daveagent-cli 1.0.0

AI-powered coding assistant with intelligent agent orchestration

🤖 DaveAgent - AI-Powered Coding Assistant

DaveAgent is an intelligent AI-powered coding assistant that works in your current directory. It uses AutoGen 0.4 to orchestrate specialized agents that help you with development tasks.

✨ Features

• 🚀 Global CLI Command: Use daveagent from any directory
• 📂 Contextual Work: Operates in your current directory automatically
• 🧠 Vector Memory with ChromaDB: Remembers conversations, code, and decisions between sessions
• 🔍 CodeSearcher: Specialized agent for searching and analyzing code
• 📎 File Mentions with @: Mention specific files with @ to give them maximum priority in context
• 🔧 45+ Integrated Tools: Filesystem, Git, JSON, CSV, Web, Memory (RAG), and more
• 🤖 Intelligent Agents: Automatic selection of the appropriate agent
• 📊 Complete Logging: Detailed logging system for debugging
• 🎨 Rich Interface: CLI with colors and formatting using Rich
• ⚡ Real-time Visualization: See the agent's thoughts and actions while it works

🎯 Use Cases

Software Development

cd my-project
daveagent

# Search code before modifying
You: /search current authentication system

# Mention specific files with @
You: @main.py fix the authentication bug in this file
You: @config.py @.env update the API configuration

# Modify with context
You: create an authentication module with JWT
You: refactor the code in services/ to use async/await
You: find all TODOs in the project

Data Analysis

cd data-project
daveagent

You: read the sales.csv file and show a summary
You: combine all CSVs in the data/ folder into one
You: convert the configuration JSON to CSV

Git Operations

cd my-repo
daveagent

You: commit the changes with a descriptive message
You: show the diff of the last 3 commits
You: create a branch feature/new-functionality

📦 Installation

Installation from PyPI (Coming Soon)

When published on PyPI:

pip install daveagent-ai
daveagent

Installation from Source Code

# 1. Clone or download the project
git clone https://github.com/davidmonterocrespo24/DaveAgent.git
cd DaveAgent

# 2. Install in development mode
pip install -e .

# 3. Use from any directory!
daveagent

Requirements

• Python 3.10 or higher
• pip (Python package manager)

Main Dependencies

• autogen-agentchat>=0.4.0 - Agent framework
• autogen-ext[openai]>=0.4.0 - Model extensions
• prompt-toolkit>=3.0.0 - Command-line interface
• rich>=13.0.0 - Formatting and colors
• pandas>=2.0.0 - Data processing

See INSTALACION.md for detailed instructions.

🚀 Usage

Basic Command

# From any directory
cd your-project
daveagent

Options

# Debug mode (detailed logs)
daveagent --debug

# View version
daveagent --version

# View help
daveagent --help

Internal Commands

Within DaveAgent, you can use these commands:

Command	Description
/help	Show command help
/search <query>	🔍 Search and analyze code
/index	🧠 Index project in vector memory
/memory	📊 Show memory statistics
@<file>	📎 Mention specific file with high priority
/debug	Enable/disable debug mode
/logs	Show logs location
/stats	Show statistics
/clear	Clear history
/new	New conversation
/exit	Exit DaveAgent

🔍 /search Command

The /search command invokes the CodeSearcher agent to search and analyze code:

You: /search authentication function
You: /search where is the TaskPlanner class used
You: /search how does the logging system work

CodeSearcher provides you with:

• 📍 Relevant files with exact locations
• 🔧 Functions found with complete code
• 📦 Important variables and constants
• 🔗 Dependencies between components
• 💡 Recommendations on what to modify

See docs/CODESEARCHER_GUIDE.md for more details.

📎 File Mentions with @

Mention specific files in your query using @:

You: @main.py explain how this file works
You: @config.py @.env update the database connection settings
You: @src/agents/code_searcher.py add docstrings to all methods

Features:

• ✅ Interactive selector with keyboard navigation (↑↓)
• ✅ Real-time search and filtering
• ✅ Mentioned files have maximum priority in context
• ✅ Supports multiple files in a single query
• ✅ Automatically excludes hidden and binary files

See docs/FILE_MENTIONS.md and examples/file_mentions_demo.md for detailed examples.

