PydanticUserError: SQLDatabaseToolkit is not fully defined; you should define Callbacks, then call SQLDatabaseToolkit.model_rebuild().

[arno756]

Checked other resources

• I added a very descriptive title to this issue.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.
• I am sure that this is a bug in LangChain rather than my code.
• The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).

Example Code

pip install "autogen-ext[azure]" python-dotenv langchain-community psycopg2 langchain_openai langchain_experimental "autogen-core" "autogen-agentchat"

from dotenv import load_dotenv
import os
from langchain_community.utilities.sql_database import SQLDatabase
import psycopg2
from langchain_openai import AzureChatOpenAI
from langchain_experimental.sql import SQLDatabaseChain
from langchain_experimental.sql.base import SQLDatabaseChain
from langchain.cache import InMemoryCache

# Load environment variables from the .env file from the same directory as notebook 
load_dotenv()

# Retrieve environment variables
POSTGRES_USER = os.getenv('POSTGRES_USER')
POSTGRES_PASSWORD = os.getenv('POSTGRES_PASSWORD')
POSTGRES_HOST = os.getenv('POSTGRES_HOST')
POSTGRES_PORT = os.getenv('POSTGRES_PORT')
POSTGRES_DB = os.getenv('POSTGRES_DB')
AZURE_OPENAI_KEY = os.getenv('AZURE_OPENAI_KEY')
AZURE_OPENAI_ENDPOINT = os.getenv('AZURE_OPENAI_ENDPOINT')
AZURE_OPENAI_DEPLOYMENT = os.getenv('AZURE_OPENAI_DEPLOYMENT')

llm_config = {
    "config_list": [
        {
        "model": AZURE_OPENAI_DEPLOYMENT,
        "temperature": 0.7,
        "api_key": AZURE_OPENAI_KEY,
        "azure_endpoint": AZURE_OPENAI_ENDPOINT,
        "api_type": "azure",
        "api_version": "2024-10-21"
        }]}

shipment_db_uri = f"postgresql+psycopg2://{POSTGRES_USER}:{POSTGRES_PASSWORD}@{POSTGRES_HOST}:{POSTGRES_PORT}/{POSTGRES_DB}"

shipment_chain = SQLDatabaseChain(llm=azure_llm, database=shipment_db, verbose=True)

Error Message and Stack Trace (if applicable)

---

PydanticUserError Traceback (most recent call last)
Cell In[22], line 116
2 llm_config = {
3 "config_list": [
4 {
(...)
111 ]
112 }
115 # Initialize the database chains
--> 116 shipment_chain = SQLDatabaseChain(llm=azure_llm, database=shipment_db, verbose=True)
117 crm_chain = SQLDatabaseChain(llm=azure_llm, database=crm_db, verbose=True)
119 # Create assistant agents

File ~/Projects/multi-AI-test/azure-postgresql-openai-langchain-autogen-demo/new_env_0.4/lib/python3.12/site-packages/langchain_core/load/serializable.py:125, in Serializable.init(self, *args, **kwargs)
123 def init(self, *args: Any, **kwargs: Any) -> None:
124 """"""
--> 125 super().init(*args, **kwargs)

[... skipping hidden 1 frame]

File ~/Projects/multi-AI-test/azure-postgresql-openai-langchain-autogen-demo/new_env_0.4/lib/python3.12/site-packages/pydantic/_internal/_mock_val_ser.py:100, in MockValSer.getattr(self, item)
98 # raise an AttributeError if item doesn't exist
99 getattr(self._val_or_ser, item)
--> 100 raise PydanticUserError(self._error_message, code=self._code)

PydanticUserError: SQLDatabaseChain is not fully defined; you should define BaseCache, then call SQLDatabaseChain.model_rebuild().

Description

I saw this issue: https://github.com/langchain-ai/langchain/pull/28297 and that one (#28297) looked like it solved but it hasn't for me. I have to go back to 2.9.2 but now it's not compatible with autogen 0.4 and I have to get to the latest version.

System Info

System Information

OS: Darwin
OS Version: Darwin Kernel Version 24.2.0: Fri Dec 6 19:02:41 PST 2024; root:xnu-11215.61.5~2/RELEASE_ARM64_T6030
Python Version: 3.12.8 (main, Dec 3 2024, 18:42:41) [Clang 16.0.0 (clang-1600.0.26.4)]

Package Information

langchain_core: 0.3.29
langchain: 0.3.14
langchain_community: 0.3.14
langsmith: 0.2.10
langchain_experimental: 0.3.4
langchain_openai: 0.3.0
langchain_text_splitters: 0.3.5

Optional packages not installed

langserve

Other Dependencies

aiohttp: 3.11.11
async-timeout: Installed. No version info available.
dataclasses-json: 0.6.7
httpx: 0.28.1
httpx-sse: 0.4.0
jsonpatch: 1.33
langsmith-pyo3: Installed. No version info available.
numpy: 2.2.1
openai: 1.59.7
orjson: 3.10.14
packaging: 24.2
pydantic: 2.10.0
pydantic-settings: 2.7.1
PyYAML: 6.0.2
requests: 2.32.3
requests-toolbelt: 1.0.0
SQLAlchemy: 2.0.37
tenacity: 9.0.0
tiktoken: 0.8.0
typing-extensions: 4.12.2
zstandard: Installed. No version info available.

[keenborder786]

You are using very old and experimental way of interacting with the database. You can do something as follow to construct a SQL Agent which talks with your data:

from langchain_community.agent_toolkits.sql.base import create_sql_agent
from langchain_openai import ChatOpenAI
from langchain_community.utilities.sql_database import SQLDatabase
from langchain_community.agent_toolkits.sql.toolkit import SQLDatabaseToolkit

agent = create_sql_agent(llm = ChatOpenAI(model = "gpt-4o-mini"),
                        agent_type = "openai-tools", 
                        verbose=True,
                        toolkit = SQLDatabaseToolkit(db = SQLDatabase.from_uri(database_uri = "postgresql+psycopg2://root:password@localhost:5432/dev_db"),llm = ChatOpenAI(model = "gpt-4o-mini")))

agent.invoke({"input": "How many calls happenend?"})