vanna-ai的参数配置改为.env的配置

api/configs/middleware/__init__.py

@@ -42,6 +42,7 @@ from .vdb.upstash_config import UpstashConfig
 from .vdb.vastbase_vector_config import VastbaseVectorConfig
 from .vdb.vikingdb_config import VikingDBConfig
 from .vdb.weaviate_config import WeaviateConfig
+from .vdb.vanna_config import VannaConfig
 
 
 class StorageConfig(BaseSettings):
@@ -323,5 +324,6 @@ class MiddlewareConfig(
     OpenGaussConfig,
     TableStoreConfig,
     DatasetQueueMonitorConfig,
+    VannaConfig,
 ):
     pass

api/configs/middleware/vdb/vanna_config.py

@@ -0,0 +1,80 @@
+from typing import Optional
+
+from pydantic import Field,PositiveInt
+from pydantic_settings import BaseSettings
+
+class VannaConfig(BaseSettings):
+    """
+    Configuration settings for Milvus vector database
+    """
+
+    VANNA_EMBEDDING_HOST: Optional[str] = Field(
+        description="vanna 向量模型地址",
+        default="http://127.0.0.1:19042",
+    )
+
+    VANNA_EMBEDDING_MODEL: Optional[str] = Field(
+        description="vanna 向量模型名称",
+        default='bge-m3',
+    )
+
+    VANNA_EMBEDDING_TYPE: Optional[str] = Field(
+        description="vanna 向量模型类型，默认是localhost,可以是ollama或其他类型",
+        default="localhost",
+    )
+
+    VANNA_LLM_TYPE: Optional[str] = Field(
+        description="vanna 语言模型类型，默认是deepseek,可以是ollama或其他类型",
+        default="deepseek",
+    )
+
+    VANNA_MODEL: str = Field(
+        description="vanna 语言模型版本，默认是deepseek-coder",
+        default="deepseek-coder",
+    )
+
+    VANNA_API_KEY: str = Field(
+        description="vanna 大模型API KEY",
+        default=None,
+    )
+
+    VANNA_SQL_TYPE: Optional[str] = Field(
+        description='vanna 训练数据库类型，默认是 postgres',
+        default="postgres",
+    )
+    VANNA_DB_USERNAME: Optional[str] = Field(
+        description='vanna 训练数据库用户名，默认是 postgres',
+        default='postgres',
+    )
+    VANNA_DB_PASSWORD: Optional[str] = Field(
+        description='vanna 训练数据库 postgres',
+        default='difyai123456',
+    )
+    VANNA_DB_HOST: Optional[str] = Field(
+        description='vanna 训练数据库地址，默认是 localhost',
+        default='localhost',
+    )
+    VANNA_DB_PORT: PositiveInt = Field(
+        description='vanna 训练数据库端口号，默认是 5432',
+        default=5432,
+    )
+    VANNA_DB_DATABASE: Optional[str] = Field(
+        description='vanna 训练数据库名称，默认是 vanna_demo',
+        default='vanna_demo',
+    )
+    VANNA_MILVUS_URI: Optional[str] = Field(
+        description='vanna 训练向量数据库地址，默认是 localhost:19530',
+        default='localhost:19530',
+    )
+    VANNA_MILVUS_USER: Optional[str] = Field(
+        description='vanna 训练向量数据库用户名，默认是 vanna_demo',
+        default='root',
+    )
+    VANNA_MILVUS_PASSWORD: Optional[str] = Field(
+        description='vanna 训练向量数据库密码，默认是 Milvus',
+        default='Milvus',
+    )
+    VANNA_MILVUS_DATABASE: str = Field(
+        description='vanna 训练向量数据库名称，默认是 vanna_demo',
+        default='vanna_demo',
+    )

