Installation¶

OneEHR requires Python 3.12+.

Quick Install (PyPI)¶

pip install oneehr

From Source¶

git clone https://github.com/MedXLab/OneEHR.git
cd OneEHR
uv venv .venv --python 3.12
uv pip install -e .

Verify The CLI¶

oneehr --help

You should see these top-level commands:

• preprocess — bin features, split patients
• train — train ML/DL models
• test — evaluate on test set
• analyze — cross-system comparison
• plot — publication-quality figures
• convert — convert raw datasets to OneEHR format

Optional Extras¶

# Development and testing
pip install oneehr[test]

# Documentation
pip install oneehr[docs]

# Survival analysis (lifelines for fast C-index)
pip install oneehr[survival]

GPU Support¶

OneEHR uses PyTorch for deep learning models. If you have a CUDA-capable GPU, install a CUDA-enabled PyTorch build. The trainer defaults to device = "auto" and will use CUDA when available.

[trainer]
device = "auto"       # auto | cpu | cuda
precision = "fp16"    # fp32 | fp16 | bf16

LIME Support (Optional)¶

For LIME-based interpretability, install the lime package:

pip install lime