OneEHR
EHR AI platform MedX-PKU/OneEHR

Installation

OneEHR requires Python 3.12 or newer.

PyPI

pip install oneehr
oneehr --help

Source Checkout

git clone https://github.com/MedX-PKU/OneEHR.git
cd OneEHR
uv venv .venv --python 3.12
source .venv/bin/activate
uv pip install -e ".[test]"
uv run oneehr --help

Use oneehr ... when the package is installed in the active environment. Use uv run oneehr ... when you are working from a checkout and want uv to run the command inside the project environment.

CLI Commands

oneehr --help

Top-level commands:

Command Purpose
preprocess Bin event data, build features, create labels, and save the split contract
train Train configured ML/DL models
test Evaluate trained models and configured systems on the test split
analyze Write comparison, feature importance, fairness, calibration, statistical test, and missing-data outputs
plot Render figures from run artifacts
convert Convert supported raw datasets into OneEHR CSV tables

Optional Packages

pip install "oneehr[survival]"  # lifelines support for survival analysis
pip install lime                # optional LIME interpretability support

GPU Setup

OneEHR uses PyTorch for deep learning models. Install a CUDA-enabled PyTorch build if you want GPU training, then keep the trainer device on auto or set it explicitly:

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

Documentation Preview

From a source checkout:

uv run --group docs mkdocs serve

MkDocs prints the local preview URL, usually http://127.0.0.1:5000/.