benchy_final.py

from benchy.api import Benchmark, BenchmarkSuite, BenchmarkRunner
import matplotlib.pyplot as plt

common_setup = """
import numpy
X = numpy.random.uniform(1,5,(1000,))
"""
statement = "cosine_distances(X, X)"


setup_bk1 = common_setup + """
import scipy.spatial.distance as ssd
X = X.reshape(-1,1)
def cosine_distances(X, Y):
    return 1. - ssd.cdist(X, Y, 'cosine')

"""
benchmark1 = Benchmark(statement, setup_bk1, name="scipy.spatial 0.8.0")

setup_bk2 = common_setup + """
from sklearn.metrics.pairwise import cosine_similarity as cosine_distances
"""
benchmark2 = Benchmark(statement, setup_bk2, name="sklearn 0.13.1")

setup_bk3 = common_setup + """
from nltk import cluster
def cosine_distances(X, Y):
    return 1. - cluster.util.cosine_distance(X, Y)

"""
benchmark3 = Benchmark(statement, setup_bk3, name="nltk.cluster")


setup_bk4 = common_setup + """
import numpy, math
def cosine_distances(X, Y):
    return 1. -  numpy.dot(X, Y) / (math.sqrt(numpy.dot(X, X)) *
                                     math.sqrt(numpy.dot(Y, Y)))
"""
benchmark4 = Benchmark(statement, setup_bk4, name="numpy")


suite = BenchmarkSuite()
suite.append(benchmark1)
suite.append(benchmark2)
suite.append(benchmark3)
suite.append(benchmark4)

runner = BenchmarkRunner(suite, '.', 'Cosine benchmarks')
n_benchs, results = runner.run()

fig = runner.plot_relative(results, horizontal=True, logy=True)
plt.savefig('%s_r.png' % runner.name.replace(' ', '_'))  # bbox_inches='tight')

runner.plot_absolute(results, horizontal=False, logy=True)
plt.savefig('%s.png' % runner.name.replace(' ', '_'))  # bbox_inches='tight')

rst_text = runner.to_rst(results, runner.name.replace(' ', '_') + 'png',
        runner.name.replace(' ', '_') + '_r.png')
with open('bench.rst', 'w') as f:
        f.write(rst_text)