HEDTools - Python

HED (Hierarchical Event Descriptors) is a framework for systematically describing both laboratory and real-world events as well as other experimental metadata. HED tags are comma-separated path strings. HED, itself, is platform-independent, extendable, and data-neutral.

This repository contains the underlying python tools that support validation, summarization, and analysis of datasets using HED tags.

Most people will simply annotate their events by creating a spreadsheet or a BIDS JSON sidecar that associates HED tags with event codes or the events themselves. If you have such a spreadsheet or a JSON, you can use the HED Online Validator currently available at https://hedtools.org to validate or transform your files without downloading any tools.

A version of the online tools corresponding to the develop branch can be found at: https://hedtools.org/hed_dev.

Installation

Use pip to install hedtools from PyPI:

    pip install hedtools

To install directly from the GitHub repository master branch:

    pip install git+https://github.com/hed-standard/hed-python/@master

The HEDTools in this repository require Python 3.8 or greater.

Relationship to other repositories

The hed-python repository contains the Python infrastructure for validating and analyzing HED. This repository has several companion repositories:

hed-web contains the web interface for HED as well as a deployable docker module that supports web services for HED.
hed-examples contains examples of using HED in Python and MATLAB. This repository also houses the HED resources. See https://www.hed-resources.org for access to these.
hed-specification contains the HED specification documents. The hed-python validator is keyed to error codes in this document.
hed-schemas contains the official HED schemas. The tools access this repository to retrieve and cache schema versions during execution. Starting with hedtools 0.2.0 local copies of the most recent schema versions are stored within the code modules for easy access.

Develop versus master versus stable branches

The hed-python repository

Branch	Meaning	Synchronized with
stable	Officially released on PyPI as a tagged version.	`stable@hed-web` `stable@hed-specification` `stable@hed-examples`
latest	Most recent usable version.	`latest@hed-web` `latest@hed-specification` `latest@hed-examples`
develop	Experimental and evolving.	`develop@hed-web` `develop@hed-specification` `develop@hed-examples`

As features are integrated, they first appear in the develop branches of the repositories. The develop branches of the repositories will be kept in sync as much as possible If an interface change in hed-python triggers a change in hed-web or hed-examples, every effort will be made to get the three types of branches (develop, latest, stable) of the respective repositories in sync.

API documentation is generated on ReadTheDocs when a new version is pushed on any of the three branches. For example, the API documentation for the latest branch can be found on hed-python.readthedocs.io/en/latest/.

To contribute

Contributions are welcome. Please use the Github issues for suggestions or bug reports. The Github pull request may also be used for contributions. These PRs should be made to the develop branch, not the master branch.

Local Settings Storage

Cached Schemas by default are stored in "home/.hedtools/" Location of "home" directory varies by OS.

Use hed.schema.set_cache_directory to change the location. The HED cache can be shared across processes.

Starting with hedtools 0.2.0 local copies of the most recent schema versions are stored within the code modules for easy access.