Right before the final evaluation (week 12)

We are approaching the soft deadline for GSoC2014. It’s about time to complete the documentation and close the iteration.

Tasks completed this week:

I managed to fix some problems connected with introducing time series management in the KCSD solvers. This meant adjusting tests and any code samples where I called KCSDs.

Defining use cases for documentation

I have added some basic use cases to the documentation, so that users can start with an example.

The samples may be found here or built from the source.

Comparing more detailed reconstructions with their matlab versions

This is done using provided datasets: test_data1.mat for KCSD1D and data_01.mat provided with KCSD2D.

1D timeseries reconstruction

from pylab import *
from scipy.io import loadmat
from pykCSD.pykCSD import KCSD


def timeplot_1D():
    el_pos = loadmat('datasets/test_data1')['elPos']
    pots = loadmat('datasets/test_data1')['pots']
    params = {'xmin': 0.0, 'xmax': 1.9, 'gdX': 0.01,
              'h': 1.0, 'R_init': (el_pos.T[1]-el_pos.T[0])}
    k = KCSD(el_pos.T, np.array(pots), params)
    k.estimate_pots()
    k.estimate_csd()

    fig = plt.figure()
    
    ax11 = fig.add_subplot(1,2,1)
    im1 = imshow(k.solver.estimated_pots, aspect="auto", interpolation="none")
    plt.colorbar(im1)
    ax11.set_title('Estimated pots')
    ax11.set_xlabel('time steps')
    ax11.set_ylabel('space')
    
    ax12 = fig.add_subplot(1,2,2)
    im2 = imshow(k.solver.estimated_csd, aspect="auto", interpolation="none")
    plt.colorbar(im2)
    ax12.set_title('Estimated CSD')
    ax12.set_xlabel('time steps')
    ax12.set_ylabel('space')

    show()

if __name__ == "__main__":
    timeplot_1D()