sparc-curation

This repo contains sparcur, a python implementation of a validator for the SPARC Data Structure (SDS).

It also contains code, files, and documentation for curation and knowledge management workflows for SPARC datasets, protocols, and anatomical connectivity.

SDS Validator

To use sparcur to validate an SDS formatted dataset run

pip install sparcur
pushd path/to/my/dataset
python -m sparcur.simple.validate

The result is written to path/to/my/dataset/curation-export.json. General issues with the dataset can be found under the path_error_report property.

Background

For a general introduction to the SPARC curpation process see background.org.

For background on the SDS (with out-of-date technical details) see this paper.

Workflows

Documentation for curation workflows can be found in workflows.org.

Developer guide

See the developer guide for examples of how to reuse and develop sparcur.

Setup

New developers or curators should start by following setup.org.

Curation viewer

The curation viewer is a GUI application written in Racket that streamlines the processes of downloading, validating, and correcting SDS formatted datasets. The setup is currently quite involved because it needs to run directly on the OS where curators work. It supports windows, macos, and linux. Once the initial setup is complete there is an update mechanism which simplifies keeping the pipelines in sync.

SCKAN

This repo contains the core of the SCKAN release pipelines as well as the documentation for running and querying SCKAN.

Related links

• SODA GUI app for creating, validating, and uploading SDS formatted datasets.
• SDS Viewer a web UI for SDS formatted datatsets via the SDS validator.
• dockerfiles/source.org spec for developer docker image for this repo. Also has specs for the image that runs the sparcron single dataset pipelines, SCKAN images, and more.
• tgbugs/musl dockerhub repo with latest build of images.
• open-physiology-viewer code for converting ApiNATOMY models to OWL/RDF needed for apinatomy pipelines.