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.ucsd.edu/hed 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.ucsd.edu/hed_dev.
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 3.7 or greater. Note: HED is continuing to support Python 3.7 until 2023, because it is needed for MATLAB versions R2019a through R2020a.
Note: The final exported interface for Python tools has not been completely frozen in 0.1.0 and is expected to undergo minor changes in interface until the release of version 1.0.0.
The hed-python repository contains the Python infrastructure for validating
and analyzing HED. This repository has several companion repositories:
hed-webcontains the web interface for HED as well as a deployable docker module that supports web services for HED.hed-examplescontains 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-specificationcontains the HED specification documents. Thehed-pythonvalidator is keyed to error codes in this document.hed-schemascontains the official HED schemas. The tools access this repository to retrieve and cache schema versions during execution. Starting withhedtools 0.2.0local copies of the most recent schema versions are stored within the code modules for easy access.
The hed-python repository
| Branch | Meaning | Synchronized with |
|---|---|---|
| stable | Officially released on PyPI as a tagged version. | stable@hed-webstable@hed-specificationstable@hed-examples |
| latest | Most recent usable version. | latest@hed-weblatest@hed-specificationlatest@hed-examples |
| develop | Experimental and evolving. | develop@hed-webdevelop@hed-specificationdevelop@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/.
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.
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.
Code climate reports: https://codeclimate.com/github/hed-standard/hed-python.