Passing argument to tool

CrewAI Community Support
1

I get this error: ValueError: Required function parameter ‘data’ not found in args_schema
data is an argument to a custom tool that I have created. In the description of Task I mention how data should be passed. I am new to crewAI, see code below.

import warnings
warnings.filterwarnings('ignore')

from crewai import Agent, Task, Crew
from dotenv import load_dotenv
import yfinance as yf
import matplotlib.pyplot as plt
import io
import os

os.environ["OPENAI_MODEL_NAME"] = 'gpt-3.5-turbo'

load_dotenv()
from crewai.tools import BaseTool
class FetchStockDataTool(BaseTool):
    name: str = "Fetch Stock Data"
    description: str = "Fetches stock price and volume data for a given ticker and date range"

    def _run(self, ticker: str, start_date: str, end_date: str):
        data = yf.download(ticker, start=start_date, end=end_date)
        return data

class PlotStockDataTool(BaseTool):
    name: str = "Plot Stock Data"
    description: str = "Plots stock price and volume data using matplotlib"

    def _run(self, data, ticker: str):
        fig, (ax1, ax2) = plt.subplots(2, 1, figsize=(12, 10), sharex=True)
        
        # Plot stock price
        ax1.plot(data.index, data['Close'], label='Close Price')
        ax1.set_title(f'{ticker} Stock Price and Volume')
        ax1.set_ylabel('Price')
        ax1.legend()
        
        # Plot volume
        ax2.bar(data.index, data['Volume'], label='Volume', color='green', alpha=0.7)
        ax2.set_xlabel('Date')
        ax2.set_ylabel('Volume')
        ax2.legend()
        
        plt.tight_layout()
        
        # Save plot to a bytes buffer
        buf = io.BytesIO()
        plt.savefig(buf, format='png')
        buf.seek(0)
        
        return buf


# Create instances of the tools
fetch_stock_tool = FetchStockDataTool()
plot_stock_tool = PlotStockDataTool()

data_analyst_agent = Agent(
    role="Data Analyst",
    goal="Monitor and analyze market data in real-time "
         "to identify trends and predict market movements.",
    backstory="Specializing in financial markets, this agent "
              "uses statistical modeling and machine learning "
              "to provide crucial insights. With a knack for data, "
              "the Data Analyst Agent is the cornerstone for "
              "informing trading decisions.",
    verbose=True,
    allow_delegation=True,

)

# Task for Data Analyst Agent: Analyze Market Data
data_analysis_task = Task(
    description=(
        "1. First, use the 'Fetch Stock Data' tool to get price and volume data for the selected stock ({stock_selection}) for the past 5 days. "
        "2. Save this data in a CSV file named {stock_selection}.csv in the current directory. The CSV file should have three columns: Date, Price, and Volume. "
        "3. Then, use the 'Plot Stock Data' tool. The output of the 'Fetch Stock Data' tool should be used as the 'data' input for the 'Plot Stock Data' tool. "
        "4. Use the stock ticker '{stock_selection}' as the ticker input for the plot tool."

    ),
    expected_output=(
        "A CSV file named {stock_selection}.csv containing 5 rows and 3 columns (Date, Price, Volume). "
        "A plot of the stock's price and volume data."
    ),
    agent=data_analyst_agent,
    tools = [fetch_stock_tool, plot_stock_tool],

)


from crewai import Crew, Process
from langchain_openai import ChatOpenAI

# Define the crew with agents and tasks
financial_trading_crew = Crew(
    agents=[data_analyst_agent],
    
    tasks=[data_analysis_task],
    
    manager_llm=ChatOpenAI(model="gpt-3.5-turbo", 
                           temperature=0.7),
    verbose=True
)

# Example data for kicking off the process
financial_trading_inputs = {
    'stock_selection': 'NVDA',
}

### this execution will take some time to run
result = financial_trading_crew.kickoff(inputs=financial_trading_inputs)
2

The problem is data doesn’t have a description or “shape” so the agent doesn’t know what to pass in. You could add a type to the data in the tool eg

def _run(self, data: str, ticker: str):

Another way is to properly describe the input as a pydantic model

class PlotStockDataToolInput(BaseModel):
    """Input schema for plotting stock data."""

    ticker: str = Field(...,
                            description="Stock ticker")
    data: Optional[str] = Field(
        None, description="Data to plot etc")

You can find more in the docs where you can see more example Tools - CrewAI

Note: my code above is for illustrative purposes only, you will need to change it to suit your needs.

3

Thanks for your assistance. Since I am new, I will need a bit more help. I am not getting the plot. I have changed the code and added your suggestions. “data” is still not being passed. Checking if there is a way to take the output of tool FetchStockDataTool (it returns data) and pass it to PlotStockDataTool? Is the issue with variable type string versus integer in data?

