Taskify

AI chatbot with functionalities such as tool usage, conversational memory, streaming, RAG etc. built using LangChain.

Project Structure

The project has a simplified structure with all core functionality in a single directory:

.
├── taskify/              # Main package directory
│   ├── __init__.py       # Package initialization and exports
│   ├── config.py         # Configuration loading and logging setup
│   ├── executor.py       # AgentExecutor for running agents
│   ├── memory.py         # Various memory implementations
│   ├── streaming.py      # Streaming functionality for async responses
│   └── tools.py          # Tool implementations for agent use
├── config.yaml           # Configuration settings
├── example.py            # Example script showing basic usage
├── requirements.txt      # Project dependencies
└── logs/                 # Directory for log files

Setup

1. Clone the repository
2. Install dependencies:

pip install -r requirements.txt

3. Set up environment variables by creating a .env file in the project root:

# Edit the .env file with your API keys
nano .env

4. Add your API keys to the .env file:

# Google Gemini API Key for LLM
GOOGLE_API_KEY=your_google_api_key_here

# SerpAPI Key for web search
SERPAPI_KEY=your_serpapi_key_here

Configuration

The project uses config.yaml for configuration settings. You can customize:

• LLM settings (model, temperature, max tokens)
• Memory type and parameters
• Logging configuration
• Agent settings (max iterations, async mode)
• Tool-specific settings
• Streaming settings

See the config.yaml file for all available options.

Running Examples

The project includes an example script to demonstrate its capabilities:

python example.py

The example script will:

1. Load the configuration from config.yaml
2. Initialize the language model and tools
3. Create an agent with the specified memory type
4. Run a series of test queries to demonstrate different features
5. Show the memory state at the end

Features

• Multiple memory implementations (Buffer, Window, Summary)
• Streaming responses for real-time output
• Various tools (calculator, datetime, web search, etc.)
• Configurable via YAML
• Comprehensive logging
• Async operation mode for improved performance