Skip to content

installation.md

Latest commit

 

History

History
148 lines (103 loc) · 2.64 KB

installation.md

File metadata and controls

148 lines (103 loc) · 2.64 KB
Raw

Installation Guide

This guide will help you set up GenomIO on your system.

Prerequisites

  • Python 3.8 or higher
  • Git
  • CUDA-compatible GPU (optional, for faster inference)

System Requirements

  • RAM: 8GB minimum, 16GB recommended
  • Storage: 10GB free space for models and data
  • GPU: 4GB VRAM for local model inference (optional)

Installation Methods

Method 1: Standard Installation

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

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

# Install dependencies
pip install -r requirements.txt

# Install the package
pip install -e .

Method 2: Development Installation

# Clone with development dependencies
git clone https://github.com/yourusername/GenomIO.git
cd GenomIO

# Create virtual environment
python -m venv genomio-dev
source genomio-dev/bin/activate

# Install with development extras
pip install -e ".[dev,notebooks]"

# Install pre-commit hooks
pre-commit install

External Dependencies

NCBI Entrez Direct (for data downloading)

# Linux/macOS
sh -c "$(curl -fsSL ftp://ftp.ncbi.nlm.nih.gov/entrez/entrezdirect/install-edirect.sh)"

# Or using conda
conda install -c bioconda entrez-direct

GPU Support (Optional)

For CUDA support:

# Install PyTorch with CUDA
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

Verification

Test your installation:

import genomio
from src.models.dnabert2 import DNABert2Model

# Test model loading
model = DNABert2Model()
print("Installation successful!")

Configuration

  1. Copy the example configuration:
cp config/config.yaml config/my_config.yaml
  1. Edit config/my_config.yaml with your settings:
data:
  results_dir: "/path/to/your/results"
  
environment:
  device: "cuda"  # or "cpu"
  cache_dir: "/path/to/cache"

Common Issues

Import Errors

# If you get import errors, make sure you're in the right directory
export PYTHONPATH="${PYTHONPATH}:$(pwd)"

CUDA Out of Memory

# Reduce batch size or use CPU
config["environment"]["device"] = "cpu"

Missing Dependencies

# Install specific missing packages
pip install transformers torch biopython langchain

Docker Installation (Alternative)

FROM python:3.9

WORKDIR /app
COPY . /app

RUN pip install -r requirements.txt
RUN pip install -e .

CMD ["python", "-c", "import genomio; print('GenomIO ready!')"]

Build and run:

docker build -t genomio .
docker run genomio