acr-harmonized-data-model

LinkML Schema for ACR Harmonized Data. This model under development is intended to capture useful cross-study data in the AnVIL data repository.

Website

https://anvilproject.github.io/acr-harmonized-data-model

Repository Structure

• examples/ - example data
• project/ - project files (do not edit these)
• src/ - source files (edit these)
  • acr_harmonized_data_model
    • schema -- LinkML schema (edit this)
    • datamodel -- generated Python datamodel
• tests/ - Python tests

Developer Documentation

You can build the model using link-ml calls or using the makefile/justfile.

Linkml tools

Tools Required

• Python 3.12 (3.13 support coming for linkml)
• LinkML 1.8.6
• Pydantic 2.10.6
• mkdocs required to generate the documentation website

Tools to install:

pip install linkml
pip install mkdocs
pip install mkdocs-mermaid2-plugin

Checking for convention compliance

It can be helpful to review the schema for compliance with linkml conventions. To run a check for all files against the default standards the following command can be run.

linkml-lint ./src/acr_harmonized_data_model

Creating the anvil project

Here is the command to generate default artifacts if there is a need.

gen-project -d anvil ./src/acr_harmonized_data_model/schema/acr_harmonized_data_model.yaml

Creating the docs

gen-doc -d docs ./src/acr_harmonized_data_model/schema/acr_harmonized_data_model.yaml
mkdocs build

Manually building the Map Dragon formatted Data Dictionary

The DD export uses ftd-dd library to generate the CSV Files. This has been integrated into the command make site, but can also be run manually.

You can install this library using pip:

pip install -r requirements.txt 

or via the makefile

make install

Once installed, the dd-files can be built manually:

linkml_extract_dd src/acr_harmonized_data_model/schema/acr_harmonized_data_model.yaml

Otherwise, just build the site as usual and the data-dictionary will be created alongside the rest of the artifacts.

make site

Resulting DD files can be found in: project/data-dictionary

Makefile / Justfile

To run commands you may use good old make or the command runner just which is a better choice on Windows. Use the make command or duty commands to generate project artefacts:

• make help or just --list: list all pre-defined tasks
• make all or just all: make everything
• make deploy or just deploy: deploys site
• make test or just test: run tests on the model and examples

Credits

This project was made with linkml-project-cookiecutter.