langchain/libs/community/langchain_community/tools/sql_database/tool.py

# flake8: noqa
"""Tools for interacting with a SQL database."""
from typing import Any, Dict, Optional

from langchain_core.pydantic_v1 import BaseModel, Extra, Field, root_validator

from langchain_core.language_models import BaseLanguageModel
from langchain_core.callbacks import (
    AsyncCallbackManagerForToolRun,
    CallbackManagerForToolRun,
)
from langchain_core.prompts import PromptTemplate
from langchain_community.utilities.sql_database import SQLDatabase
from langchain_core.tools import BaseTool
from langchain_community.tools.sql_database.prompt import QUERY_CHECKER


class BaseSQLDatabaseTool(BaseModel):
    """Base tool for interacting with a SQL database."""

    db: SQLDatabase = Field(exclude=True)

    class Config(BaseTool.Config):
        pass


class QuerySQLDataBaseTool(BaseSQLDatabaseTool, BaseTool):
    """Tool for querying a SQL database."""

    name: str = "sql_db_query"
    description: str = """
    Input to this tool is a detailed and correct SQL query, output is a result from the database.
    If the query is not correct, an error message will be returned.
    If an error is returned, rewrite the query, check the query, and try again.
    """

    def _run(
        self,
        query: str,
        run_manager: Optional[CallbackManagerForToolRun] = None,
    ) -> str:
        """Execute the query, return the results or an error message."""
        return self.db.run_no_throw(query)


class InfoSQLDatabaseTool(BaseSQLDatabaseTool, BaseTool):
    """Tool for getting metadata about a SQL database."""

    name: str = "sql_db_schema"
    description: str = """
    Input to this tool is a comma-separated list of tables, output is the schema and sample rows for those tables.    

    Example Input: "table1, table2, table3"
    """

    def _run(
        self,
        table_names: str,
        run_manager: Optional[CallbackManagerForToolRun] = None,
    ) -> str:
        """Get the schema for tables in a comma-separated list."""
        return self.db.get_table_info_no_throw(
            [t.strip() for t in table_names.split(",")]
        )


class ListSQLDatabaseTool(BaseSQLDatabaseTool, BaseTool):
    """Tool for getting tables names."""

    name: str = "sql_db_list_tables"
    description: str = "Input is an empty string, output is a comma separated list of tables in the database."

    def _run(
        self,
        tool_input: str = "",
        run_manager: Optional[CallbackManagerForToolRun] = None,
    ) -> str:
        """Get the schema for a specific table."""
        return ", ".join(self.db.get_usable_table_names())


class QuerySQLCheckerTool(BaseSQLDatabaseTool, BaseTool):
    """Use an LLM to check if a query is correct.
    Adapted from https://www.patterns.app/blog/2023/01/18/crunchbot-sql-analyst-gpt/"""

    template: str = QUERY_CHECKER
    llm: BaseLanguageModel
    llm_chain: Any = Field(init=False)
    name: str = "sql_db_query_checker"
    description: str = """
    Use this tool to double check if your query is correct before executing it.
    Always use this tool before executing a query with sql_db_query!
    """

    @root_validator(pre=True)
    def initialize_llm_chain(cls, values: Dict[str, Any]) -> Dict[str, Any]:
        if "llm_chain" not in values:
            from langchain.chains.llm import LLMChain

            values["llm_chain"] = LLMChain(
                llm=values.get("llm"),
                prompt=PromptTemplate(
                    template=QUERY_CHECKER, input_variables=["dialect", "query"]
                ),
            )

        if values["llm_chain"].prompt.input_variables != ["dialect", "query"]:
            raise ValueError(
                "LLM chain for QueryCheckerTool must have input variables ['query', 'dialect']"
            )

        return values

    def _run(
        self,
        query: str,
        run_manager: Optional[CallbackManagerForToolRun] = None,
    ) -> str:
        """Use the LLM to check the query."""
        return self.llm_chain.predict(
            query=query,
            dialect=self.db.dialect,
            callbacks=run_manager.get_child() if run_manager else None,
        )

    async def _arun(
        self,
        query: str,
        run_manager: Optional[AsyncCallbackManagerForToolRun] = None,
    ) -> str:
        return await self.llm_chain.apredict(
            query=query,
            dialect=self.db.dialect,
            callbacks=run_manager.get_child() if run_manager else None,
        )