test_iterutils.py

Go to the documentation of this file.

import doctest
import functools
import math
import sys
from lal import iterutils
 
 
#
# check examples in documentation
#
 
 
failures = doctest.testmod(iterutils)[0]
if failures:
        sys.exit(bool(failures))
 
 
#
# check CDF of randindex.  requires matplotlib for visual check of results.
# without matplotlib, simply runs the code to test for failures
#
 
 
def gen_randindex_results(n, N):
        randindex = functools.partial(next, iter(iterutils.randindex(10, 100, n)))
        counts = [0.] * 100
        p = [float("-inf")] * 100
        for i in range(N):
                x, lnP = randindex()
                counts[x] += 1.
                p[x] = lnP
        for i in range(len(counts)):
                counts[i] /= N
                p[i] = math.exp(p[i])
        return list(range(100)), counts, p
 
try:
        import matplotlib
        matplotlib.use("Agg")
        from matplotlib import figure
        from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
        fig = figure.Figure()
        FigureCanvas(fig)
        axes = fig.gca()
        axes.loglog()
except ImportError:
        axes = None
 
N = 1000000
x, counts, p = gen_randindex_results(.5, N)
if axes is not None:
        axes.plot(x, counts, "k+", label = "Observed fractions (n=0.5)")
        axes.plot(x, p, "k-", label = "Reported probabilities (n=0.5)")
x, counts, p = gen_randindex_results(1., N)
if axes is not None:
        axes.plot(x, counts, "b+", label = "Observed fractions (n=1)")
        axes.plot(x, p, "b-", label = "Reported probabilities (n=1)")
x, counts, p = gen_randindex_results(2., N)
if axes is not None:
        axes.plot(x, counts, "g+", label = "Observed fractions (n=2)")
        axes.plot(x, p, "g-", label = "Reported probabilities (n=2)")
 
if axes is not None:
        axes.set_xlim((9., 100.))
        axes.set_title("randindex(10, 100, n), %d samples" % N)
        axes.legend()
        fig.savefig("iterutils_randindex_histogram.png")