xnat-dashboards/xnat-dashboards

Dashboards for XNAT

Dashboards for XNAT

Dashboards for XNAT is a framework for extensive exploration, review and monitoring of large neuroimaging datasets hosted on an XNAT instance. The project was initiated in 2020 during the Google Summer of Code program by Mohammad Asif Hashmi.

The project is now maintained by Greg Operto, Jordi Huguet and Marina Garcia Prat of the BarcelonaBeta Brain Research Center (BBRC).

The project was designed to be compatible with any generic XNAT instance. However, to date the development version includes some features that are specific to BBRC. While future work will alleviate that specificity, it is worth underlining that the project provides a framework to build tailored visualizations rather than ready-made dashboards to use off the shelf.

Built with

• Flask
• pyxnat
• pandas
• plotly
• bootstrap

Prerequisite

A running XNAT instance and read access rights to it are required prior to using Dashboards for XNAT.

Installation

Installation from source is advised at this current stage of development of the project.

However, version v0.3 of Dashboards for XNAT can be installed using pip. That version was released at the end of the Google Summer of Code 2020.

pip install xnat-dashboards

Tests

Running the test suite requires the following:

• python-nose v1.2.1+
• coveralls

Run the tests with the following command:

nosetests xnat-dashboards/tests

Documentation

While the project has gone under a significant refactoring process over the last months and require new efforts on documentation, the original materials are still available along with a user guide and a developer guide.

xnat-dashboards's Issues

[`ScanTypeGraph`] Some scan types are missing

Some random scan types were missing from ScanTypeGraph, both at global and project-specific level. Those scan types missing are not excluded from the whitelist.json.

Sorting dicts by using aggregate values (ie. integers) as keys is not a good idea, since they can very well be non-unique and thus unsuited as keys. We lose graph items having the very same aggregate values (ie. scans of a given type). See below,

XNAT-Dashboards/dashboards/data/graph.py

Line 27 in e9e03ff

sd = OrderedDict(sorted({v: k for k, v in g['count'].items()}.items()))