fix ci

pingcap · Feb 17, 2025 · 2b39d38 · 2b39d38
1 parent 7bcb7f8
commit 2b39d38
Show file tree

Hide file tree

Showing 4 changed files with 13 additions and 10 deletions.
diff --git a/backend/app/api/admin_routes/knowledge_base/models.py b/backend/app/api/admin_routes/knowledge_base/models.py
@@ -14,7 +14,7 @@
 )
 from app.exceptions import KBNoVectorIndexConfigured
 from app.models import KgIndexStatus
-from app.models.knowledge_base import IndexMethod, ChunkingConfig, GeneralChunkingConfig
+from app.models.knowledge_base import IndexMethod, GeneralChunkingConfig, ChunkingConfig
 
 
 class KnowledgeBaseCreate(BaseModel):
@@ -56,12 +56,13 @@ class KnowledgeBaseDetail(BaseModel):
 
     id: int
     name: str
-    description: str
+    description: Optional[str] = None
     documents_total: int
     data_sources_total: int
     # Notice: By default, SQLModel will not serialize list type relationships.
     # https://github.com/fastapi/sqlmodel/issues/37#issuecomment-2093607242
     data_sources: list[KBDataSource]
+    chunking_config: Optional[ChunkingConfig] = None
     index_methods: list[IndexMethod]
     llm_id: int | None = None
     llm: LLMDescriptor | None = None
@@ -79,7 +80,7 @@ class KnowledgeBaseItem(BaseModel):
 
     id: int
     name: str
-    description: str
+    description: Optional[str] = None
     documents_total: int
     data_sources_total: int
     index_methods: list[IndexMethod]

diff --git a/backend/app/api/admin_routes/knowledge_base/routes.py b/backend/app/api/admin_routes/knowledge_base/routes.py
@@ -75,7 +75,7 @@ def create_knowledge_base(
             index_methods=create.index_methods,
             llm_id=create.llm_id,
             embedding_model_id=create.embedding_model_id,
-            chunking_config=create.chunking_config,
+            chunking_config=create.chunking_config.model_dump(),
             data_sources=data_sources,
             created_by=user.id,
             updated_by=user.id,

diff --git a/backend/app/models/knowledge_base.py b/backend/app/models/knowledge_base.py
@@ -96,29 +96,31 @@ class ChunkingMode(str, enum.Enum):
     ADVANCED = "advanced"
 
 
-class ChunkingConfig(BaseModel):
+class BaseChunkingConfig(BaseModel):
     mode: ChunkingMode = Field(default=ChunkingMode.GENERAL)
 
 
-class GeneralChunkingConfig(ChunkingConfig):
+class GeneralChunkingConfig(BaseChunkingConfig):
     mode: ChunkingMode = Field(default=ChunkingMode.GENERAL)
     chunk_size: int = Field(default=DEFAULT_CHUNK_SIZE, gt=0)
     chunk_overlap: int = Field(default=SENTENCE_CHUNK_OVERLAP, gt=0)
     paragraph_separator: str = Field(default=DEFAULT_PARAGRAPH_SEP)
 
 
-class AdvancedChunkingConfig(ChunkingConfig):
+class AdvancedChunkingConfig(BaseChunkingConfig):
     mode: ChunkingMode = Field(default=ChunkingMode.ADVANCED)
     rules: Dict[MimeTypes, ChunkSplitterConfig] = Field(default_factory=list)
 
 
+ChunkingConfig = Union[GeneralChunkingConfig | AdvancedChunkingConfig]
+
 # Knowledge Base Model
 
 
 class KnowledgeBase(SQLModel, table=True):
     id: Optional[int] = Field(default=None, primary_key=True)
     name: str = Field(max_length=255, nullable=False)
-    description: str = Field(sa_column=Column(MEDIUMTEXT))
+    description: Optional[str] = Field(sa_column=Column(MEDIUMTEXT), default=None)
 
     # The config for chunking, the process to break down the document into smaller chunks.
     chunking_config: Dict = Field(

diff --git a/frontend/app/src/api/knowledge-base.ts b/frontend/app/src/api/knowledge-base.ts
@@ -63,7 +63,7 @@ export interface UpdateKnowledgeBaseParams {
 export interface KnowledgeBaseSummary {
   id: number;
   name: string;
-  description: string;
+  description: string | null;
   index_methods: KnowledgeBaseIndexMethod[];
   documents_total?: number;
   data_sources_total?: number;
@@ -94,7 +94,7 @@ export type KnowledgeGraphDocumentChunk = z.infer<typeof knowledgeGraphDocumentC
 const knowledgeBaseSummarySchema = z.object({
   id: z.number(),
   name: z.string(),
-  description: z.string(),
+  description: z.string().nullable(),
   index_methods: z.enum(['vector', 'knowledge_graph']).array(),
   documents_total: z.number().optional(),
   data_sources_total: z.number().optional(),