regmaxs/regmaxsn/scripts/analysis/regErrorVsAnisotropicscaling.py

import os
import numpy as np
from matplotlib import pyplot as plt
import seaborn as sns
import json
from regmaxsn.core.RegMaxSPars import RegMaxSParNames
from regmaxsn.core.misc import parFileCheck
from regmaxsn.core.matplotlibRCParams import mplPars

homeFolder = os.path.expanduser('~')

plt.ion()


# Example colFunc, takes refSWC and testSWC and returns an object that can be passed to matplotlib plotting argument
# color. For example return objects could be 'b' (blue), 'r' (red) and  [0, 0, 1] (green).
# def colFunc(refSWC, testSWC):
#
#     testInd = int(os.path.split(testSWC)[1][25:-4])
#
#     if testInd in range(2, 5):
#         return 'r'
#     else:
#         return 'b'

colFunc = None

def regErrorVsAIScaling(parFile, colFunc=None):
    # Axis 1: neuron pairs; Axis 2: (reg accuracy, anisotropic scaling)

    parsList = parFileCheck(parFile, RegMaxSParNames)
    translErrStats = np.empty((len(parsList), 2))
    for parInd, par in enumerate(parsList):

        refSWC = par['refSWC']
        testSWC = par['testSWC']
        resFile = par['resFile']
        thresh = par['gridSizes'][-1]

        print('Doing ' + repr((refSWC, resFile)))

        origJSON = testSWC[:-4] + '.json'

        if os.path.isfile(origJSON):
            with open(origJSON, 'r') as fle:
                pars = json.load(fle)
                scales = np.array(pars['scale'])
        else:
            raise(IOError('File not found: {}'.format(origJSON)))

        scalesOrdered = np.sort(scales)
        scalesRelative = np.mean([scalesOrdered[0] / scalesOrdered[1],
                                  scalesOrdered[0] / scalesOrdered[2],
                                  scalesOrdered[1] / scalesOrdered[2]])

        refPts = np.loadtxt(refSWC)[:, 2:5]
        testPts = np.loadtxt(resFile)[:, 2:5]

        if refPts.shape[0] != testPts.shape[0]:

            print('Number of points do not match for ' + refSWC + 'and' + testSWC)
            continue


        ptDiff = np.linalg.norm(refPts - testPts, axis=1)
        translErrStats[parInd, 0] = 100 * sum(ptDiff <= thresh) / float(ptDiff.shape[0])
        translErrStats[parInd, 1] = scalesRelative

    sns.set(rc=mplPars)
    with sns.axes_style('whitegrid'):
        fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(10, 8))

        for parInd, vals in enumerate(translErrStats):
            try:
                if colFunc:
                    col = colFunc(parsList[parInd]['refSWC'], parsList[parInd]['testSWC'])
                else:
                    col = 'b'

                ax.plot(vals[1], vals[0], color=col, marker='o', ls='None', ms=10)
            except Exception as e:
                raise(Exception('Problem with plotting. There could be a problem with argument colFunc'))

    ax.set_xlabel('measure of anisotropic scaling')
    ax.set_ylabel('\% points closer than \nthe lowest grid size')
    ax.set_ylim(-10, 110)

    fig.tight_layout()

    return fig

if __name__ == '__main__':

    import sys
    assert len(sys.argv) == 2, 'Improper usage! Please use as \'python refErrorVsAnisotropicscaling.py parFile\''

    parFile = sys.argv[1]
    fig = regErrorVsAIScaling(parFile, colFunc)
    raw_input('Press any key to close figures and quit:')