🧠 Vector Memory System

DaveAgent uses ChromaDB to maintain persistent memory between sessions:

# Index your project once
You: /index
📚 Indexing project in vector memory...
✅ Indexing completed!
  • Indexed files: 45
  • Chunks created: 234

# View memory statistics
You: /memory
🧠 Vector Memory Statistics

📚 Active memory system with 4 collections:
  • Conversations: Conversation history
  • Codebase: Indexed source code
  • Decisions: Architectural decisions
  • Preferences: User preferences

Memory Benefits:

• 💬 Conversations: Remembers previous interactions and maintains context
• 📝 Code Base: Semantic searches in your code without grep
• 🎯 Decisions: Maintains consistency in architectural decisions
• ⚙️ Preferences: Learns your preferred coding style

Agents use memory automatically:

• CodeSearcher: Queries indexed code for faster searches
• Coder: Remembers previous solutions and style preferences
• PlanningAgent: Maintains consistency with past decisions

See docs/MEMORY_SYSTEM.md for complete documentation and examples/memory_usage_example.py for usage examples.

🛠️ Available Tools

Filesystem (7 tools)

• read_file - Read files with line range support
• write_file - Write/create files
• edit_file - Edit files with search & replace
• list_dir - List directory contents
• delete_file - Delete files safely
• file_search - Search for files by name
• glob_search - Search files using glob patterns

Git (8 tools)

• git_status - Repository status
• git_add - Stage files
• git_commit - Create commits
• git_push - Push changes to remote
• git_pull - Pull changes from remote
• git_log - View commit history
• git_branch - Manage branches
• git_diff - View differences

JSON (8 tools)

• read_json - Read and parse JSON
• write_json - Write JSON files
• merge_json_files - Combine multiple JSONs
• validate_json - Validate JSON syntax
• format_json - Format/prettify JSON
• json_get_value - Extract values by path
• json_set_value - Set values by path
• json_to_text - Convert to readable text

CSV (7 tools)

• read_csv - Read CSV files
• write_csv - Write CSV files
• csv_info - Get CSV information/statistics
• filter_csv - Filter rows by conditions
• merge_csv_files - Combine multiple CSVs
• csv_to_json - Convert CSV to JSON
• sort_csv - Sort rows by column

Web (7 tools)

• wiki_search - Search Wikipedia articles
• wiki_summary - Get article summary
• wiki_content - Get full article content
• wiki_page_info - Get page metadata
• wiki_random - Get random article
• wiki_set_language - Change Wikipedia language
• web_search - General web search

Analysis (5 tools)

• analyze_python_file - Analyze Python code structure
• find_function_definition - Find function/class definitions
• list_all_functions - List all functions in file
• grep_search - Search text with patterns
• run_terminal_cmd - Execute shell commands

Memory (RAG) (8 tools)

• query_conversation_memory - Search conversation history
• query_codebase_memory - Search indexed code
• query_decision_memory - Search architectural decisions
• query_preferences_memory - Search user preferences
• query_user_memory - Search user information
• save_user_info - Store user information
• save_decision - Record architectural decision
• save_preference - Save user preference

📖 Examples

Example 1: Use CodeSearcher before modifying

cd my-project
daveagent

# First, search for context
You: /search existing utilities system

# The agent shows functions, files and current structure
# Now modify with context

You: create a utils.py module with functions for:
    - validate email
    - format dates
    - calculate MD5 hash

DaveAgent first analyzes the existing code and then creates the my-project/utils.py file with the requested functions, avoiding duplicates and maintaining consistency.

Example 2: Analyze a Project

cd existing-project
daveagent

You: analyze the project structure and give me a summary
You: how many functions are there in total?
You: find all files that use the requests library

Example 3: Data Operations

cd data
daveagent

You: read the sales.csv file and show the top 10 sales
You: create a new CSV with only 2024 sales
You: convert the config.json file to CSV

🐛 Debugging and Logs

View Logs

# Start with detailed logs
daveagent --debug

# Within DaveAgent
You: /logs
📄 Log file: logs/daveagent_20250131_154022.log

Log Location

Logs are saved in:

logs/
└── daveagent_YYYYMMDD_HHMMSS.log

Each file contains detailed logs with format:

2025-01-31 15:40:22 | DaveAgent | INFO | process_user_request:257 | 📝 New request...

See LOGGING_GUIDE.md for more details.

