analyze.py

""" prints mean of iter and tol and std """
import numpy as np
import extractor as ex


def analyze(ref=''):
    """ analyze """
    #
    files = ['Picard'+ref+'.txt',
             'MHDtConvectionDiffusion'+ref+'.txt',
             'DivGrad'+ref+'.txt',
             'ModeNonlinearOp_ConvectionDiffusionVOp'+ref+'.txt',
             'ConvectionDiffusionVOp'+ref+'.txt']
    #
    for fil in files:
        stats = ex.extract(fil, ex.BelosIterPattern)
        count = len(stats[:, 0])
        print(fil, ' (', count, ')')
        if count != 0:
            print('total: ', sum(stats[:, 0]))
            print('iter: ', np.percentile(stats[:, 0], 10),
                  np.percentile(stats[:, 0], 50), np.percentile(stats[:, 0],
                                                                90))
            print('iter: ', np.mean(stats[:, 0]), '+- ', np.std(stats[:, 0]))
            print('max: ', np.max(stats[:, 0]))
            print('tol: ', np.percentile(stats[:, 1], 10),
                  np.percentile(stats[:, 1], 50), np.percentile(stats[:, 1],
                                                                90))
            print('tol: ', np.mean(stats[:, 1]), '+- ', np.std(stats[:, 1]))
        print()


def compares(paths=None, ref=''):
    """ compares different folders """
    if paths is None:
        paths = ['./']
    #
    files = ['Picard'+ref+'.txt',
             # 'MHDtConvectionDiffusion'+ref+'.txt',
             'DivGrad'+ref+'.txt',
             'ModeNonlinearOp_ConvectionDiffusionVOp'+ref+'.txt',
             'ConvectionDiffusionVOp'+ref+'.txt']
    #
    npath = len(paths)
    for fil in files:
        iter_tot = np.zeros(npath)
        iter_stds = np.zeros(npath)
        iter_max = np.zeros(npath)
        iter_means = np.zeros(npath)
        tol_means = np.zeros(npath)
        tol_stds = np.zeros(npath)
        counts = np.zeros(npath)
        print(fil, ': ', paths)
        for i, path in enumerate(paths):
            stats = ex.extract(path+fil, ex.BelosIterPattern)
            counts[i] = len(stats[:, 0])
            if counts[i] != 0:
                iter_tot[i] = np.sum(stats[:, 0])
                iter_means[i] = np.mean(stats[:, 0])
                iter_max[i] = np.max(stats[:, 0])
                tol_means[i] = np.mean(stats[:, 1])
                iter_stds[i] = np.std(stats[:, 0])
                tol_stds[i] = np.std(stats[:, 1])
        print('counts: ', counts)
        print('total:  ', iter_tot)
        print('iter:   ', iter_means)
        print('+-      ', iter_stds)
        print('max     ', iter_max)
        print('tol:    ', tol_means)
        print('+-      ', tol_stds)
        print()
    print()


if __name__ == "__main__":
    analyze()