def setup_axes(fig, header):
from mpl_toolkits.axes_grid import make_axes_locatable
ax0 = pywcsgrid2.subplot(111, wcs=header)
divider = make_axes_locatable(ax0)
gh1 = pywcsgrid2.GridHelperSimple(wcs=header, axis_nums=[0, 2])
ax_v = divider.new_vertical(1.5, pad=0.1, sharex=ax0,
axes_class=pywcsgrid2.Axes,
grid_helper=gh1)
fig.add_axes(ax_v)
gh2 = pywcsgrid2.GridHelperSimple(wcs=header, axis_nums=[2, 1])
ax_h = divider.new_horizontal(1.5, pad=0.1, sharey=ax0,
axes_class=pywcsgrid2.Axes,
grid_helper=gh2)
fig.add_axes(ax_h)
ax_h.axis["left"].toggle(label=False, ticklabels=False)
ax_v.axis["bottom"].toggle(label=False, ticklabels=False)
return ax0, ax_v, ax_h
def drawmap(DATA,TITLESTRING,PROD,UNITS):
F = plt.gcf() # Gets the current figure
m.drawstates(color='k', linewidth=1.25)
m.drawcoastlines(color='k')
m.drawcountries(color='k', linewidth=1.25)
#m.readshapefile(shapefile='/data/geog/shapefiles/fe_2007_40_county.shp',name='COUNTY',drawbounds='True')
#m.readshapefile(shapefile='/data/geog/shapefiles/fe_2007_48_county.shp',name='COUNTY',drawbounds='True')
#plt.suptitle('%s' % UNITS, fontsize = 11, x = 0.08, y = 0.105)
plt.title('UW WRF-ARW %s (%s) Valid: %s' % (TITLESTRING, UNITS, curtimestring), \
fontsize=11,bbox=dict(facecolor='white', alpha=0.65),\
x=0.5,y=.95,weight = 'demibold',style='oblique', \
stretch='normal', family='sans-serif')
# Code to make the colorbar outside of the main axis, on the bottom, and lined up
ax = plt.gca() # Gets the current axes
divider = make_axes_locatable(ax) # Lets us move axes around
cax = divider.append_axes("bottom", size="2%",pad=-0.02,axes_class=maxes.Axes) # Adds an axis for the colorbar
F.add_axes(cax) # Adds the new axis to the figure as the current working axis
bar = plt.colorbar(DATA,cax=cax,orientation='horizontal',format='%4.2f',extend='both') # Plots colorbar in new axis
bar.ax.xaxis.set_major_locator(matplotlib.ticker.MultipleLocator(base=1.0)) # Make the colorbars numbers nice
bar.update_ticks()
file_id = '%s_%s_f%02d' % (dom, PROD, time+restart_time)
filename = '%s.png' % (file_id)
plt.savefig(filename,bbox_inches='tight') # Saves the figure with small margins
plt.close()
#if export_flag == 1:
# Convert the figure to a gif file
os.system('convert -render -flatten %s %s.gif' % (filename, file_id))
os.system('rm -f %s' % filename)
def plot(axes, net):
classname = net.__class__.__name__
axes.set_xticks([])
axes.set_yticks([])
divider = make_axes_locatable(axes)
subaxes = divider.new_vertical(1.0, pad=0.4, sharex=axes)
fig.add_axes(subaxes)
subaxes.set_xticks([])
subaxes.yaxis.set_major_locator(matplotlib.ticker.MaxNLocator(2))
subaxes.yaxis.set_ticks_position('right')
subaxes.set_ylabel('Distortion')
subaxes.set_xlabel('Time')
Y = net.distortion[::1]
X = np.arange(len(Y))/float(len(Y)-1)
subaxes.plot(X,Y)
if classname == 'NG':
plt.title('Neural Gas', fontsize=20)
elif classname == 'SOM':
plt.title('Self-Organizing Map', fontsize=20)
elif classname == 'DSOM':
