tomwalters@138: #!/usr/bin/python
tomwalters@138: """
tomwalters@138: plot_munged_results.py
tomwalters@138: 
tomwalters@138: """
tomwalters@138: 
tomwalters@138: import numpy as np
tomwalters@138: import pylab as p
tomwalters@140: import matplotlib as mpl
tomwalters@140: mpl.use('PDF')
tomwalters@138: import matplotlib.pyplot as plt
tomwalters@138: 
tomwalters@140: 
tomwalters@138: f=open("results_test_all.csv","r")
tomwalters@138: results = dict()
tomwalters@138: for line in f:
tomwalters@138:   if line[0] != "#":
tomwalters@138:     values = line.strip().split(",")
tomwalters@138:     results.setdefault(values[3],dict())
tomwalters@138:     results[values[3]].setdefault(values[0], dict())
tomwalters@138:     results[values[3]][values[0]].setdefault(values[1], dict())
tomwalters@139:     results[values[3]][values[0]][values[1]].setdefault(int(values[4]), dict())
tomwalters@139:     results[values[3]][values[0]][values[1]][int(values[4])].setdefault(int(values[5]), dict())
tomwalters@139:     results[values[3]][values[0]][values[1]][int(values[4])][int(values[5])].setdefault(int(values[6]), dict())
tomwalters@138:     if values[2] == 'clean':
tomwalters@139:        snr = 50
tomwalters@138:     else:
tomwalters@138:        snr = int(values[2])
tomwalters@140:        results[values[3]][values[0]][values[1]][int(values[4])][int(values[5])][int(values[6])][snr] = float(values[7])
tomwalters@138: #    results[values[3]].append((values[1],values[2],values[2],values[4]))
tomwalters@138: 
tomwalters@138: ax = plt.subplot(111)
tomwalters@138: 
tomwalters@138: train_set = 'inner'
tomwalters@138: lines = []
tomwalters@138: labels = []
tomwalters@139: hmm_iterations = 2
tomwalters@139: hmm_states = 4
tomwalters@139: hmm_components = 4
tomwalters@138: for feature_type in ('mfcc', 'mfcc_vtln', 'aim'):
tomwalters@138:   for feature_subtype in results[train_set][feature_type].keys():
tomwalters@139:     this_line = results[train_set][feature_type][feature_subtype][hmm_states][hmm_components][hmm_iterations].items()
tomwalters@138:     this_line.sort(cmp=lambda x,y: x[0] - y[0])
tomwalters@138:     xs, ys = zip(*this_line)
tomwalters@138:     xs = list(xs)
tomwalters@138:     ys = list(ys)
tomwalters@138:     line, = ax.plot(xs,ys,'-o',linewidth=2)
tomwalters@138:     lines.append(line)
tomwalters@138:     labels.append(feature_type + "_" + feature_subtype)
tomwalters@138: p.legend(lines, labels, 'upper left', shadow=True)
tomwalters@138: p.xlabel('SNR/dB')
tomwalters@138: p.ylabel('Recognition performance %')
tomwalters@140: plt.savefig(output_file)