# What is niio?
niio is a Python package for quickly loading common neuroimaging data. It's a class of methods that act as wrappers for common input formats like those offered by Nifti, Matlab, Nibabel, h5py, and pickle libraries. Current functionality includes loading, saving, and converting between various formats.
While the above libraries work fine, I wanted a way to access the data without needing to remember / write out specifically what attributes and methods would specifically return the data. As an example, with nibabel, for Gifti formattd data:
```python import nibabel as nb func_file = '../data/func.gii' func = nb.load(func_file)
# get access to the data array func_data = nb.darrays[0].data
# and if func_file has multiple data arrays... for i in np.arange(num_darrays):
func_data = nb.load(darrays[i]).data
I wanted a convenient way just to get all the data at once.
# How to use?
`bash
git clone https://github.com/kristianeschenburg/niio.git
cd .
pip install .
`
# Example usage:
We begin by loading in the necessary modules. As named, loaded is used for loading data, write is used for saving data, and convert is used for converting between file formats.
`python
from niio import loaded, write, convert
`
### Matlab files Let's start by loading in a conventional Matlab .mat file.
`python
ex1 = loaded.load('../data/matrix.mat')
print(ex1.shape)
`
We see that ex1 is a (5,10) matrix. However, what we can't see immediately, is that `matrix.mat` contains two arrays, named mat1 and mat2. If we know what our specific arrays are called, we can load each one by name:
`python
ex1 = loaded.load('../data/matrix.mat', dataset='mat1')
ex2 = loaded.load('../data/matrix.mat', dataset='mat2')
print(ex1.shape)
print(ex2.shape)
`
Generally, we'll only save one array per .mat file, so this is more for convenience.
### Gifti Files
`python
func = loaded.load('../data/func.gii')
print(func), print(func.shape)
`
And if you know which data array you might be interested in:
`python
func = loaded.load('../data/func.gii',dataset=1)
print(func), print(func.shape)
`