BabyAGI with Tools

This notebook builds on top of baby agi, but shows how you can swap out the execution chain. The previous execution chain was just an LLM which made stuff up. By swapping it out with an agent that has access to tools, we can hopefully get real reliable information

Install and Import Required Modules

import os
from collections import deque
from typing import Dict, List, Optional, Any

from langchain import LLMChain, OpenAI, PromptTemplate
from langchain.embeddings import OpenAIEmbeddings
from langchain.llms import BaseLLM
from langchain.vectorstores.base import VectorStore
from pydantic import BaseModel, Field
from langchain.chains.base import Chain

Connect to the Vector Store

Depending on what vectorstore you use, this step may look different.

%pip install faiss-cpu > /dev/null
%pip install google-search-results > /dev/null
from langchain.vectorstores import FAISS
from langchain.docstore import InMemoryDocstore

# Define your embedding model
embeddings_model = OpenAIEmbeddings()
# Initialize the vectorstore as empty
import faiss

embedding_size = 1536
index = faiss.IndexFlatL2(embedding_size)
vectorstore = FAISS(embeddings_model.embed_query, index, InMemoryDocstore({}), {})

Define the Chains

BabyAGI relies on three LLM chains:

• Task creation chain to select new tasks to add to the list
• Task prioritization chain to re-prioritize tasks
• Execution Chain to execute the tasks

NOTE: in this notebook, the Execution chain will now be an agent.

class TaskCreationChain(LLMChain):
    """Chain to generates tasks."""

    @classmethod
    def from_llm(cls, llm: BaseLLM, verbose: bool = True) -> LLMChain:
        """Get the response parser."""
        task_creation_template = (
            "You are an task creation AI that uses the result of an execution agent"
            " to create new tasks with the following objective: {objective},"
            " The last completed task has the result: {result}."
            " This result was based on this task description: {task_description}."
            " These are incomplete tasks: {incomplete_tasks}."
            " Based on the result, create new tasks to be completed"
            " by the AI system that do not overlap with incomplete tasks."
            " Return the tasks as an array."
        )
        prompt = PromptTemplate(
            template=task_creation_template,
            input_variables=[
                "result",
                "task_description",
                "incomplete_tasks",
                "objective",
            ],
        )
        return cls(prompt=prompt, llm=llm, verbose=verbose)

class TaskPrioritizationChain(LLMChain):
    """Chain to prioritize tasks."""

    @classmethod
    def from_llm(cls, llm: BaseLLM, verbose: bool = True) -> LLMChain:
        """Get the response parser."""
        task_prioritization_template = (
            "You are an task prioritization AI tasked with cleaning the formatting of and reprioritizing"
            " the following tasks: {task_names}."
            " Consider the ultimate objective of your team: {objective}."
            " Do not remove any tasks. Return the result as a numbered list, like:"
            " #. First task"
            " #. Second task"
            " Start the task list with number {next_task_id}."
        )
        prompt = PromptTemplate(
            template=task_prioritization_template,
            input_variables=["task_names", "next_task_id", "objective"],
        )
        return cls(prompt=prompt, llm=llm, verbose=verbose)

from langchain.agents import ZeroShotAgent, Tool, AgentExecutor
from langchain import OpenAI, SerpAPIWrapper, LLMChain

todo_prompt = PromptTemplate.from_template(
    "You are a planner who is an expert at coming up with a todo list for a given objective. Come up with a todo list for this objective: {objective}"
)
todo_chain = LLMChain(llm=OpenAI(temperature=0), prompt=todo_prompt)
search = SerpAPIWrapper()
tools = [
    Tool(
        name="Search",
        func=search.run,
        description="useful for when you need to answer questions about current events",
    ),
    Tool(
        name="TODO",
        func=todo_chain.run,
        description="useful for when you need to come up with todo lists. Input: an objective to create a todo list for. Output: a todo list for that objective. Please be very clear what the objective is!",
    ),
]


prefix = """You are an AI who performs one task based on the following objective: {objective}. Take into account these previously completed tasks: {context}."""
suffix = """Question: {task}
{agent_scratchpad}"""
prompt = ZeroShotAgent.create_prompt(
    tools,
    prefix=prefix,
    suffix=suffix,
    input_variables=["objective", "task", "context", "agent_scratchpad"],
)

