CodeAce

CodeAce is a Python package that helps you analyze and understand your codebase using Large Language Models (LLMs). It provides an intuitive interface to map your codebase and query it using natural language.

Features

• 🤖 Multiple LLM providers support (Azure OpenAI, OpenAI, Google Gemini, Anthropic Claude)
• 🔍 Smart code search and analysis
• 💡 Natural language queries about your code
• 📝 Automatic code documentation

Installation

pip install codeace

Quick Start

from dotenv import load_dotenv
from codeace import CoreAgent, MappingAgent

# Load environment variables
load_dotenv()

# Initialize agents
src_path = "path/to/your/code"
model_name = "azure"  # or "openai", "gemini", "anthropic"

# Map your codebase (do this once)
# Note: Mapping process automatically uses GPT-4o-mini for efficient processing
mapping_agent = MappingAgent(model_name=model_name, src_path=src_path)
mapping_agent.run_mapping_process()

# Query your code (uses full model capabilities)
core_agent = CoreAgent(model_name=model_name, src_path=src_path)
result = core_agent.run_core_process("Explain how the error handling works in this codebase")
print(result)

Advanced Usage

Mapping Process

The mapping process uses Azure OpenAI's GPT-4o-mini model for optimal performance and cost efficiency. This process:

• Scans your codebase
• Analyzes each file
• Generates descriptions and function lists
• Creates a searchable index
• Builds a project summary

# Initialize mapping agent
mapping_agent = MappingAgent(model_name="azure", src_path=src_path)

# Run mapping with progress updates
for status in mapping_agent.run_mapping_process():
    print(status)  # Shows progress of file processing

# Optional: Force remapping of all files
mapping_agent.run_mapping_process(override=True)

# Optional: Map without generating summary
mapping_agent.run_mapping_process(generate_summary=False)

Context Management

CodeAce supports rich context management to improve code analysis:

# Initialize agent
core_agent = CoreAgent(model_name="azure", src_path=src_path)

# Add documentation context from file
core_agent.add_extra_context_by_path("path/to/documentation.md")

# Add custom context directly
core_agent.add_extra_context("Additional context information")

# Context is automatically used to improve prompts
improved_prompt = core_agent.improve_user_prompt(user_query)

# Context is considered during code analysis
result = core_agent.process_code_query(user_query, relevant_files)

Multi-Codebase Analysis

You can analyze dependencies across multiple codebases:

# Initialize agents for different codebases
main_agent = CoreAgent(model_name="azure", src_path=main_src)
modules_agent = CoreAgent(model_name="azure", src_path=modules_src)

# Analyze dependencies
dependencies_result = modules_agent.process_dependencies_query(query, files)

# Use dependencies context in main analysis
main_agent.add_extra_context(dependencies_result)
result = main_agent.process_code_query(query, files)

Environment Setup

1. Create a .env file in your project root directory
2. Add the required environment variables based on your chosen LLM provider:

OpenAI

OPENAI_API_KEY=your_api_key_here

Azure OpenAI

AZ_OPENAI_API_KEY=your_azure_api_key_here
AZ_OPENAI_API_BASE=your_azure_endpoint_here
AZ_OPENAI_API_VERSION=your_api_version_here
AZ_OPENAI_LLM_4_O=your_deployment_name_here

Google (Gemini)

GOOGLE_API_KEY=your_google_api_key_here

Anthropic

ANTHROPIC_API_KEY=your_anthropic_api_key_here

Supported LLM Providers

Choose the appropriate model_name when initializing agents:

• "azure": Azure OpenAI (GPT-4o)
• "openai": OpenAI API (GPT-4)
• "gemini": Google Gemini Pro
• "anthropic": Anthropic Claude
• "ollama": Local Ollama models

Requirements

• Python 3.8+

• Required dependencies (installed automatically):

  • langchain
  • pydantic
  • python-dotenv
  • tiktoken
  • (Provider-specific packages based on your choice)

• Built with LangChain

• Supports multiple LLM providers