A workflow demonstrating how the Programming Agent can analyze, generate, execute, and debug code.

​
Quick Start

1

Install Package

First, install the PraisonAI Agents package:

pip install praisonaiagents
2

Set API Key

Set your OpenAI API key as an environment variable: 

export OPENAI_API_KEY=your_api_key_here
3

Create Script

Create a new file code_assistant.py: 

from praisonaiagents import Agent, Tools
from praisonaiagents.tools import execute_code, analyze_code, format_code, lint_code, disassemble_code # Code Tools
from praisonaiagents.tools import execute_command, list_processes, kill_process, get_system_info # Shell Tools
from praisonaiagents.tools import duckduckgo # Web Search Tool

agent = Agent(
    instructions="You are a Programming Agent", self_reflect=True, min_reflect=5, max_reflect=10, 
    tools=[execute_code, analyze_code, format_code, lint_code, disassemble_code, execute_command, list_processes, kill_process, get_system_info, duckduckgo]
)
agent.start(
    "Write a python script using yfinance to find the stock price of Tesla"
    "First check if required packages are installed"
    "Run it using execute_code"
    "execute_command if you want to run any terminal command"
    "search internet using duckduckgo if you want to know update python package information"
    "Analyse the output using analyze_code and fix error if required"
    "if no package is installed, install it"
    "then run the code"
)

​
Understanding Code Development

The Programming Agent combines multiple tools for comprehensive code development:

  1. Code Tools:

    • execute_code: Run Python code
    • analyze_code: Analyze code structure
    • format_code: Format code to standards
    • lint_code: Check code quality
    • disassemble_code: View bytecode

  2. Shell Tools:

    • execute_command: Run terminal commands
    • list_processes: View running processes
    • kill_process: Terminate processes
    • get_system_info: System information

  3. Web Tools:

    • duckduckgo: Search for programming resources

​
Features

Code Execution

Run and test code directly.

Code Analysis

Analyze and improve code quality.

System Integration

Execute system commands and manage processes.

Package Management

Handle dependencies and installations.

​
Example Usage 

# Example: Create and run a data analysis script
response = agent.start("""
    1. Create a script that:
       - Uses pandas for data analysis
       - Reads a CSV file
       - Performs basic statistics
       - Creates a visualization
    2. Check and install required packages
    3. Execute and analyze the results
""")

​
Next Steps