chore: apply ty checks on api code with script and ci action (#24653)

api/core/rag/datasource/vdb/analyticdb/analyticdb_vector_openapi.py

@@ -192,8 +192,8 @@ class AnalyticdbVectorOpenAPI:
             collection=self._collection_name,
             metrics=self.config.metrics,
             include_values=True,
-            vector=None,
-            content=None,
+            vector=None,  # ty: ignore [invalid-argument-type]
+            content=None,  # ty: ignore [invalid-argument-type]
             top_k=1,
             filter=f"ref_doc_id='{id}'",
         )
@@ -211,7 +211,7 @@ class AnalyticdbVectorOpenAPI:
             namespace=self.config.namespace,
             namespace_password=self.config.namespace_password,
             collection=self._collection_name,
-            collection_data=None,
+            collection_data=None,  # ty: ignore [invalid-argument-type]
             collection_data_filter=f"ref_doc_id IN {ids_str}",
         )
         self._client.delete_collection_data(request)
@@ -225,7 +225,7 @@ class AnalyticdbVectorOpenAPI:
             namespace=self.config.namespace,
             namespace_password=self.config.namespace_password,
             collection=self._collection_name,
-            collection_data=None,
+            collection_data=None,  # ty: ignore [invalid-argument-type]
             collection_data_filter=f"metadata_ ->> '{key}' = '{value}'",
         )
         self._client.delete_collection_data(request)
@@ -249,7 +249,7 @@ class AnalyticdbVectorOpenAPI:
             include_values=kwargs.pop("include_values", True),
             metrics=self.config.metrics,
             vector=query_vector,
-            content=None,
+            content=None,  # ty: ignore [invalid-argument-type]
             top_k=kwargs.get("top_k", 4),
             filter=where_clause,
         )
@@ -285,7 +285,7 @@ class AnalyticdbVectorOpenAPI:
             collection=self._collection_name,
             include_values=kwargs.pop("include_values", True),
             metrics=self.config.metrics,
-            vector=None,
+            vector=None,  # ty: ignore [invalid-argument-type]
             content=query,
             top_k=kwargs.get("top_k", 4),
             filter=where_clause,

api/core/rag/datasource/vdb/clickzetta/clickzetta_vector.py

@@ -12,7 +12,7 @@ import clickzetta  # type: ignore
 from pydantic import BaseModel, model_validator
 
 if TYPE_CHECKING:
-    from clickzetta import Connection
+    from clickzetta.connector.v0.connection import Connection  # type: ignore
 
 from configs import dify_config
 from core.rag.datasource.vdb.field import Field

api/core/rag/datasource/vdb/couchbase/couchbase_vector.py

@@ -306,7 +306,7 @@ class CouchbaseVector(BaseVector):
     def search_by_full_text(self, query: str, **kwargs: Any) -> list[Document]:
         top_k = kwargs.get("top_k", 4)
         try:
-            CBrequest = search.SearchRequest.create(search.QueryStringQuery("text:" + query))
+            CBrequest = search.SearchRequest.create(search.QueryStringQuery("text:" + query))  # ty: ignore [too-many-positional-arguments]
             search_iter = self._scope.search(
                 self._collection_name + "_search", CBrequest, SearchOptions(limit=top_k, fields=["*"])
             )

api/core/rag/datasource/vdb/elasticsearch/elasticsearch_vector.py

@@ -138,7 +138,7 @@ class ElasticSearchVector(BaseVector):
             if not client.ping():
                 raise ConnectionError("Failed to connect to Elasticsearch")
 
-        except requests.exceptions.ConnectionError as e:
+        except requests.ConnectionError as e:
             raise ConnectionError(f"Vector database connection error: {str(e)}")
         except Exception as e:
             raise ConnectionError(f"Elasticsearch client initialization failed: {str(e)}")

api/core/rag/datasource/vdb/milvus/milvus_vector.py

@@ -376,7 +376,12 @@ class MilvusVector(BaseVector):
         if config.token:
             client = MilvusClient(uri=config.uri, token=config.token, db_name=config.database)
         else:
-            client = MilvusClient(uri=config.uri, user=config.user, password=config.password, db_name=config.database)
+            client = MilvusClient(
+                uri=config.uri,
+                user=config.user or "",
+                password=config.password or "",
+                db_name=config.database,
+            )
         return client
 
