Installation

Get started with MRRA by installing it in your Python environment.

Prerequisites

MRRA requires Python version 3.10 or higher to run properly.

Python >= 3.10
pip package manager

Basic Installation

# Basic installation
pip install mrra

# Using conda (when available)
conda install -c conda-forge mrra

# Using Poetry
poetry add mrra

Installation Variants

MRRA provides several installation options depending on your needs:

Development Installation

# Include development tools
pip install mrra[dev]

GPU Support

# Include GPU acceleration support
pip install mrra[gpu]

Complete Installation

# Include all optional dependencies
pip install mrra[all]

Verify Installation

After installation, verify that MRRA is properly installed:

# Check if installed correctly
import mrra
print(f"MRRA version: {mrra.__version__}")

# Test basic functionality
from mrra import AgentBuilder
agent_builder = AgentBuilder()
print("✅ MRRA installed successfully!")

Development Installation

If you want to contribute to MRRA or run the latest development version:

# Clone the repository
git clone https://github.com/yourusername/mrra.git
cd mrra

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

# Run tests to verify
pytest

Post-Installation Setup

After installation, you'll need to configure your LLM provider and other services:

1. Configure LLM Provider

Set up your preferred LLM provider:

from mrra import AgentBuilder

# OpenAI Configuration
agent = AgentBuilder().with_llm(
    provider="openai",
    model="gpt-4o-mini", 
    api_key="your-openai-api-key"
)

# Qwen Configuration  
agent = AgentBuilder().with_llm(
    provider="qwen",
    model="qwen-plus",
    base_url="https://dashscope.aliyuncs.com/compatible-mode/v1",
    api_key="your-qwen-api-key"
)

2. Test Your Setup

Run a quick test to ensure everything is working:

from mrra import AgentBuilder
from mrra.core.types import TrajectoryBatch
import pandas as pd

# Create sample data
df = pd.DataFrame({
    'user_id': ['user_1', 'user_1'], 
    'timestamp': ['2023-01-01 09:00:00', '2023-01-01 12:00:00'],
    'latitude': [31.2304, 31.2404],
    'longitude': [121.4737, 121.4837],
    'poi_id': ['poi_1', 'poi_2']
})

# Test basic prediction
agent = (AgentBuilder()
         .with_llm(provider="openai", model="gpt-4o-mini", api_key="your-key")
         .build())

trajectory_batch = TrajectoryBatch(df)
result = agent.predict_next(user_id="user_1", history=trajectory_batch)
print(f"✅ Prediction successful: {result.predicted_location}")

Environment Variables

You can also configure MRRA using environment variables:

# Set environment variables
export MRRA_OPENAI_API_KEY="your-openai-key"
export MRRA_QWEN_API_KEY="your-qwen-key"  
export MRRA_DEFAULT_PROVIDER="openai"
export MRRA_DEFAULT_MODEL="gpt-4o-mini"

Troubleshooting Installation Issues

Dependency Conflicts

If you encounter dependency conflicts:

# Create a clean virtual environment
python -m venv mrra-env
source mrra-env/bin/activate  # On Windows: mrra-env\Scripts\activate

# Install MRRA
pip install mrra

GPU Support Issues

For GPU acceleration issues:

# Install with specific CUDA version
pip install mrra[gpu] --extra-index-url https://download.pytorch.org/whl/cu118

# Verify GPU support
python -c "import torch; print(f'CUDA available: {torch.cuda.is_available()}')"

Import Errors

If you encounter import errors:

# Check installation
pip show mrra

# Reinstall if necessary
pip uninstall mrra
pip install mrra --no-cache-dir

Updating MRRA

Keep MRRA up to date with the latest features:

# Update to latest version
pip install --upgrade mrra

# Update with all dependencies  
pip install --upgrade mrra[all]

Uninstallation

To remove MRRA from your system:

# Uninstall MRRA
pip uninstall mrra

# Clean up dependencies (if needed)
pip autoremove

Configuration files and cached data will remain after uninstallation. You can manually clean them up if needed.

Next Steps

Now that MRRA is installed, proceed to:

Configuration - Set up your LLM providers and services
Examples - See real-world usage examples
Development - Contributing and development guidelines

Installation

On this page