codex-weather-agent 1.0.0

Production-ready AI-powered conversational weather assistant built with LangGraph

🌤️ Codex Weather Agent

An AI-powered conversational weather assistant built with LangGraph, supporting multiple LLM providers (OpenAI, Azure OpenAI, Gemini, Claude) and providing real-time weather data through a production-ready Python API.

✨ Features

• 🤖 Conversational AI: Natural language interactions powered by LangGraph workflows
• 🌍 Multiple LLM Support: OpenAI, Azure OpenAI, Google Gemini, and Anthropic Claude
• 🌦️ Comprehensive Weather Data: Current conditions, forecasts, and air quality
• 📍 Location Intelligence: Automatic location detection and coordinate lookup
• ⚙️ Flexible Configuration: Environment variables and programmatic setup
• 🔧 Easy Integration: Production-ready Python library for applications
• 📊 Proper Logging: Structured logging with configurable levels
• 🛡️ Error Handling: Robust error handling and validation

🚀 Installation

From PyPI

pip install codex-weather-agent

From Source

git clone https://github.com/CodexJitin/Weather-Agent.git
cd Weather-Agent
pip install -e .

⚙️ Configuration

1. Get API Keys

You'll need:

• OpenWeatherMap API Key: Get it here (Required)
• LLM Provider API Key: Choose from OpenAI, Azure OpenAI, Google Gemini, or Anthropic Claude

2. Environment Setup

Create a .env file (copy from .env.example):

# Weather API (Required)  
OPENWEATHER_API_KEY=your_openweather_api_key_here

# Choose your LLM provider
LLM_PROVIDER=openai

# OpenAI (default)
OPENAI_API_KEY=your_openai_api_key_here
OPENAI_MODEL=gpt-3.5-turbo

# Or Azure OpenAI
AZURE_OPENAI_API_KEY=your_azure_key_here
AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com/
AZURE_OPENAI_DEPLOYMENT_NAME=your_deployment_name
AZURE_OPENAI_MODEL=gpt-35-turbo

# Or Google Gemini
GOOGLE_API_KEY=your_google_api_key_here
GEMINI_MODEL=gemini-pro

# Or Anthropic Claude
ANTHROPIC_API_KEY=your_anthropic_api_key_here
CLAUDE_MODEL=claude-3-sonnet-20240229

🎯 Usage

Basic Library Usage

from codex_weather_agent import WeatherAgent
from codex_weather_agent.logging_config import setup_logging

# Set up logging (optional)
setup_logging(level="INFO")

# Method 1: Use environment variables (recommended)
agent = WeatherAgent(provider_name="openai")
response = agent.query("What's the weather in Tokyo?")
print(response)

# Method 2: Pass API keys directly
agent = WeatherAgent(
    provider_name="openai",
    provider_config={
        "api_key": "your-openai-api-key",
        "model": "gpt-4"
    },
    openweather_api_key="your-openweather-api-key"
)

# Method 3: Use configuration object
from codex_weather_agent import WeatherAgentConfig

config = WeatherAgentConfig(
    llm_provider="openai",
    openweather_api_key="your-openweather-api-key"
)
config.update_llm_config(api_key="your-openai-api-key", model="gpt-4")
agent = WeatherAgent.from_config(config)
response = agent.query("Air quality in Delhi?")

Advanced Configuration

from weather_agent_langgraph import WeatherAgent

# Configure different LLM providers
configs = {
    "openai": {
        "api_key": "sk-...",
        "model": "gpt-4",
        "temperature": 0
    },
    "azure": {
        "api_key": "your-key",
        "azure_endpoint": "https://your-resource.openai.azure.com/",
        "deployment_name": "gpt-35-turbo",
        "model": "gpt-35-turbo"
    },
    "gemini": {
        "api_key": "your-google-key",
        "model": "gemini-pro"
    },
    "claude": {
        "api_key": "your-anthropic-key", 
        "model": "claude-3-sonnet-20240229"
    }
}

# Create agent with specific provider
agent = WeatherAgent(
    provider_name="gemini",
    provider_config=configs["gemini"]
)

🌟 Example Queries

The agent understands natural language queries:

• "What's the weather in New York?"
• "Give me the air quality for Delhi"
• "5-day forecast for London"
• "Is it raining in my location?"
• "Compare weather between Tokyo and Paris"
• "What should I wear in Mumbai today?"

🏗️ Architecture

The package uses a LangGraph workflow with the following components:

• Agent: Main orchestrator class
• Workflow: LangGraph state machine for query processing
• Tools: Weather data collection (OpenWeatherMap API)
• LLM Providers: Abstraction layer for different AI models
• State Management: Conversation and tool execution tracking

┌─────────────────┐    ┌──────────────────┐    ┌─────────────────┐
│   User Query    │───▶│  LangGraph       │───▶│   AI Response   │
│                 │    │  Workflow        │    │                 │
└─────────────────┘    └──────────────────┘    └─────────────────┘
                              │
                              ▼
                    ┌──────────────────┐
                    │  Weather Tools   │
                    │  • Current       │
                    │  • Forecast      │
                    │  • Air Quality   │
                    │  • Location      │
                    └──────────────────┘

🛠️ Development

# Clone the repository
git clone https://github.com/CodexJitin/Codex-Weather-Agent.git
cd Codex-Weather-Agent

# Install in development mode
pip install -e ".[dev]"

# Run tests
pytest

# Format code
black src/
isort src/

📋 Requirements

• Python 3.8+
• OpenWeatherMap API key
• One of the supported LLM provider API keys:
  • OpenAI API key
  • Azure OpenAI credentials
  • Google AI Studio API key (for Gemini)
  • Anthropic API key (for Claude)

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request. For major changes, please open an issue first to discuss what you would like to change.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• LangChain and LangGraph for the AI framework
• OpenWeatherMap for weather data
• Rich for beautiful CLI interfaces

📞 Support

• 📫 Issues: GitHub Issues
• 💬 Discussions: GitHub Discussions

---

🌤️ Weather Agent LangGraph
AI-powered weather intelligence at your fingertips
_{Built with ❤️ by CodexJitin}