Fix metadata config (#14480)

### What problem does this PR solve?

Fix metadata config

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)

api/apps/services/dataset_api_service.py

@@ -650,25 +650,8 @@ def get_auto_metadata(dataset_id: str, tenant_id: str):
     kb = KnowledgebaseService.get_or_none(id=dataset_id, tenant_id=tenant_id)
     if kb is None:
         return False, f"User '{tenant_id}' lacks permission for dataset '{dataset_id}'"
-
     parser_cfg = kb.parser_config or {}
-    metadata = parser_cfg.get("metadata") or []
-    enabled = parser_cfg.get("enable_metadata", bool(metadata))
-    # Normalize to AutoMetadataConfig-like JSON
-    fields = []
-    for f in metadata:
-        if not isinstance(f, dict):
-            continue
-        fields.append(
-            {
-                "name": f.get("name", ""),
-                "type": f.get("type", ""),
-                "description": f.get("description"),
-                "examples": f.get("examples"),
-                "restrict_values": f.get("restrict_values", False),
-            }
-        )
-    return True, {"enabled": enabled, "fields": fields}
+    return True, {"metadata": parser_cfg.get("metadata") or [], "built_in_metadata": parser_cfg.get("built_in_metadata") or []}
 
 
 async def update_auto_metadata(dataset_id: str, tenant_id: str, cfg: dict):
@@ -685,24 +668,13 @@ async def update_auto_metadata(dataset_id: str, tenant_id: str, cfg: dict):
         return False, f"User '{tenant_id}' lacks permission for dataset '{dataset_id}'"
 
     parser_cfg = kb.parser_config or {}
-    fields = []
-    for f in cfg.get("fields", []):
-        fields.append(
-            {
-                "name": f.get("name", ""),
-                "type": f.get("type", ""),
-                "description": f.get("description"),
-                "examples": f.get("examples"),
-                "restrict_values": f.get("restrict_values", False),
-            }
-        )
-    parser_cfg["metadata"] = fields
-    parser_cfg["enable_metadata"] = cfg.get("enabled", True)
+    parser_cfg["metadata"] = cfg.get("metadata")
+    parser_cfg["built_in_metadata"] = cfg.get("built_in_metadata")
 
     if not KnowledgebaseService.update_by_id(kb.id, {"parser_config": parser_cfg}):
         return False, "Update auto-metadata error.(Database error)"
 
-    return True, {"enabled": parser_cfg["enable_metadata"], "fields": fields}
+    return True, cfg
 
 
 def delete_tags(dataset_id: str, tenant_id: str, tags: list[str]):

api/utils/validation_utils.py

@@ -364,18 +364,17 @@ class ParentChildConfig(Base):
 class AutoMetadataField(Base):
     """Schema for a single auto-metadata field configuration."""
 
-    name: Annotated[str, StringConstraints(strip_whitespace=True, min_length=1, max_length=255), Field(...)]
-    type: Annotated[Literal["string", "list", "time"], Field(...)]
+    key: Annotated[str, StringConstraints(strip_whitespace=True, min_length=1, max_length=255), Field(...)]
+    type: Annotated[Literal["string", "list", "time", "number"], Field(...)]
     description: Annotated[str | None, Field(default=None, max_length=65535)]
-    examples: Annotated[list[str] | None, Field(default=None)]
-    restrict_values: Annotated[bool, Field(default=False)]
+    enum: Annotated[list[str] | None, Field(default=None)]
 
 
 class AutoMetadataConfig(Base):
     """Top-level auto-metadata configuration attached to a dataset."""
 
-    enabled: Annotated[bool, Field(default=True)]
-    fields: Annotated[list[AutoMetadataField], Field(default_factory=list)]
+    metadata: Annotated[list[AutoMetadataField], Field(default_factory=list)]
+    built_in_metadata: Annotated[list[AutoMetadataField], Field(default_factory=list)]
 
 
 class ParserConfig(Base):

test/testcases/test_http_api/test_file_management_within_dataset/test_metadata_retrieval.py