Define the BabyAGI Controller

BabyAGI composes the chains defined above in a (potentially-)infinite loop.

def get_next_task(
    task_creation_chain: LLMChain,
    result: Dict,
    task_description: str,
    task_list: List[str],
    objective: str,
) -> List[Dict]:
    """Get the next task."""
    incomplete_tasks = ", ".join(task_list)
    response = task_creation_chain.run(
        result=result,
        task_description=task_description,
        incomplete_tasks=incomplete_tasks,
        objective=objective,
    )
    new_tasks = response.split("\n")
    return [{"task_name": task_name} for task_name in new_tasks if task_name.strip()]

def prioritize_tasks(
    task_prioritization_chain: LLMChain,
    this_task_id: int,
    task_list: List[Dict],
    objective: str,
) -> List[Dict]:
    """Prioritize tasks."""
    task_names = [t["task_name"] for t in task_list]
    next_task_id = int(this_task_id) + 1
    response = task_prioritization_chain.run(
        task_names=task_names, next_task_id=next_task_id, objective=objective
    )
    new_tasks = response.split("\n")
    prioritized_task_list = []
    for task_string in new_tasks:
        if not task_string.strip():
            continue
        task_parts = task_string.strip().split(".", 1)
        if len(task_parts) == 2:
            task_id = task_parts[0].strip()
            task_name = task_parts[1].strip()
            prioritized_task_list.append({"task_id": task_id, "task_name": task_name})
    return prioritized_task_list

def _get_top_tasks(vectorstore, query: str, k: int) -> List[str]:
    """Get the top k tasks based on the query."""
    results = vectorstore.similarity_search_with_score(query, k=k)
    if not results:
        return []
    sorted_results, _ = zip(*sorted(results, key=lambda x: x[1], reverse=True))
    return [str(item.metadata["task"]) for item in sorted_results]


def execute_task(
    vectorstore, execution_chain: LLMChain, objective: str, task: str, k: int = 5
) -> str:
    """Execute a task."""
    context = _get_top_tasks(vectorstore, query=objective, k=k)
    return execution_chain.run(objective=objective, context=context, task=task)

class BabyAGI(Chain, BaseModel):
    """Controller model for the BabyAGI agent."""

    task_list: deque = Field(default_factory=deque)
    task_creation_chain: TaskCreationChain = Field(...)
    task_prioritization_chain: TaskPrioritizationChain = Field(...)
    execution_chain: AgentExecutor = Field(...)
    task_id_counter: int = Field(1)
    vectorstore: VectorStore = Field(init=False)
    max_iterations: Optional[int] = None

    class Config:
        """Configuration for this pydantic object."""

        arbitrary_types_allowed = True

    def add_task(self, task: Dict):
        self.task_list.append(task)

    def print_task_list(self):
        print("\033[95m\033[1m" + "\n*****TASK LIST*****\n" + "\033[0m\033[0m")
        for t in self.task_list:
            print(str(t["task_id"]) + ": " + t["task_name"])

    def print_next_task(self, task: Dict):
        print("\033[92m\033[1m" + "\n*****NEXT TASK*****\n" + "\033[0m\033[0m")
        print(str(task["task_id"]) + ": " + task["task_name"])

    def print_task_result(self, result: str):
        print("\033[93m\033[1m" + "\n*****TASK RESULT*****\n" + "\033[0m\033[0m")
        print(result)

    @property
    def input_keys(self) -> List[str]:
        return ["objective"]

    @property
    def output_keys(self) -> List[str]:
        return []

    def _call(self, inputs: Dict[str, Any]) -> Dict[str, Any]:
        """Run the agent."""
        objective = inputs["objective"]
        first_task = inputs.get("first_task", "Make a todo list")
        self.add_task({"task_id": 1, "task_name": first_task})
        num_iters = 0
        while True:
            if self.task_list:
                self.print_task_list()

