Defining custom tools
One option for creating a tool that runs custom code is to use a DynamicTool
.
The DynamicTool
and DynamicStructuredTool
classes takes as input a name, a description, and a function.
Importantly, the name and the description will be used by the language model to determine when to call this function and with what parameters,
so make sure to set these to some values the language model can reason about!
The provided function is what will the agent will actually call. When an error occurs, the function should, when possible, return a string representing an error, rather than throwing an error. This allows the error to be passed to the LLM and the LLM can decide how to handle it. If an error is thrown, then execution of the agent will stop.
DynamicStructuredTool
s allow you to specify more complex inputs as Zod schemas for the agent to populate. However, note that more complex schemas require
better models and agents. See this guide for a complete list of agent types.
See below for an example of defining and using DynamicTool
s.
- npm
- Yarn
- pnpm
npm install @langchain/openai
yarn add @langchain/openai
pnpm add @langchain/openai
import { ChatOpenAI } from "@langchain/openai";
import type { ChatPromptTemplate } from "@langchain/core/prompts";
import { createOpenAIFunctionsAgent, AgentExecutor } from "langchain/agents";
import { pull } from "langchain/hub";
import { z } from "zod";
import { DynamicTool, DynamicStructuredTool } from "@langchain/core/tools";
const llm = new ChatOpenAI({
model: "gpt-3.5-turbo",
temperature: 0,
});
const tools = [
new DynamicTool({
name: "FOO",
description:
"call this to get the value of foo. input should be an empty string.",
func: async () => "baz",
}),
new DynamicStructuredTool({
name: "random-number-generator",
description: "generates a random number between two input numbers",
schema: z.object({
low: z.number().describe("The lower bound of the generated number"),
high: z.number().describe("The upper bound of the generated number"),
}),
func: async ({ low, high }) =>
(Math.random() * (high - low) + low).toString(), // Outputs still must be strings
}),
];
// Get the prompt to use - you can modify this!\
// If you want to see the prompt in full, you can at:
// https://smith.langchain.com/hub/hwchase17/openai-functions-agent
const prompt = await pull<ChatPromptTemplate>(
"hwchase17/openai-functions-agent"
);
const agent = await createOpenAIFunctionsAgent({
llm,
tools,
prompt,
});
const agentExecutor = new AgentExecutor({
agent,
tools,
verbose: true,
});
const result = await agentExecutor.invoke({
input: `What is the value of foo?`,
});
console.log(`Got output ${result.output}`);
/*
[chain/start] [1:chain:AgentExecutor] Entering Chain run with input: {
"input": "What is the value of foo?"
}
[agent/action] [1:chain:AgentExecutor] Agent selected action: {
"tool": "FOO",
"toolInput": {},
"log": "Invoking \"FOO\" with {}\n",
"messageLog": [
{
"lc": 1,
"type": "constructor",
"id": [
"langchain_core",
"messages",
"AIMessage"
],
"kwargs": {
"content": "",
"additional_kwargs": {
"function_call": {
"name": "FOO",
"arguments": "{}"
}
}
}
}
]
}
[tool/start] [1:chain:AgentExecutor > 8:tool:FOO] Entering Tool run with input: "undefined"
[tool/end] [1:chain:AgentExecutor > 8:tool:FOO] [113ms] Exiting Tool run with output: "baz"
[chain/end] [1:chain:AgentExecutor] [3.36s] Exiting Chain run with output: {
"input": "What is the value of foo?",
"output": "The value of foo is \"baz\"."
}
Got output The value of foo is "baz".
*/
const result2 = await agentExecutor.invoke({
input: `Generate a random number between 1 and 10.`,
});
console.log(`Got output ${result2.output}`);
/*
[chain/start] [1:chain:AgentExecutor] Entering Chain run with input: {
"input": "Generate a random number between 1 and 10."
}
[agent/action] [1:chain:AgentExecutor] Agent selected action: {
"tool": "random-number-generator",
"toolInput": {
"low": 1,
"high": 10
},
"log": "Invoking \"random-number-generator\" with {\n \"low\": 1,\n \"high\": 10\n}\n",
"messageLog": [
{
"lc": 1,
"type": "constructor",
"id": [
"langchain_core",
"messages",
"AIMessage"
],
"kwargs": {
"content": "",
"additional_kwargs": {
"function_call": {
"name": "random-number-generator",
"arguments": "{\n \"low\": 1,\n \"high\": 10\n}"
}
}
}
}
]
}
[tool/start] [1:chain:AgentExecutor > 8:tool:random-number-generator] Entering Tool run with input: "{"low":1,"high":10}"
[tool/end] [1:chain:AgentExecutor > 8:tool:random-number-generator] [58ms] Exiting Tool run with output: "2.4757639017769293"
[chain/end] [1:chain:AgentExecutor] [3.32s] Exiting Chain run with output: {
"input": "Generate a random number between 1 and 10.",
"output": "The random number generated between 1 and 10 is 2.476."
}
Got output The random number generated between 1 and 10 is 2.476.
*/
API Reference:
- ChatOpenAI from
@langchain/openai
- ChatPromptTemplate from
@langchain/core/prompts
- createOpenAIFunctionsAgent from
langchain/agents
- AgentExecutor from
langchain/agents
- pull from
langchain/hub
- DynamicTool from
@langchain/core/tools
- DynamicStructuredTool from
@langchain/core/tools