "Field: <ids> - Message: <Invalid UUID1 format> - Value:", 3),
+            ("not json", 101, "Invalid request payload: expected object, got str", 3),
+            (lambda r: {"ids": r[0]}, 101, "Field: <ids> - Message: <Input should be a valid list> - Value", 3),
+            (lambda r: {"ids": r}, 0, "", 0),
        ],
    )
    def test_basic_scenarios(self, WebApiAuth, add_documents_func, payload, expected_code, expected_message, remaining):
        kb_id, document_ids = add_documents_func
        if callable(payload):
            payload = payload(document_ids)
-        res = delete_document(WebApiAuth, payload)
+        res = delete_document(WebApiAuth, kb_id, payload)
        assert res["code"] == expected_code, res
        if res["code"] != 0:
-            assert res["message"] == expected_message, res
+            assert expected_message in res["message"], res

        res = list_documents(WebApiAuth, {"kb_id": kb_id})
        assert len(res["data"]["docs"]) == remaining, res
@@ -69,57 +70,46 @@ class TestDocumentsDeletion:

    @pytest.mark.p2
    def test_repeated_deletion(self, WebApiAuth, add_documents_func):
-        _, document_ids = add_documents_func
+        kb_id, document_ids = add_documents_func
        for doc_id in document_ids:
-            res = delete_document(WebApiAuth, {"doc_id": doc_id})
+            res = delete_document(WebApiAuth, kb_id, {"ids": [doc_id]})
            assert res["code"] == 0, res

        for doc_id in document_ids:
-            res = delete_document(WebApiAuth, {"doc_id": doc_id})
-            assert res["code"] == 109, res
-            assert res["message"] == "No authorization.", res
+            res = delete_document(WebApiAuth, kb_id, {"ids": [doc_id]})
+            assert res["code"] == 102, res
+            assert res["message"] == "Document not found!", res
+
+    @pytest.mark.p2
+    def test_delete_all(self, WebApiAuth, add_documents_func):
+        kb_id, document_ids = add_documents_func
+
+        res = delete_document(WebApiAuth, kb_id, {"delete_all": True})
+        assert res["code"] == 0, res
+
+        res = list_documents(WebApiAuth, {"kb_id": kb_id})
+        assert len(res["data"]["docs"]) == 0, res
+        assert res["data"]["total"] == 0, res


@pytest.mark.p2
-class TestDocumentsDeletionUnit:
-    def test_rm_string_doc_id_normalization_success_unit(self, document_app_module, monkeypatch):
-        module = document_app_module
-        captured = {}
-
-        async def fake_request_json():
-            return {"doc_id": "doc1"}
-
-        async def fake_thread_pool_exec(func, doc_ids, user_id):
-            captured["func"] = func
-            captured["doc_ids"] = doc_ids
-            captured["user_id"] = user_id
-            return None
-
-        monkeypatch.setattr(module, "get_request_json", fake_request_json)
-        monkeypatch.setattr(module.DocumentService, "accessible4deletion", lambda *_args, **_kwargs: True)
-        monkeypatch.setattr(module, "thread_pool_exec", fake_thread_pool_exec)
-        res = _run(module.rm.__wrapped__())
-        assert res["code"] == 0
-        assert res["data"] is True
-        assert captured["func"] == module.FileService.delete_docs
-        assert captured["doc_ids"] == ["doc1"]
-        assert captured["user_id"] == module.current_user.id
-
-
-@pytest.mark.p3
 def test_concurrent_deletion(WebApiAuth, add_dataset, tmp_path):
    count = 100
    kb_id = add_dataset
    document_ids = bulk_upload_documents(WebApiAuth, kb_id, count, tmp_path)

    with ThreadPoolExecutor(max_workers=5) as executor:
-        futures = [executor.submit(delete_document, WebApiAuth, {"doc_id": document_ids[i]}) for i in range(count)]
+        futures = [executor.submit(delete_document, WebApiAuth, kb_id, {"ids": [document_ids[i]]}) for i in range(count)]
    responses = list(as_completed(futures))
    assert len(responses) == count, responses
    assert all(future.result()["code"] == 0 for future in futures), responses