🏗️ Architecture

DaveAgent/
├── src/
│   ├── agents/          # Specialized agents
│   │   ├── task_planner.py      # Task planning
│   │   ├── task_executor.py     # Task execution
│   │   └── code_searcher.py     # 🔍 Code search
│   ├── config/          # Configuration and prompts
│   ├── interfaces/      # CLI interface
│   ├── managers/        # Conversation management
│   ├── tools/           # 45+ tools
│   │   ├── filesystem/
│   │   ├── git/
│   │   ├── data/       # JSON, CSV
│   │   ├── web/        # Wikipedia, web search
│   │   └── analysis/   # Code analysis, grep, terminal
│   ├── utils/          # Utilities (logger)
│   └── cli.py          # CLI entry point
├── docs/               # Documentation
│   └── CODESEARCHER_GUIDE.md  # CodeSearcher Guide
└── main.py             # Main application

🔧 Configuration

API Key

DaveAgent uses DeepSeek by default. To change the model:

1. Edit main.py:

self.model_client = OpenAIChatCompletionClient(
    model="gpt-4",  # Change here
    api_key="your-api-key",
    # ...
)

2. Or use environment variables in .daveagent/.env:

DAVEAGENT_API_KEY=your-api-key
DAVEAGENT_MODEL=gpt-4
DAVEAGENT_BASE_URL=https://api.openai.com/v1

SSL Issues (Corporate Networks)

If you experience SSL certificate errors:

1. Method 1: Environment variable in .daveagent/.env:

DAVEAGENT_SSL_VERIFY=false

2. Method 2: Command line argument:

daveagent --no-ssl-verify
# or
daveagent --ssl-verify=false

3. Method 3: System environment variable:

export DAVEAGENT_SSL_VERIFY=false  # Linux/macOS
set DAVEAGENT_SSL_VERIFY=false     # Windows

🤝 Contributing

Contributions are welcome! To contribute:

1. Fork the repository
2. Create a branch: git checkout -b feature/new-functionality
3. Commit your changes: git commit -m 'Add new functionality'
4. Push to the branch: git push origin feature/new-functionality
5. Open a Pull Request

Development

# Install with development dependencies
pip install -e ".[dev]"

# Run tests
pytest

# Format code
black src/

# Check types
mypy src/

📚 Documentation

User Guides

• Installation Guide - Detailed installation
• CodeSearcher Guide - 🔍 Code search and analysis
• File Mentions Guide - 📎 Mention files with @
• File Mentions Demo - Interactive examples
• Logging Guide - Logging system
• Real-time Visualization - See agent thoughts
• Changes Made - Change history
• Implemented Improvements - Technical analysis

For Developers

• Publish to PyPI - Complete guide to publish to PyPI
• PyPI Quick Start - Publish in 10 minutes
• Agent Integration - Agent team architecture

🧪 Evaluation with SWE-bench (Linux)

To evaluate the agent's performance using the SWE-bench Verified standard, we have included an automated script that works in Linux environments (or WSL2).

Prerequisites

• Linux or WSL2 environment
• Docker installed and running (required for evaluation harness)
• Python 3.10+

Execution

The setup_and_run_linux.sh script automates the entire process:

1. Compiles and installs the agent
2. Runs inference on 10 test tasks
3. Runs the official evaluation using Docker

# 1. Grant execution permissions
chmod +x setup_and_run_linux.sh

# 2. Run the script
./setup_and_run_linux.sh

Note: The complete evaluation may take time depending on your connection speed and CPU.

🐛 Known Issues

See CAMBIOS_REALIZADOS.md for resolved issues.

If you encounter an issue:

1. Check existing issues
2. Create a new issue with details

📝 License

This project is under the MIT License. See LICENSE for more details.

🙏 Acknowledgments

• AutoGen - Agent framework
• Rich - Terminal formatting
• Prompt Toolkit - Interactive CLI

📞 Contact & Community

• Discord Community: Join our Discord server - Get help, report bugs, suggest improvements, and collaborate with other users
• GitHub: https://github.com/davidmonterocrespo24/DaveAgent
• Issues: https://github.com/davidmonterocrespo24/DaveAgent/issues
• Email: [e-mail]

Join Our Discord Community

We encourage you to join our Discord server to:

• 💬 Get Help: Ask questions and get support from the community
• 🐛 Report Bugs: Help us improve by reporting issues you encounter
• 💡 Suggest Features: Share your ideas for new features and improvements
• 🤝 Collaborate: Connect with other users and contributors
• 📢 Stay Updated: Get the latest news and updates about DaveAgent

Click here to join: https://discord.gg/2dRTd4Cv

---

Made with ❤️ using AutoGen 0.4