sbxreader

Python module to read Neurolabware Scanbox files.

Usage

Open a scanbox file

from sbxreader import sbx_memmap
dat = sbx_memmap(`filename.sbx`)
# This memory maps the file, the shape is:
# NFRAMES x NPLANES x NCHANNELS x HEIGHT x WIDTH
# To access frames treat it like a numpy array
# BUT create a new array when getting the frame:
# that is needed because the file is memory mapped.

stack = np.array(dat[:20])

# access metadata as a dictionary
print(dat.metadata)

# number of invalid columns when recording in bidirectional mode
print(dat.ndeadcols)

Note: that np.array(dat[:]) is important to have the data properly formatted; dat[:] is the inverted dataset. This is because of how the dataset is saved.

Important you can also use dat.get_stack(nframes, offset_frame) to load data. This may be faster but on windows the offset_frame needs to be incremented sequentially otherwise there is a limit to the number of frames that can be read.

Command line tool

The reader includes a file viewer to explore raw data. From the command line (not the python terminal) do:

sbxviewer <filename>

This requires additional dependencies: opencv-python PyQt5 and pyqtgraph The dependencies are installed using pip if not present.

Get metadata from the file

from sbxreader import sbx_get_metadata
metadata = sbx_get_metadata(`filename.sbx`)

print(metadata) # dictionary with the recording metadata

Parse the mat file

from sbxreader import sbx_get_info
info = sbx_get_info(`filename.sbx`)

# Returns a scipy.io matlab structure 

Instalation

pip install sbxreader

Source code is in https://github.com/jcouto/sbxreader.git

To install from source:

1. clone the repository
2. run the command python setup.py develop