nsrr

nsrr - a Python-based client library is available for users to access NSSR Cloud resources. This library is compatible with Mac, Linux and Windows (tested on win10 PowerShell with admin privileges).

Installation

pip install nsrr

If both version of Python i.e., Python2.x and Python3.x are installed in the OS then you can use the command below to call Python3 based pip using:

pip3 install nsrr or python3 -m pip install nsrr

Usage

This section of the documentation covers usage of 'CFS' dataset as an example.

To learn about different parameters, use help argument:

nsrr --help

To list approved datasets access of a user:

nsrr --list-access

To list all the files of the dataset:

nsrr cfs --list-files

To list all the directories of the dataset:

nsrr cfs --list-directories

To download based on a folder or file path:

nsrr -d cfs/forms
nsrr -d cfs/dataset/cfs-data-dictionary-0.5.0-variables.csv
nsrr -d cfs/polysomnography/annotations-events-nsrr

To download entire dataset:

nsrr -d cfs

To decompress EDFZ files into EDF files, and delete the compressed originals:

nsrr -d cfs --decompress

To learn more about EDFZ file format, visit EDFZ: working with compressed EDFs{:target="_blank"}.

To list all the subjects of a specific dataset:

nsrr cfs --list-subjects

To download subject specific files from a dataset:

nsrr -d cfs --subject 800002

To provide password in non-interactive way:

nsrr -d cfs --token-file token.txt

Data Integrity check is performed via the following two options.

• (Recommended) md5 checksum value is unique to every file. This option verifies that the downloaded file is same as being served by NSRR using md5 checksum value comparison.
• file size check to match with download size of the file hosted by NSRR.

To skip memory intensive data-integrity check:

nsrr cfs -d --no-md5

To forcefully download the whole dataset:

nsrr -d cfs --force

To list the version of the nsrr library:

nsrr -v

Developer guide

Prerequisites

Following installation are necessary to start development,

• Python (version >=3.6)
• Auth server is running

Initialization

Update Auth server address in the 'nsrr.py' file

Build and publish package

Delete any existing distributions in the dist folder:

rm -rf dist/*

Update setup.py, nsrr/main.py and nsrr/init.py to bump version number:

ex: vi nsrr/__init__.py
__version__ = "x.x.x"

Run build command:

python3 setup.py sdist bdist_wheel

Update test pypi with the latest version:

twine upload --repository-url https://test.pypi.org/legacy/ dist/*

Upload pypi with the latest version:

twine upload -u <username> -p <password> dist/*

Notes:

1. It is recommended to use Python version 3.8.x
2. Compatible with Windows (tested on win10 powershell with admin privileges), Mac and Linux systems
3. Data Integrity check is performed via the following two options
   • (Recommended) md5 checksum value is unique to every file. This option verifies that the downloaded file is exactly the same as being served by NSRR using md5 checksum value comparison. Use '--no-md5' to skip this option
   • file size check to match with download size of the file hosted by NSRR