pycaravel: This module is designed to simplify continuous integration (CI) of data from multiple projects.

Note

Go to the end to download the full example code.

BIDS directory parsing (user mode)¶

Credit: A Grigis

pycaravel is a Python package that enables you to parse various source of data. In this tutorial you will learn how to parse and search in a BIDS directory.

First checks¶

In order to test if pycaravel package is installed on your machine, you can check the package version.

import os
from pprint import pprint

import caravel

print(caravel.__version__)
print(caravel.info())

Create a parser for your project¶

The package provides a common interface to parse a BIDS directory. The parsing rules are defined by projects in the module, so we will need to specify the project name you are working on. For the moment it is not possible to specify these rules via the API.

cwdir = os.path.dirname(os.path.realpath(__file__))
parser = caravel.get_parser(
    project="hbn", confdir=os.path.join(cwdir, os.pardir, "conf"),
    layoutdir=cwdir)

You can now list the available configurations for your project, and the available layout representations pre-generated. Note that these representations are sorted by dates, and that the latest one will be used.

pprint(parser.conf)
pprint(parser.representation)

You can export the whole ‘sourcedata’ layout in a pandas DataFrame.

print(parser.export_layout("sourcedata"))

It is also possible to filter this dataset. You need first to list all the available filtering keys, then list all the available values for the filtering key(s) of interest, and finally filter your dataset.

print(parser.list_keys("sourcedata"))
print(parser.list_values("sourcedata", "subject"))
print(parser.list_values("sourcedata", "modality"))
search1 = parser.filter_layout("sourcedata", subject="NDARLU606ZDD",
                               modality="anat|func")
print(search1)
search2 = parser.filter_layout("sourcedata", extension="tsv")
print(search2)

Finally you may want to ask the system to load the filtered data. Only a couple of file extensions are supported. If no loader has been found the filename is returned

if not search1.empty:
    data1 = parser.load_data("sourcedata", search1)
    pprint(data1)
if not search2.empty:
    data2 = parser.load_data("sourcedata", search2)
    pprint(data2)

And for the derivatives¶

We can do the same for the derivatives

print(parser.export_layout("derivatives"))
print(parser.list_keys("derivatives"))
print(parser.list_values("derivatives", "process"))
search3 = parser.filter_layout("derivatives", subject="NDARAE199TDD",
                               process="scalars")
print(search3)
data3 = parser.load_data("derivatives", search3)
pprint(data3)
search3bis = parser.filter_layout("derivatives", process="tbss_stats",
                                  extension="tsv")
print(search3bis)
data3bis = parser.load_data("derivatives", search3bis)
pprint(data3bis)

And for the clinical data¶

We can do the same for the clinical data denoted in BIDS phenotype

print(parser.export_layout("phenotype"))
search4 = parser.filter_layout("phenotype", extension="tsv")
print(search4)
data4 = parser.load_data("phenotype", search4)
pprint(data4)

Estimated memory usage: 0 MB

Gallery generated by Sphinx-Gallery