Source Code for Module parsimony.utils.plot

  1  # -*- coding: utf-8 -*- 
  2  """ 
  3  Created on Mon Jan  6 14:13:42 2014 
  4   
  5  Copyright (c) 2013-2014, CEA/DSV/I2BM/Neurospin. All rights reserved. 
  6   
  7  @author:  Edouard Duchesnay, Tommy Löfstedt 
  8  @email:   edouard.duchesnay@cea.fr, lofstedt.tommy@gmail.com 
  9  @license: BSD 3-clause. 
 10  """ 
 11  import numpy as np 
 12  import scipy.stats as ss 
 13   
 14  __all__ = ["plot_map2d", "plot_classes"] 
 15   
 16  COLORS = ["b", "g", "r", "c", "m", "y", "k", "w"] 
 17  COLORS_FULL = ["blue", "green", "red", "cyan", "magenta", "yellow", "black", 
 18                 "white"] 
 19  MARKERS = ["+", ".", "o", "*", "p", "s", "x", "D", "h", "^"] 
 20  LINE_STYLE = ["-", "--", "--.", ":"] 
 21   
 22   
 23 -def plot_map2d(map2d, plot=None, title=None, limits=None, 
 24                 center_cmap=True): 
 25      import matplotlib.pyplot as plt 
 26   
 27      if plot is None: 
 28          plot = plt 
 29   
 30      map2d = map2d.squeeze() 
 31   
 32      if len(map2d.shape) != 2: 
 33          raise ValueError("input map is not 2D") 
 34   
 35      if np.asarray(limits).size is 2: 
 36          mx = limits[0] 
 37          mi = limits[1] 
 38      else: 
 39          mx = map2d.max() 
 40          mi = map2d.min() 
 41   
 42      if center_cmap: 
 43          mx = np.abs([mi, mx]).max() 
 44          mi = -mx 
 45   
 46      cax = plot.matshow(map2d, cmap=plt.cm.coolwarm) 
 47      frame = plt.gca() 
 48      frame.get_xaxis().set_visible(False) 
 49      frame.get_yaxis().set_visible(False) 
 50      #k = 1 
 51      #while (10 ** k * mx) < 1 and k < 10: 
 52      #    k += 1 
 53      #ticks = np.array([-mi, -mi / 4 - mi / 2, 0, mx / 2, mx / 2, 
 54      #                  mx]).round(k + 2) 
 55      cbar = plt.colorbar(cax)  # , ticks=ticks) 
 56      cbar.set_clim(vmin=mi, vmax=mx) 
 57   
 58      if title is not None: 
 59          plt.title(title) 
 60   
 61   
 62 -def plot_classes(X, classes, title=None, xlabel=None, ylabel=None, show=True): 
 63   
 64      import matplotlib.pyplot as plot 
 65   
 66      if isinstance(classes, np.ndarray): 
 67          classes = classes.ravel().tolist() 
 68   
 69      cls = list(set(classes)) 
 70   
 71      # TODO: Add the other cases. 
 72      if X.shape[1] == 2: 
 73   
 74          for i in xrange(len(cls)): 
 75              c = cls[i] 
 76              cl = np.array(classes) == c 
 77  #            print cl.shape 
 78  #            print X[cl, 0].shape 
 79  #            print X[cl, 1].shape 
 80              plot.plot(X[cl, 0], X[cl, 1], 
 81                        color=COLORS[i % len(COLORS)], 
 82                        marker='.', 
 83                        markersize=15, 
 84                        linestyle="None") 
 85   
 86          if title is not None: 
 87              plot.title(title, fontsize=22) 
 88   
 89          if xlabel is not None: 
 90              plot.xlabel(xlabel, fontsize=16) 
 91          if ylabel is not None: 
 92              plot.ylabel(ylabel, fontsize=16) 
 93   
 94          if show: 
 95              plot.show() 
 96   
 97   
 98 -def plot_errorbars(X, classes=None, means=None, alpha=0.05, 
 99                     title=None, xlabel=None, ylabel=None, 
100                     colors=None, 
101                     show=True, latex=True): 
102   
103      import matplotlib.pyplot as plot 
104   
105      B, n = X.shape 
106      if classes is None: 
107          classes = np.array([1] * n) 
108      classes = np.array(classes).reshape((n, 1)) 
109   
110      if colors is None: 
111          colors = COLORS 
112   
113      data_mu = np.mean(X, axis=0) 
114      data_df = np.array([B - 1] * n) 
115      data_sd = np.std(X, axis=0) 
116   
117      x = np.arange(1, n + 1) 
118   
119      labels, cls_inverse = np.unique(classes, return_inverse=True) 
120      labels = labels.ravel().tolist() 
121   
122  #    plot.figure() 
123      if latex: 
124          plot.rc('text', usetex=True) 
125          plot.rc('font', family='serif') 
126      if means is not None: 
127          plot.plot(x, means, '*', 
128                    markerfacecolor="black", markeredgecolor="black", 
129                    markersize=10) 
130   
131      ci = ss.t.ppf(1.0 - alpha / 2.0, data_df) * data_sd / np.sqrt(B) 
132   
133      for i in xrange(len(labels)): 
134          ind = np.where(classes == labels[i])[0] 
135   
136          plot.errorbar(x[ind], 
137                        data_mu[ind], 
138                        yerr=ci[ind], 
139                        fmt='o' + colors[i % len(colors)], 
140                        color=colors[i % len(colors)], 
141                        ecolor=colors[i % len(colors)], 
142                        elinewidth=2, 
143                        markeredgewidth=2, 
144                        markeredgecolor=colors[i % len(colors)], 
145                        capsize=5) 
146   
147      plot.xlim((0, n + 1)) 
148      mn = np.min(data_mu - ci) 
149      mx = np.max(data_mu + ci) 
150      d = mx - mn 
151      plot.ylim((mn - d * 0.05, mx + d * 0.05)) 
152   
153      if xlabel is not None: 
154          plot.xlabel(xlabel) 
155      if ylabel is not None: 
156          plot.ylabel(ylabel) 
157      if title is not None: 
158          plot.title(title) 
159      if show: 
160          plot.show() 
161