mostly conforms to pylint/pep8. just need to remove duplicate code, w…

…hich likely amounts to removing redundant sripts.

examples/basic_example.py

@@ -1,4 +1,9 @@
 #!/usr/bin/env python
+'''
+Basic example which creates a dummy test dataset and a
+benchmark collection.  This example 'trains' on the benchmark
+set and displays the results of the comparison.
+'''
 
 import numpy
 
@@ -9,16 +14,17 @@
     'Uniform': numpy.random.uniform(size=100)
 }
 
-n_benchmark_collections = 5
-benchmark_labels = ['Benchmark_%d' % i for i in range(n_benchmark_collections)]
+N_BENCHMARK_COLLECTIONS = 5
+benchmark_labels = ['Benchmark_%d' % i for i in range(N_BENCHMARK_COLLECTIONS)]
 benchmark_data = {
     benchmark_label:{'Gaussian': numpy.random.normal(size=100),
                      'Uniform': numpy.random.uniform(size=100)}
     for benchmark_label in benchmark_labels
 }
 
 # histogram the data
-test_histograms = {name:numpy.histogram(data)[0] for name, data in test_data.items()}
+test_histograms = {name:numpy.histogram(data)[0]
+                   for name, data in test_data.items()}
 benchmark_histograms = dict()
 for name, bm_data in benchmark_data.items():
     benchmark_histograms[name] = {n:numpy.histogram(data)[0]

examples/classic_fit_example.py

@@ -1,5 +1,9 @@
 #!/usr/bin/env python
 
+'''
+Example illustrating the classic method, meaning fitting by hand.
+'''
+
 import numpy
 import pylab
 import scipy.optimize
@@ -16,25 +20,32 @@
         data.append(time)
 
 def gauss(x, *p):
-    A, mu, sigma = p
-    return A*numpy.exp(-(x-mu)**2/(2.*sigma**2))
+    '''
+    The gaussian distribution.
+    '''
+    amplitude, mean, sigma = p
+    return amplitude*numpy.exp(-(x-mean)**2/(2.*sigma**2))
 
 hist, bin_edges = numpy.histogram(data)
 bin_centers = (bin_edges[:-1] + bin_edges[1:])/2
 p0 = [1000., 0., 1.]
 p1 = [1000., 0., 1.]
 print(hist)
 print(bin_centers)
-coeff, var_matrix = scipy.optimize.curve_fit(gauss, bin_centers, hist, p0=p1)
 
-print('Fitted amplitude = %f' % coeff[0])
-print('Fitted mean = %f' % coeff[1])
-print('Fitted standard deviation = %f' % coeff[2])
+result = scipy.optimize.curve_fit(gauss, bin_centers, hist, p0=p1)
+fitted_amplitude = result[0]
+fitted_mean = result[1]
+fitted_std_dev = result[2]
+
+print('Fitted amplitude = %f' % fitted_amplitude)
+print('Fitted mean = %f' % fitted_mean)
+print('Fitted standard deviation = %f' % fitted_std_dev)
 
+coeff = tuple(result[:3])
 hist_fit = gauss(bin_centers, *coeff)
 pylab.hist(data)
 pylab.plot(bin_centers, gauss(bin_centers, *p0), label='Initial')
 pylab.plot(bin_centers, hist_fit, label='Fitted data')
 
 pylab.show()
-