mirror of
https://github.com/langchain-ai/langgraph-example.git
synced 2026-07-01 16:12:43 -04:00
Harrison Chase
201 lines
6.7 KiB
Plaintext
201 lines
6.7 KiB
Plaintext
{
|
|
"cells": [
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "68c0837d-c40a-4209-9f88-5d08c00c31b0",
|
|
"metadata": {},
|
|
"source": [
|
|
"# How to create agents with configuration\n",
|
|
"\n",
|
|
"One of the benefits of LangGraph API is that it lets you create agents with different configurations.\n",
|
|
"This is useful when you want to:\n",
|
|
"\n",
|
|
"- Define a cognitive architecture once as a LangGraph\n",
|
|
"- Let that LangGraph be configurable across some attributes (for example, system message or LLM to use)\n",
|
|
"- Let users create agents with arbitrary configurations, save them, and then use them in the future\n",
|
|
"\n",
|
|
"In this guide we will show how to do that for the default agent we have built in.\n",
|
|
"\n",
|
|
"If you look at the agent we defined, you can see that inside the `call_model` node we have created the model based on some configuration. That node looks like:\n",
|
|
"\n",
|
|
"```python\n",
|
|
"def call_model(state, config):\n",
|
|
" messages = state[\"messages\"]\n",
|
|
" model_name = config.get('configurable', {}).get(\"model_name\", \"anthropic\")\n",
|
|
" model = _get_model(model_name)\n",
|
|
" response = model.invoke(messages)\n",
|
|
" # We return a list, because this will get added to the existing list\n",
|
|
" return {\"messages\": [response]}\n",
|
|
"```\n",
|
|
"\n",
|
|
"We are looking inside the config for a `model_name` parameter (which defaults to `anthropic` if none is found).\n",
|
|
"That means that by default we are using Anthropic as our model provider.\n",
|
|
"In this example we will see an example of how to create an example agent that is configured to use OpenAI.\n",
|
|
"\n",
|
|
"We've also communicated to the graph that it should expect configuration with this key. \n",
|
|
"We've done this by passing `config_schema` when constructing the graph, eg:\n",
|
|
"\n",
|
|
"```python\n",
|
|
"class GraphConfig(TypedDict):\n",
|
|
" model_name: Literal[\"anthropic\", \"openai\"]\n",
|
|
"\n",
|
|
"\n",
|
|
"# Define a new graph\n",
|
|
"workflow = StateGraph(AgentState, config_schema=GraphConfig)\n",
|
|
"```"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 16,
|
|
"id": "f69c9a4f-2ef9-4998-827b-fe86d12bfd76",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"from langgraph_sdk import get_client\n",
|
|
"\n",
|
|
"client = get_client()"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 14,
|
|
"id": "9a37bfb5-7331-4004-8054-508838e54f18",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"# First, let's check what valid configuration can be\n",
|
|
"# We can do this by getting the default assistant\n",
|
|
"# There should always be a default assistant with no configuration\n",
|
|
"assistants = await client.assistants.search()\n",
|
|
"assistants = [a for a in assistants if not a['config']]\n",
|
|
"base_assistant = assistants[0]"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 17,
|
|
"id": "70193a08-127c-44b3-a102-10db260d7e3b",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"data": {
|
|
"text/plain": [
|
|
"{'model_name': {'title': 'Model Name',\n",
|
|
" 'enum': ['anthropic', 'openai'],\n",
|
|
" 'type': 'string'}}"
|
|
]
|
|
},
|
|
"execution_count": 17,
|
|
"metadata": {},
|
|
"output_type": "execute_result"
|
|
}
|
|
],
|
|
"source": [
|
|
"# We can now call `.get_schemas` to get schemas associated with this graph\n",
|
|
"schemas = await client.assistants.get_schemas(assistant_id=base_assistant[\"assistant_id\"])\n",
|
|
"# There are multiple types of schemas\n",
|
|
"# We can get the `config_schema` to look at the the configurable parameters\n",
|
|
"schemas['config_schema']['definitions']['Configurable']['properties']"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 18,
|
|
"id": "99be5aee-9a6b-4515-b72f-ba135a893c65",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"assistant = await client.assistants.create(graph_id=\"agent\", config={\"configurable\": {\"model_name\": \"openai\"}})"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "4f10d346-69e6-44f4-8ff0-ef539ba938df",
|
|
"metadata": {},
|
|
"source": [
|
|
"We can see that this assistant has saved the config"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 20,
|
|
"id": "3898ca35-eb2c-4b12-97ea-e0cc6a7c6a2e",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"data": {
|
|
"text/plain": [
|
|
"{'assistant_id': '40a3a2bf-5319-4fae-a2ac-05e075615cdc',\n",
|
|
" 'graph_id': 'agent',\n",
|
|
" 'config': {'configurable': {'model_name': 'openai'}},\n",
|
|
" 'created_at': '2024-06-05T23:12:30.519458+00:00',\n",
|
|
" 'updated_at': '2024-06-05T23:12:30.519458+00:00',\n",
|
|
" 'metadata': {}}"
|
|
]
|
|
},
|
|
"execution_count": 20,
|
|
"metadata": {},
|
|
"output_type": "execute_result"
|
|
}
|
|
],
|
|
"source": [
|
|
"assistant"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 21,
|
|
"id": "68ed7a1b-74be-4560-8c55-c76d49d3d348",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"StreamPart(event='metadata', data={'run_id': '1ef23911-c23b-6d8c-b1dc-94bb982ca7b1'})\n",
|
|
"StreamPart(event='values', data={'messages': [{'role': 'user', 'content': 'who made you?'}]})\n",
|
|
"StreamPart(event='values', data={'messages': [{'content': 'who made you?', 'additional_kwargs': {}, 'response_metadata': {}, 'type': 'human', 'name': None, 'id': 'ed93c1c9-80d6-4f2b-a048-ef859ea533f9', 'example': False}, {'content': 'I was created by OpenAI, a research organization focused on developing and advancing artificial intelligence technology.', 'additional_kwargs': {}, 'response_metadata': {'finish_reason': 'stop'}, 'type': 'ai', 'name': None, 'id': 'run-6560cd65-5c9c-434b-8835-0baadc684760', 'example': False, 'tool_calls': [], 'invalid_tool_calls': [], 'usage_metadata': None}]})\n",
|
|
"StreamPart(event='end', data=None)\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"thread = await client.threads.create()\n",
|
|
"input = {\"messages\": [{\"role\": \"user\", \"content\": \"who made you?\"}]}\n",
|
|
"async for event in client.runs.stream(thread['thread_id'], assistant['assistant_id'], input=input):\n",
|
|
" print(event)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "666d78f1-019a-433e-839e-52d2ebb3d9c8",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": []
|
|
}
|
|
],
|
|
"metadata": {
|
|
"kernelspec": {
|
|
"display_name": "Python 3 (ipykernel)",
|
|
"language": "python",
|
|
"name": "python3"
|
|
},
|
|
"language_info": {
|
|
"codemirror_mode": {
|
|
"name": "ipython",
|
|
"version": 3
|
|
},
|
|
"file_extension": ".py",
|
|
"mimetype": "text/x-python",
|
|
"name": "python",
|
|
"nbconvert_exporter": "python",
|
|
"pygments_lexer": "ipython3",
|
|
"version": "3.11.1"
|
|
}
|
|
},
|
|
"nbformat": 4,
|
|
"nbformat_minor": 5
|
|
}
|