Data (engineering) tooling for the OMOP CDM

Why this?

Current definitions for the OMOP Common Data Model, and utilities to generate definitions for a few SQL/DDL dialects are written in R.

This package is currently an experiment to facilitate the use of Python, a popular language in data engineering circles, and its rich ecosystem of packages for it, to implement data pipelines aiming at using OMOP.

Initially, this package extract the OMOP CDM definitions from CSV files distributed with the R package CommonDataModel into an intermediate abstraction that can be used to translate OMOP into SQLAlchemy, an Object-Relational Mapping (ORM) for Python. Example of this are currently here: doc/render

Installation

From this source tree:

pip install `.[sqlalchemy]'

CLI

The CDM definitions in the R package CommonDataModel can be converted to Python modules that use SQLAlchemy from command line. That

topomop-convert \
  -s path/to/sourcetree/CommonDataModel/inst/csv/ \
  -d path/to/python/conversion.py \
  v5.4

topomop-convert \
  -s $(Rscript -e "cat(system.file(\"csv\", package = \"CommonDataModel\"))")
  path/to/sourcetree/CommonDataModel/inst/csv/ \
  -d path/to/python/conversion.py \
  v5.4