@@ -70,11 +70,10 @@ def add_dataset_with_metadata(HttpApiAuth):
         headers={"Content-Type": "application/json"},
         auth=HttpApiAuth,
         json={
-            "enabled": False,
-            "fields": [
-                {"name": "character", "type": "string", "description": "Historical figure name"},
-                {"name": "era", "type": "string", "description": "Historical era"},
-                {"name": "achievements", "type": "list", "description": "Major achievements"},
+            "metadata": [
+                {"key": "character", "type": "string", "description": "Historical figure name"},
+                {"key": "era", "type": "string", "description": "Historical era"},
+                {"key": "achievements", "type": "list", "description": "Major achievements"},
             ]
         }
     ).json()

test/testcases/test_sdk_api/test_dataset_mangement/test_auto_metadata.py

@@ -1,126 +0,0 @@
-#
-#  Copyright 2025 The InfiniFlow Authors. All Rights Reserved.
-#
-#  Licensed under the Apache License, Version 2.0 (the "License");
-#  you may not use this file except in compliance with the License.
-#  You may obtain a copy of the License at
-#  http://www.apache.org/licenses/LICENSE-2.0
-#  Unless required by applicable law or agreed to in writing, software
-#  distributed under the License is distributed on an "AS IS" BASIS,
-#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-#  See the License for the specific language governing permissions and
-#  limitations under the License.
-#
-
-import pytest
-
-
-@pytest.mark.usefixtures("clear_datasets")
-class TestAutoMetadataOnCreate:
-    @pytest.mark.p1
-    def test_create_dataset_with_auto_metadata(self, client):
-        payload = {
-            "name": "auto_metadata_create",
-            "auto_metadata_config": {
-                "enabled": True,
-                "fields": [
-                    {
-                        "name": "author",
-                        "type": "string",
-                        "description": "The author of the document",
-                        "examples": ["John Doe", "Jane Smith"],
-                        "restrict_values": False,
-                    },
-                    {
-                        "name": "category",
-                        "type": "list",
-                        "description": "Document category",
-                        "examples": ["Technical", "Business"],
-                        "restrict_values": True,
-                    },
-                ],
-            },
-        }
-        dataset = client.create_dataset(**payload)
-        # The SDK should expose parser_config via internal properties or metadata;
-        # we rely on the HTTP API for verification via get_auto_metadata.
-        cfg = dataset.get_auto_metadata()
-        assert cfg["enabled"] is True
-        assert len(cfg["fields"]) == 2
-        names = {f["name"] for f in cfg["fields"]}
-        assert names == {"author", "category"}
-
-
-@pytest.mark.usefixtures("clear_datasets")
-class TestAutoMetadataOnUpdate:
-    @pytest.mark.p1
-    def test_update_auto_metadata_via_dataset_update(self, client, add_dataset_func):
-        dataset = add_dataset_func
-
-        # Initially set auto-metadata via dataset.update
-        payload = {
-            "auto_metadata_config": {
-                "enabled": True,
-                "fields": [
-                    {
-                        "name": "tags",
-                        "type": "list",
-                        "description": "Document tags",
-                        "examples": ["AI", "ML", "RAG"],
-                        "restrict_values": False,
-                    }
-                ],
-            }
-        }
-        dataset.update(payload)
-
-        cfg = dataset.get_auto_metadata()
-        assert cfg["enabled"] is True
-        assert len(cfg["fields"]) == 1
-        assert cfg["fields"][0]["name"] == "tags"
-        assert cfg["fields"][0]["type"] == "list"
-
-        # Disable auto-metadata and replace fields
-        update_cfg = {
-            "enabled": False,
-            "fields": [
-                {
-                    "name": "year",
-                    "type": "time",
-                    "description": "Publication year",
-                    "examples": None,
-                    "restrict_values": False,
-                }
-            ],
-        }
-        dataset.update_auto_metadata(**update_cfg)
-
-        cfg2 = dataset.get_auto_metadata()
-        assert cfg2["enabled"] is False
-        assert len(cfg2["fields"]) == 1
-        assert cfg2["fields"][0]["name"] == "year"
-        assert cfg2["fields"][0]["type"] == "time"
-
-
-@pytest.mark.usefixtures("clear_datasets")
-class TestAutoMetadataValidation:
-    @pytest.mark.p2
-    def test_invalid_field_type_rejected(self, client):
-        payload = {
-            "name": "auto_metadata_invalid_type",
-            "auto_metadata_config": {
-                "enabled": True,
-                "fields": [
-                    {
-                        "name": "invalid_type",
-                        "type": "unknown",  # invalid literal
-                    }
-                ],
-            },
-        }
-        with pytest.raises(Exception) as exc_info:
-            client.create_dataset(**payload)
-        msg = str(exc_info.value)
-        # Pydantic literal_error message should appear
-        assert "Input should be" in msg or "literal_error" in msg
-

web/src/pages/dataset/components/metedata/hooks/use-manage-modal.ts

@@ -98,6 +98,7 @@ export const util = {
     return data.map((item) => {
       return {
         key: item.field,
+        type: item.valueType?.toLowerCase(),
         description: item.description,
         enum: item.values,
       };