plt.title('Dynamic Self-Organizing Map', fontsize=20)
axes.axis([0,1,0,1])
axes.set_aspect(1)
codebook = net.codebook
axes.imshow(codebook, interpolation='nearest')
#interpolation='bicubic')
if classname == 'NG':
axes.text(0.5, -0.01,
r'$\lambda_i = %.3f,\lambda_f = %.3f, \varepsilon_i=%.3f, \varepsilon_f=%.3f$' % (
net.sigma_i, net.sigma_f, net.lrate_i, net.lrate_f),
fontsize=16,
horizontalalignment='center',
verticalalignment='top',
transform = axes.transAxes)
if classname == 'SOM':
axes.text(0.5, -0.01,
r'$\sigma_i = %.3f,\sigma_f = %.3f, \varepsilon_i=%.3f, \varepsilon_f=%.3f$' % (
net.sigma_i, net.sigma_f, net.lrate_i, net.lrate_f),
fontsize=16,
horizontalalignment='center',
verticalalignment='top',
transform = axes.transAxes)
elif classname == 'DSOM':
axes.text(0.5, -0.01,
r'$elasticity = %.2f$' % (net.elasticity),
fontsize=16,
horizontalalignment='center',
verticalalignment='top',
transform = axes.transAxes)
def demo_images_side_by_sied(ax):
from mpl_toolkits.axes_grid import make_axes_locatable
divider = make_axes_locatable(ax)
Z, extent = get_demo_image()
ax2 = divider.new_horizontal(size="100%", pad=0.05)
fig1 = ax.get_figure()
fig1.add_axes(ax2)
ax.imshow(Z, extent=extent, interpolation="nearest")
ax2.imshow(Z, extent=extent, interpolation="nearest")
for tl in ax2.get_yticklabels():
tl.set_visible(False)
def plot_dist(self, axes):
''' Plot network on given axes
'''
classname = self.__class__.__name__
fig = plt.gcf()
divider = make_axes_locatable(axes)
axes.axis([0,1,0,.5])
Y = self.distortion[::1]
X = np.arange(len(Y))/float(len(Y)-1)
axes.plot(X,Y)
def demo_locatable_axes_easy(ax):
from mpl_toolkits.axes_grid import make_axes_locatable
divider = make_axes_locatable(ax)
ax_cb = divider.new_horizontal(size="5%", pad=0.05)
fig1 = ax.get_figure()
fig1.add_axes(ax_cb)
Z, extent = get_demo_image()
im = ax.imshow(Z, extent=extent, interpolation="nearest")
plt.colorbar(im, cax=ax_cb)
ax_cb.yaxis.tick_right()
for tl in ax_cb.get_yticklabels():
tl.set_visible(False)
ax_cb.yaxis.tick_right()
def QC_Chromosome_Plot(calls, title=None, outfile=None):
x = np.array(calls.calls["num_probes"].values)
y = np.array(calls.calls["median_svdzrpkm"].values)
fig = plt.figure(1, figsize=(9,9))
from mpl_toolkits.axes_grid import make_axes_locatable
axScatter = plt.subplot(111)
divider = make_axes_locatable(axScatter)
# create a new axes with a height of 1.2 inch above the axScatter
axHistx = divider.new_vertical(1.2, pad=0.1, sharex=axScatter)
# create a new axes with a width of 1.2 inch on the right side of the
# axScatter
axHisty = divider.new_horizontal(1.2, pad=0.1, sharey=axScatter)
fig.add_axes(axHistx)
fig.add_axes(axHisty)
# make some labels invisible
plt.setp(axHistx.get_xticklabels() + axHisty.get_yticklabels(),
visible=False)
# the scatter plot:
#axScatter.scatter(x[x<0], y[x<0], lw=0, alpha=0.3, color="r")
axScatter.scatter(x[x>=0], y[x>=0], lw=0, alpha=0.3, color="b")
#axScatter.set_aspect(1.)
