Welcome to Neurogister, a DVC driven data registry for neuroimaging. It provides datasets for process validation and benchmarking, educational purposes, and more.
Neurogister is delivered through Pypi. To install it, simply run:
pip install neurogister
You can also clone the repository and install it from source. For that, you'll need hatch installed. Then, run the following commands :
git clone https://github.com/AlexVCaron/neurogister.git
hatch env create
hatch shell
For now, neurogister doesn't offer much. It is mostly designed to serve as a test dataset creator and fetcher for scilpy and nf-scil.
You are free to roam it's current API. You can use the info, pull and list commands to get data from the registry. Try it and tell us about it !
Once cloned, you can use dvc directly to interact with the registry. For example, to get the list of all datasets, run:
dvc list -R
To get a list of commands, run dvc --help.
The data is stored in a DVC registry, stored online and versioned through this repository. The dataset registry is stored in the data/ folder and the package store in the store/ folder.
Registry
Inside the registry, you'll find several collection of datasets, in a more or less organized fashion,though each collection is stored in a separate folder. Files collected in a same folder should have a common role or exist in the same context or space.
Store
Comparatively to the registry, the store is a strictly organized collection of packages. Each entry is a curated vendor, registered with neurogister, that contains a set of validated and versioned packages with purpose. Every package in the store points to a collection of datasets in the registry.
- Data structure schema
- Reorganize store in logical folders (departments)
- Create PR workflow for adding new data (in data/)
- Create PR workflow for adding new packages (in store/...)