+    res = list_documents(WebApiAuth, {"kb_id": kb_id})
+    assert len(res["data"]["docs"]) == 0, res
+    assert res["data"]["total"] == 0, res

-@pytest.mark.p3
+
+@pytest.mark.p2
 def test_delete_100(WebApiAuth, add_dataset, tmp_path):
    documents_num = 100
    kb_id = add_dataset
@@ -128,7 +118,7 @@ def test_delete_100(WebApiAuth, add_dataset, tmp_path):
    assert res["data"]["total"] == documents_num, res

    for doc_id in document_ids:
-        res = delete_document(WebApiAuth, {"doc_id": doc_id})
+        res = delete_document(WebApiAuth, kb_id, {"ids": [doc_id]})
        assert res["code"] == 0, res

    res = list_documents(WebApiAuth, {"kb_id": kb_id})
--- a/web/src/hooks/use-document-request.ts
+++ b/web/src/hooks/use-document-request.ts
@@ -16,6 +16,7 @@ import {
 import i18n from '@/locales/config';
 import { EMPTY_METADATA_FIELD } from '@/pages/dataset/dataset/use-select-filters';
 import kbService, {
+  deleteDocument,
  documentFilter,
  listDocument,
  renameDocument,
@@ -315,6 +316,7 @@ export const useRunDocument = () => {

 export const useRemoveDocument = () => {
  const queryClient = useQueryClient();
+  const { id: datasetId } = useParams();
  const {
    data,
    isPending: loading,
@@ -322,7 +324,8 @@ export const useRemoveDocument = () => {
  } = useMutation({
    mutationKey: [DocumentApiAction.RemoveDocument],
    mutationFn: async (documentIds: string | string[]) => {
-      const { data } = await kbService.documentRm({ doc_id: documentIds });
+      const ids = Array.isArray(documentIds) ? documentIds : [documentIds];
+      const { data } = await deleteDocument(datasetId!, ids);
      if (data.code === 0) {
        message.success(i18n.t('message.deleted'));
        queryClient.invalidateQueries({
@@ -435,7 +438,7 @@ export const useSetDocumentMeta = () => {
        }
        return data?.code;
      } catch (error) {
-        message.error('error');
+        message.error('error:' + error);
      }
    },
  });
--- a/web/src/services/knowledge-service.ts
+++ b/web/src/services/knowledge-service.ts
@@ -18,8 +18,6 @@ const {
  kbList,
  getDocumentList,
  documentChangeStatus,
-  documentRm,
-  documentDelete,
  documentCreate,
  documentChangeParser,
  documentThumbnails,
@@ -72,10 +70,6 @@ const methods = {
    url: documentChangeStatus,
    method: 'post',
  },
-  documentRm: {
-    url: documentRm,
-    method: 'post',
-  },
  documentCreate: {
    url: documentCreate,
    method: 'post',
@@ -137,10 +131,6 @@ const methods = {
    url: knowledgeGraph,
    method: 'get',
  },
-  documentDelete: {
-    url: documentDelete,
-    method: 'delete',
-  },
  listTagByKnowledgeIds: {
    url: listTagByKnowledgeIds,
    method: 'get',
@@ -276,6 +266,9 @@ export const renameDocument = (
  data: { name?: string },
 ) => request.patch(api.documentRename(datasetId, documentId), { data });

+export const deleteDocument = (datasetId: string, documentIds: string[]) =>
+  request.delete(api.documentDelete(datasetId), { data: { ids: documentIds } });
+
 export const getMetaDataService = ({
  kb_id,
  doc_ids,
--- a/web/src/utils/api.ts
+++ b/web/src/utils/api.ts
@@ -109,8 +109,8 @@ export default {
  getDocumentList: (datasetId: string) =>
    `${restAPIv1}/datasets/${datasetId}/documents`,
  documentChangeStatus: `${webAPI}/document/change_status`,
-  documentRm: `${webAPI}/document/rm`,
-  documentDelete: `${webAPI}/api/document`,
+  documentDelete: (datasetId: string) =>
+    `${restAPIv1}/datasets/${datasetId}/documents`,
  documentRename: (datasetId: string, documentId: string) =>
    `${restAPIv1}/datasets/${datasetId}/documents/${documentId}`,
  documentCreate: `${webAPI}/document/create`,