# Installation

This guide provides different ways to install `spMetaTME`, whether you are using it as a library in your Python scripts or as a standalone CLI tool.

## Prerequisites

Before installing `spMetaTME`, ensure your system meets these requirements:

### Python
*   **Version**: Python 3.9 or higher (3.10+ recommended).
*   **Package Manager**: `pip` or `uv` (recommended for faster dependency resolution).

### Additional Tools (Recommended)
*   **R Statistics**: Required for spatial domain clustering (specifically used by the `BayesSpace` package).
*   **NVIDIA GPU**: Highly recommended for faster model inference and training (requires `CUDA` and compatible drivers).

---

## Standard Installation

### Using `uv` (Recommended)
If you are using the modern `uv` package manager:

```bash
uv add spmetatme
```

### Using `pip`
To install `spMetaTME` directly from PyPI:

```bash
pip install spmetatme
```

---

## Installing from Source

If you want to contribute to the package or use the latest developmental version, you can install it from GitHub.

1.  **Clone the repository**:
    ```bash
    git clone https://github.com/SurajRepo/spmetatme.git
    cd spmetatme
    ```

2.  **Install in editable mode**:
    ```bash
    pip install -e .
    ```

---

## Verifying Installation

Once installed, you can verify that the CLI is working by checking the version:

```bash
spmetatme --version
```

Or by listing the available metabolic models for inference:

```bash
spmetatme infer --list-models
```

---

## Troubleshooting

### R Dependency Issues
If spatial clustering fails, ensure R is correctly installed and accessible in your system's `PATH`. You can verify this by typing `R --version` in your terminal.

### GPU Support
`spMetaTME` uses PyTorch. If you have an NVIDIA GPU but it's not being utilized, check your PyTorch installation:

```python
import torch
print(torch.cuda.is_available())
```
If this returns `False`, you may need to install the version of `torch` that matches your CUDA version. See [pytorch.org](https://pytorch.org/) for instructions.