                # Step 1: Pull the first task
                task = self.task_list.popleft()
                self.print_next_task(task)

                # Step 2: Execute the task
                result = execute_task(
                    self.vectorstore, self.execution_chain, objective, task["task_name"]
                )
                this_task_id = int(task["task_id"])
                self.print_task_result(result)

                # Step 3: Store the result in Pinecone
                result_id = f"result_{task['task_id']}"
                self.vectorstore.add_texts(
                    texts=[result],
                    metadatas=[{"task": task["task_name"]}],
                    ids=[result_id],
                )

                # Step 4: Create new tasks and reprioritize task list
                new_tasks = get_next_task(
                    self.task_creation_chain,
                    result,
                    task["task_name"],
                    [t["task_name"] for t in self.task_list],
                    objective,
                )
                for new_task in new_tasks:
                    self.task_id_counter += 1
                    new_task.update({"task_id": self.task_id_counter})
                    self.add_task(new_task)
                self.task_list = deque(
                    prioritize_tasks(
                        self.task_prioritization_chain,
                        this_task_id,
                        list(self.task_list),
                        objective,
                    )
                )
            num_iters += 1
            if self.max_iterations is not None and num_iters == self.max_iterations:
                print(
                    "\033[91m\033[1m" + "\n*****TASK ENDING*****\n" + "\033[0m\033[0m"
                )
                break
        return {}

    @classmethod
    def from_llm(
        cls, llm: BaseLLM, vectorstore: VectorStore, verbose: bool = False, **kwargs
    ) -> "BabyAGI":
        """Initialize the BabyAGI Controller."""
        task_creation_chain = TaskCreationChain.from_llm(llm, verbose=verbose)
        task_prioritization_chain = TaskPrioritizationChain.from_llm(
            llm, verbose=verbose
        )
        llm_chain = LLMChain(llm=llm, prompt=prompt)
        tool_names = [tool.name for tool in tools]
        agent = ZeroShotAgent(llm_chain=llm_chain, allowed_tools=tool_names)
        agent_executor = AgentExecutor.from_agent_and_tools(
            agent=agent, tools=tools, verbose=True
        )
        return cls(
            task_creation_chain=task_creation_chain,
            task_prioritization_chain=task_prioritization_chain,
            execution_chain=agent_executor,
            vectorstore=vectorstore,
            **kwargs,
        )

Run the BabyAGI

Now it's time to create the BabyAGI controller and watch it try to accomplish your objective.

OBJECTIVE = "Write a weather report for SF today"

llm = OpenAI(temperature=0)

# Logging of LLMChains
verbose = False
# If None, will keep on going forever
max_iterations: Optional[int] = 3
baby_agi = BabyAGI.from_llm(
    llm=llm, vectorstore=vectorstore, verbose=verbose, max_iterations=max_iterations
)

baby_agi({"objective": OBJECTIVE})


*****TASK LIST*****

1: Make a todo list

*****NEXT TASK*****

1: Make a todo list


> Entering new AgentExecutor chain...
Thought: I need to gather data on the current weather conditions in SF
Action: Search
Action Input: Current weather conditions in SF
Observation: High 67F. Winds WNW at 10 to 15 mph. Clear to partly cloudy.
Thought: I need to make a todo list
Action: TODO
Action Input: Write a weather report for SF today
Observation: 

1. Research current weather conditions in San Francisco
2. Gather data on temperature, humidity, wind speed, and other relevant weather conditions
3. Analyze data to determine current weather trends
4. Write a brief introduction to the weather report
5. Describe current weather conditions in San Francisco
6. Discuss any upcoming weather changes
7. Summarize the weather report
8. Proofread and edit the report
9. Submit the report
Thought: I now know the final answer
Final Answer: A weather report for SF today should include research on current weather conditions in San Francisco, gathering data on temperature, humidity, wind speed, and other relevant weather conditions, analyzing data to determine current weather trends, writing a brief introduction to the weather report, describing current weather conditions in San Francisco, discussing any upcoming weather changes, summarizing the weather report, proofreading and editing the report, and submitting the report.