 

api/core/rag/datasource/vdb/vikingdb/vikingdb_vector.py

@@ -32,9 +32,9 @@ class VikingDBConfig(BaseModel):
     scheme: str
     connection_timeout: int
     socket_timeout: int
-    index_type: str = IndexType.HNSW
-    distance: str = DistanceType.L2
-    quant: str = QuantType.Float
+    index_type: str = str(IndexType.HNSW)
+    distance: str = str(DistanceType.L2)
+    quant: str = str(QuantType.Float)
 
 
 class VikingDBVector(BaseVector):

api/core/rag/datasource/vdb/weaviate/weaviate_vector.py

@@ -37,22 +37,22 @@ class WeaviateVector(BaseVector):
         self._attributes = attributes
 
     def _init_client(self, config: WeaviateConfig) -> weaviate.Client:
-        auth_config = weaviate.auth.AuthApiKey(api_key=config.api_key)
+        auth_config = weaviate.AuthApiKey(api_key=config.api_key or "")
 
-        weaviate.connect.connection.has_grpc = False
+        weaviate.connect.connection.has_grpc = False  # ty: ignore [unresolved-attribute]
 
         # Fix to minimize the performance impact of the deprecation check in weaviate-client 3.24.0,
         # by changing the connection timeout to pypi.org from 1 second to 0.001 seconds.
         # TODO: This can be removed once weaviate-client is updated to 3.26.7 or higher,
         #       which does not contain the deprecation check.
-        if hasattr(weaviate.connect.connection, "PYPI_TIMEOUT"):
-            weaviate.connect.connection.PYPI_TIMEOUT = 0.001
+        if hasattr(weaviate.connect.connection, "PYPI_TIMEOUT"):  # ty: ignore [unresolved-attribute]
+            weaviate.connect.connection.PYPI_TIMEOUT = 0.001  # ty: ignore [unresolved-attribute]
 
         try:
             client = weaviate.Client(
                 url=config.endpoint, auth_client_secret=auth_config, timeout_config=(5, 60), startup_period=None
             )
-        except requests.exceptions.ConnectionError:
+        except requests.ConnectionError:
             raise ConnectionError("Vector database connection error")
 
         client.batch.configure(

api/core/rag/retrieval/router/multi_dataset_function_call_router.py

@@ -1,4 +1,4 @@
-from typing import Union, cast
+from typing import Union
 
 from core.app.entities.app_invoke_entities import ModelConfigWithCredentialsEntity
 from core.model_manager import ModelInstance
@@ -28,14 +28,11 @@ class FunctionCallMultiDatasetRouter:
                 SystemPromptMessage(content="You are a helpful AI assistant."),
                 UserPromptMessage(content=query),
             ]
-            result = cast(
-                LLMResult,
-                model_instance.invoke_llm(
-                    prompt_messages=prompt_messages,
-                    tools=dataset_tools,
-                    stream=False,
-                    model_parameters={"temperature": 0.2, "top_p": 0.3, "max_tokens": 1500},
-                ),
+            result: LLMResult = model_instance.invoke_llm(
+                prompt_messages=prompt_messages,
+                tools=dataset_tools,
+                stream=False,
+                model_parameters={"temperature": 0.2, "top_p": 0.3, "max_tokens": 1500},
             )
             if result.message.tool_calls:
                 # get retrieval model config

api/core/rag/retrieval/router/multi_dataset_react_route.py

@@ -1,5 +1,5 @@
 from collections.abc import Generator, Sequence
-from typing import Union, cast
+from typing import Union
 
 from core.app.entities.app_invoke_entities import ModelConfigWithCredentialsEntity
 from core.model_manager import ModelInstance
@@ -150,15 +150,12 @@ class ReactMultiDatasetRouter:
         :param stop: stop
         :return:
         """
-        invoke_result = cast(
-            Generator[LLMResult, None, None],
-            model_instance.invoke_llm(
-                prompt_messages=prompt_messages,
-                model_parameters=completion_param,
-                stop=stop,
-                stream=True,
-                user=user_id,
-            ),
+        invoke_result: Generator[LLMResult, None, None] = model_instance.invoke_llm(
+            prompt_messages=prompt_messages,
+            model_parameters=completion_param,
+            stop=stop,
+            stream=True,
+            user=user_id,
         )
 
         # handle invoke result