api/extensions/ext_vanna_server.py

@@ -15,8 +15,11 @@ from datetime import datetime
 class Config:
     def __init__(self, supplier):
         self.embedding_supplier = "SiliconFlow"
-        self.milvus_uri = dify_config.MILVUS_URI
+        self.milvus_uri = dify_config.VANNA_MILVUS_URI
-        self.milvus_database = 'vanna_demo'
+        self.milvus_database = dify_config.VANNA_MILVUS_DATABASE
+        self.embedding_host = dify_config.VANNA_EMBEDDING_HOST
+        self.embedding_model = dify_config.VANNA_EMBEDDING_MODEL
+        self.embedding_type = dify_config.VANNA_EMBEDDING_TYPE
         self.supplier = supplier
         # self.llm_type = 'tongyi'
         # self.model = 'qwen-max'
@@ -24,16 +27,16 @@ class Config:
         # 本地模型
         # self.ollama_host = 'http://wsd.wisdomidata.com:19042'
         # self.model = 'qwen2:7b'
-        self.llm_type = 'deepseek'
+        self.llm_type = dify_config.VANNA_LLM_TYPE
-        self.model = 'deepseek-coder'
+        self.model = dify_config.VANNA_MODEL
-        self.api_key = 'sk-0382990b7a90496c889774b1d3843f90'
+        self.api_key = dify_config.VANNA_API_KEY
-        self.sql_type = 'postgres'
+        self.sql_type = dify_config.VANNA_SQL_TYPE
         self.sql_config = {
-            "host": dify_config.DB_HOST,
+            "host": dify_config.VANNA_DB_HOST,
-            "dbname": 'vanna_demo',
+            "dbname": dify_config.VANNA_DB_DATABASE,
-            "user": dify_config.DB_USERNAME,
+            "user": dify_config.VANNA_DB_USERNAME,
-            "password": dify_config.DB_PASSWORD,
+            "password": dify_config.VANNA_DB_PASSWORD,
-            "port": dify_config.DB_PORT
+            "port": dify_config.VANNA_DB_PORT
         }
 
 # 存储不同的 VannaServer 实例
@@ -214,7 +217,7 @@ def init_app(app: DifyApp):
 
     @app.route('/api/training/data/import', methods=['POST'])
     def training_data_import():
-       
+
         if 'file' not in request.files:
                 return jsonify({"type": "error", "error": "未上传文件"}), 400
 
@@ -235,4 +238,4 @@ def init_app(app: DifyApp):
             return jsonify({'status': 'success'}), 200
 
         except Exception as e:
-            return jsonify({"type": "error", "error": f"文件解析失败: {str(e)}"}), 500
+            return jsonify({"type": "error", "error": f"文件解析失败: {str(e)}"}), 500

api/extensions/utils/vanna_text2sql.py

@@ -69,16 +69,19 @@ class VannaServer:
         milvus_database = config["milvus_database"] if "milvus_database" in config else "test"
         milvus_client = MilvusClient(uri=milvus_uri,db_name=milvus_database)
 
+        embedding_type = config["embedding_type"]
         embedding_host = config["embedding_host"] if "embedding_host" in config else 'http://wsd.wisdomidata.com:19042'
         embedding_model = config["embedding_model"] if "embedding_model" in config else "bge-m3" # BAAI/bge-m3
-        # embedding_function = model.dense.SentenceTransformerEmbeddingFunction(
+        if embedding_type == "ollama":
-        #     model_name=embedding_model,
+            embedding_function = CustomEmbeddingFunction({
-        #     device='cpu'  # 'cpu' or 'cuda:0'
+                "host": embedding_host,
-        # )
+                "embed_model": embedding_model
-        embedding_function = CustomEmbeddingFunction({
+            })
-            "host": embedding_host,
+        else:
-            "embed_model": embedding_model
+            embedding_function = model.dense.SentenceTransformerEmbeddingFunction(
-        })
+                model_name=embedding_model,
+                device='cpu'  # 'cpu' or 'cuda:0'
+            )
         chat_llm = Ollama
         if llm_type == "ollama":
             config = {