> Finished chain.

*****TASK RESULT*****

A weather report for SF today should include research on current weather conditions in San Francisco, gathering data on temperature, humidity, wind speed, and other relevant weather conditions, analyzing data to determine current weather trends, writing a brief introduction to the weather report, describing current weather conditions in San Francisco, discussing any upcoming weather changes, summarizing the weather report, proofreading and editing the report, and submitting the report.

*****TASK LIST*****

2: Gather data on temperature, humidity, wind speed, and other relevant weather conditions
3: Analyze data to determine current weather trends
4: Write a brief introduction to the weather report
5: Describe current weather conditions in San Francisco
6: Discuss any upcoming weather changes
7: Summarize the weather report
8: Proofread and edit the report
9: Submit the report
1: Research current weather conditions in San Francisco

*****NEXT TASK*****

2: Gather data on temperature, humidity, wind speed, and other relevant weather conditions


> Entering new AgentExecutor chain...
Thought: I need to search for the current weather conditions in SF
Action: Search
Action Input: Current weather conditions in SF
Observation: High 67F. Winds WNW at 10 to 15 mph. Clear to partly cloudy.
Thought: I need to make a todo list
Action: TODO
Action Input: Create a weather report for SF today
Observation: 

1. Gather current weather data for SF, including temperature, wind speed, humidity, and precipitation.
2. Research historical weather data for SF to compare current conditions.
3. Analyze current and historical data to determine any trends or patterns.
4. Create a visual representation of the data, such as a graph or chart.
5. Write a summary of the weather report, including key findings and any relevant information.
6. Publish the weather report on a website or other platform.
Thought: I now know the final answer
Final Answer: Today in San Francisco, the temperature is 67F with winds WNW at 10 to 15 mph. The sky is clear to partly cloudy.

> Finished chain.

*****TASK RESULT*****

Today in San Francisco, the temperature is 67F with winds WNW at 10 to 15 mph. The sky is clear to partly cloudy.

*****TASK LIST*****

3: Research current weather conditions in San Francisco
4: Compare the current weather conditions in San Francisco to the average for this time of year.
5: Identify any potential weather-related hazards in the area.
6: Research any historical weather patterns in San Francisco.
7: Analyze data to determine current weather trends
8: Include any relevant data from nearby cities in the report.
9: Include any relevant data from the National Weather Service in the report.
10: Include any relevant data from local news sources in the report.
11: Include any relevant data from online weather sources in the report.
12: Include any relevant data from local meteorologists in the report.
13: Include any relevant data from local weather stations in the report.
14: Include any relevant data from satellite images in the report.
15: Describe current weather conditions in San Francisco
16: Discuss any upcoming weather changes
17: Write a brief introduction to the weather report
18: Summarize the weather report
19: Proofread and edit the report
20: Submit the report

*****NEXT TASK*****

3: Research current weather conditions in San Francisco


> Entering new AgentExecutor chain...
Thought: I need to search for current weather conditions in San Francisco
Action: Search
Action Input: Current weather conditions in San Francisco
Observation: TodaySun 04/09 High 67 · 1% Precip. ; TonightSun 04/09 Low 49 · 9% Precip. ; TomorrowMon 04/10 High 64 · 11% Precip.
Thought: I now know the final answer
Final Answer: Today in San Francisco, the high temperature is 67 degrees with 1% chance of precipitation. The low temperature tonight is 49 degrees with 9% chance of precipitation. Tomorrow's high temperature is 64 degrees with 11% chance of precipitation.

> Finished chain.

*****TASK RESULT*****

Today in San Francisco, the high temperature is 67 degrees with 1% chance of precipitation. The low temperature tonight is 49 degrees with 9% chance of precipitation. Tomorrow's high temperature is 64 degrees with 11% chance of precipitation.

*****TASK ENDING*****






{'objective': 'Write a weather report for SF today'}