I get this error:

I encountered an error while trying to use the tool. This was the error: string indices must be integers.
Tool Plot Stock Data accepts these inputs: Tool Name: Plot Stock Data
Tool Arguments: {'data': {'description': None, 'type': 'str'}, 'ticker': {'description': None, 'type': 'str'}}      
Tool Description: Plots stock price and volume data using matplotlib

Code:

import warnings
warnings.filterwarnings('ignore')

from crewai import Agent, Task, Crew
from dotenv import load_dotenv
import yfinance as yf
import matplotlib.pyplot as plt
import io
import os 

os.environ["OPENAI_MODEL_NAME"] = 'gpt-3.5-turbo'

load_dotenv()
from crewai.tools import BaseTool
from pydantic import BaseModel, Field
class FetchStockDataTool(BaseTool):
    name: str = "Fetch Stock Data"
    description: str = "Fetches stock price and volume data for a given ticker and date range"
    ticker: str = Field(..., description="Stock ticker symbol")
    def _run(self, ticker: str, start_date: str, end_date: str):
        data = yf.download(ticker, start=start_date, end=end_date)
        return data

class PlotStockDataTool(BaseTool):
    name: str = "Plot Stock Data"
    description: str = "Plots stock price and volume data using matplotlib"
    ticker: str = Field(..., description="Stock ticker symbol")
    data: str = Field(None, description="Stock data to plot in three columns, date, Close price, volume")
    def _run(self, data: str, ticker: str):
        fig, (ax1, ax2) = plt.subplots(2, 1, figsize=(12, 10), sharex=True)
        
        # Plot stock price
        ax1.plot(data.index, data['Close'], label='Close Price')
        ax1.set_title(f'{ticker} Stock Price and Volume')
        ax1.set_ylabel('Price')
        ax1.legend()
        
        # Plot volume
        ax2.bar(data.index, data['Volume'], label='Volume', color='green', alpha=0.7)
        ax2.set_xlabel('Date')
        ax2.set_ylabel('Volume')
        ax2.legend()
        
        plt.tight_layout()
        
        # Save plot to a bytes buffer
        buf = io.BytesIO()
        plt.savefig(buf, format='png')
        buf.seek(0)
        
        return buf


# Create instances of the tools
fetch_stock_tool = FetchStockDataTool(ticker="{stock_selection}")
plot_stock_tool = PlotStockDataTool(ticker="{stock_selection}")

data_analyst_agent = Agent(
    role="Data Analyst",
    goal="Monitor and analyze market data in real-time "
         "to identify trends and predict market movements.",
    backstory="Specializing in financial markets, this agent "
              "uses statistical modeling and machine learning "
              "to provide crucial insights. With a knack for data, "
              "the Data Analyst Agent is the cornerstone for "
              "informing trading decisions.",
    verbose=True,
    allow_delegation=True,

)

# Task for Data Analyst Agent: Analyze Market Data
data_analysis_task = Task(
    description=(
        "1. First, use the 'Fetch Stock Data' tool to get price and volume data for the selected stock ({stock_selection}) for the past 5 days. "
        "2. Save this data in a CSV file named {stock_selection}.csv in the current directory. The CSV file should have three columns: Date, Price, and Volume. "
        "3. Then, use the 'Plot Stock Data' tool. The output of the 'Fetch Stock Data' tool should be used as the 'data' input for the 'Plot Stock Data' tool. "
        "4. Use the stock ticker '{stock_selection}' as the ticker input for the plot tool."

    ),
    expected_output=(
        "A CSV file named {stock_selection}.csv containing 5 rows and 3 columns (Date, Price, Volume). "
        "A plot of the stock's price and volume data."
    ),
    agent=data_analyst_agent,
    tools = [fetch_stock_tool, plot_stock_tool],

)


from crewai import Crew, Process
from langchain_openai import ChatOpenAI

# Define the crew with agents and tasks
financial_trading_crew = Crew(
    agents=[data_analyst_agent #, 
            #trading_strategy_agent, 
            #execution_agent
            #, risk_management_agent
            ],
    
    tasks=[data_analysis_task #, 
           #strategy_development_task, 
           #execution_planning_task
           #, risk_assessment_task
           ],
    
    manager_llm=ChatOpenAI(model="gpt-3.5-turbo", 
                           temperature=0.7),
    verbose=True
)

# Example data for kicking off the process
financial_trading_inputs = {
    'stock_selection': 'NVDA',
}

### this execution will take some time to run
result = financial_trading_crew.kickoff(inputs=financial_trading_inputs)
4

You need to set data in the shape you want it to be. I used string just as an example, so it will not work.

I assume data is a dict and not a string. Since I don’t know what you want to pass in I can’t write the code for you.