medpython 1.0.4

Medial EarlySign Python

Medial EarlySign Python Library

Our platform is designed to transform complex, semi-structured Electronic Medical Records (EMR) into machine-learning-ready data and reproducible model pipelines. The framework is optimized for the unique challenges of sparse, time-series EMR data, delivering low memory usage and high-speed processing at scale.

It was conceived as a TensorFlow for machine learning on medical data.

All software is now open-sourced under the MIT license. Some of the models developed by Medial EarlySign that are currently in production are available exclusively through our partners.

The framework was battle-tested in production across multiple healthcare sites and was a key component of an award-winning submission to the CMS AI Health Outcomes Challenge.

Why Use This Platform?

• High-Performance Processing: Engineered for large-scale, sparse EMR time-series data where general-purpose libraries like pandas fall short.
• Reusable Pipelines: Save valuable engineering time by providing shareable, tested pipelines and methods.
• Built-in Safeguards: Mitigate common pitfalls like data leakage and time-series-specific overfitting.
• Production-Ready: Designed for easy deployment using Docker or minimal distroless Linux images.
• State of The Art Algorithms Developed state of the art algorithms for EMR use cases, explainablity and fairness and more...

Core Components

The platform is built on three key pillars:

• MedRepository: A compact, efficient data repository and API for storing and accessing EMR signals. Querying categorical signals like perscriptions and diagnosis in an easy and efficient API.
• MedModel: An end-to-end machine learning pipeline that takes data from MedRepository or JSON EMR inputs to produce predictions and explainability outputs. It supports both training and inference.
• Medial Tools: A suite of utilities for training, evaluation, and workflow management, including bootstrap analysis, fairness checks, and explainability.

Setup

You can quickly install the package using pip:

pip install medpython

System Requirements

• Supported Systems: This pre-built version is available for modern Linux distributions (specifically manylinux2014 equivalents, such as CentOS >= 7 or Ubuntu >= 13.04). The software also compiles in Windows but you will need to install Boost yourself. I hope shortl to provide windows pre-compiled builds through conda.
• Python: Requires Python 3.10 through 3.14

Compilation for Other Systems If you're using an older Linux or a different platform/Python version >= 3.8, you'll need to compile the package yourself.

• Note on Compilation: Ensure the Boost libraries are installed. For a local setup, set the environment variable BOOST_DISABLE_STATIC=1 to link against shared Boost libraries (The reason is that your system static libraries weren't compiled with -fPIC flag, so you can't use them inside python module):

export BOOST_DISABLE_STATIC=1

You can also set Boost installation directory with BOOST_ROOT environment variable if it is not part of the system libraries. Currently the library was compiled/tested in linux and windows only.

Usage

import med
from AlgoMarker import AlgoMarker
from ETL_Infra import prepare_final_signals, prepare_dicts, finish_prepare_load, create_train_signal

More information on usage:

• ETL_Infra - To load data into medpython data format
  • API Documentation
• AlgoMarker
  • From usage in deployment/lighter API interface. Usefull for accessing prebuild models. API Documentation
• med library
  • Full library from training new models or using with existing models. Complete Example: From Data to Model
  • API Documentation

Getting Started

• Build a new model: Follow the step-by-step Tutorials to build a model from scratch.
• Use an existing model: Browse the collection of Models.