Truncated SVD

from __future__ import print_function
from scipy.sparse import *
from sklearn.decomposition import TruncatedSVD
import matplotlib.pyplot as plt
from matplotlib import cm as CM
import numpy as np

import time

class Timer(object):
    def __init__(self, name):
        print("%s: " % name, end="")
    def __enter__(self):
        self.t0 = time.time()
    def __exit__(self, *args):
        print("%.3fs" % (time.time() - self.t0))

def read_dimension(f):
    r = 0;
    c = 0
    for line in f:
        c = max(c, max([int(x) for x in line.split()]))
        r += 1
    return (r,c+1)

def read_sparse_matrix(f):
    i = 0
    for line in f:
        for value in line.split():
            yield (i, int(value), 1)
        i += 1
            
if __name__ == '__main__':
    filename = r'q.txt'

    with Timer("Dimension"):
        d = read_dimension(open(filename))
    print(d)

    with Timer("Create sparse matrix"):
        m = lil_matrix(d)
        for (r,c,v) in read_sparse_matrix(open(filename)):
            m[r,c] = v
        mc = m.tocsr()

    with Timer("ARPACK SVD"):
        X1 = TruncatedSVD(n_components=2, algorithm="arpack").fit_transform(mc)

    plt.hexbin(*zip(*X1), cmap=CM.jet, bins=None)
    cb = plt.colorbar()
    cb.set_label('mean value')
    plt.show()