def test_format_method_comparison_heatmaps(self):
obs = format_method_comparison_heatmaps(self.real_results1,
self.sim_results1, [Adonis(), Anosim()])
self.assertEqual(obs, exp_method_comparison_heatmaps1)
self.assertRaises(ValueError, format_method_comparison_heatmaps,
self.real_results2, {}, [Adonis(), Anosim()])
self.assertRaises(ValueError, format_method_comparison_heatmaps,
self.real_results1, self.sim_results2, [Adonis(), Anosim()])
def create_method_comparison_heatmaps(in_dir, workflow, heatmap_methods):
"""Generates heatmaps showing the correlation between each pair of methods.
Generates two heatmaps (one for Pearson correlation, one for Spearman
correlation). Uses all available results (e.g. all even sampling depths,
metrics, and datasets) that match between each pair of methods as input to
the correlation coefficient methods. This includes both real data (e.g.
original datasets and shuffled datasets) as well as simulated data.
"""
real_data_results = _collate_real_data_results(in_dir, workflow)
sim_data_results = _collate_simulated_data_results(in_dir, workflow)
heatmap_data = format_method_comparison_heatmaps(real_data_results,
sim_data_results,
heatmap_methods)
for correlation_method, data in heatmap_data.items():
# Generate the heatmap. Code based on
# http://matplotlib.org/users/tight_layout_guide.html and
# http://psaffrey.wordpress.com/2010/07/05/chromosome-interactions-
# heatmaps-and-matplotlib/
fig = figure()
ax = subplot(111)
cmap = cm.get_cmap()
# default value is 'k' (black)
cmap.set_bad('w')
im = ax.imshow(data, cmap=cmap, interpolation='nearest')
labels = [method.DisplayName for method in heatmap_methods]
colorbar(im, use_gridspec=True)
xticks(range(len(labels)), labels, rotation=90)
yticks(range(len(labels)), labels)
for loc, spine in ax.spines.items():
if loc in ['right','top']:
# don't draw spine
spine.set_color('none')
ax.xaxis.set_ticks_position('bottom')
ax.yaxis.set_ticks_position('left')
ax.grid(True, which='minor')
tight_layout()
savefig(join(in_dir, '%s_heatmap.pdf' % correlation_method),
format='pdf')
savefig(join(in_dir, '%s_heatmap.png' % correlation_method),
format='png', dpi=1000)