axScatter.set_xscale("symlog")
axHistx.set_xscale("symlog")
axScatter.set_xlabel("Size of call (# of probes)")
axScatter.set_ylabel("Signal Strength (Median SVD-ZRPKM)")
axHistx.hist(x, bins=np.arange(0,np.max(x)), histtype="stepfilled", lw=0,align='left')
axHisty.hist(y, bins=200, orientation='horizontal', histtype="stepfilled", lw=0)
for tl in axHistx.get_xticklabels():
tl.set_visible(False)
axHisty.set_xticklabels(["%d" % i for i in axHisty.get_xticks()], rotation=-90)
if title is not None:
axHistx.set_title(title)
if outfile is not None:
plt.savefig(outfile)
def _stabilityassessment(headers, data1d, dist, fig_correlmatrices, correlmatrixaxes, std_multiplier,
correlmatrix_colormap,
correlmatrix_filename, logarithmic_correlmatrix=True, cormaptest=True):
# calculate and plot correlation matrix
cmatrix, badidx, rowavg = correlmatrix(data1d, std_multiplier, logarithmic_correlmatrix)
rowavgmean = rowavg.mean()
rowavgstd = rowavg.std()
writemarkdown('#### Assessing sample stability')
writemarkdown("- Mean of row averages: " + str(rowavgmean))
writemarkdown("- Std of row averages: " + str(rowavgstd) + ' (%.2f %%)' % (rowavgstd / rowavgmean * 100))
img = correlmatrixaxes.imshow(cmatrix, interpolation='nearest', cmap=matplotlib.cm.get_cmap(correlmatrix_colormap))
cax = make_axes_locatable(correlmatrixaxes).append_axes('right', size="5%", pad=0.1)
fig_correlmatrices.colorbar(img, cax=cax)
fsns = [h.fsn for h in headers]
correlmatrixaxes.set_title('%.2f mm' % dist)
correlmatrixaxes.set_xticks(list(range(len(data1d))))
correlmatrixaxes.set_xticklabels([str(f) for f in fsns], rotation='vertical')
correlmatrixaxes.set_yticks(list(range(len(data1d))))
correlmatrixaxes.set_yticklabels([str(f) for f in fsns])
np.savez_compressed(correlmatrix_filename,
correlmatrix=cmatrix, fsns=np.array(fsns))
# Report table on sample stability
tab = [['FSN', 'Date', 'Discrepancy', 'Relative discrepancy ((x-mean(x))/std(x))', 'Quality', 'Quality (cormap)']]
badfsns = []
badfsns_datcmp = []
if cormaptest:
matC, matp, matpadj, datcmp_ok = datcmp(*data1d)
else:
datcmp_ok = [not x for x in badidx]
for h, bad, discr, dcmp_ok in zip(headers, badidx, rowavg, datcmp_ok):
tab.append([h.fsn, h.date.isoformat(), discr, (discr - rowavgmean) / rowavgstd,
["\u2713", "\u2718\u2718\u2718\u2718\u2718"][bad],
["\u2713", "\u2718\u2718\u2718\u2718\u2718"][dcmp_ok != 1]])
if bad:
badfsns.append(h.fsn)
if (not dcmp_ok and not np.isnan(dcmp_ok)):
badfsns_datcmp.append(h.fsn)
tab = ipy_table.IpyTable(tab)
tab.apply_theme('basic')
return badfsns, badfsns_datcmp, tab, rowavg
def plot(self, axes):
''' Plot network on given axes
'''
classname = self.__class__.__name__
fig = plt.gcf()
divider = make_axes_locatable(axes)
axes.axis([0,1,0,1])
# Plot samples
axes.scatter(self.samples[:,0], self.samples[:,1], s=1, color='g', alpha=0.5)
C = self.adj
Cx,Cy = C[...,0], C[...,1]
if classname != 'SSk':
for i in range(C.shape[0]):
axes.plot (Cx[i,:], Cy[i,:], 'k', alpha=1, lw=1.5)
for i in range(C.shape[1]):
axes.plot (Cx[:,i], Cy[:,i], 'k', alpha=1, lw=1.5)
def plot_eigenvectors(ax, Y, idx, colormap):
from matplotlib.ticker import MaxNLocator
from mpl_toolkits.axes_grid import make_axes_locatable
divider = make_axes_locatable(ax)
ax2 = divider.new_vertical(size="100%", pad=0.05)
fig1 = ax.get_figure()
fig1.add_axes(ax2)
ax2.set_title("Eigenvectors", fontsize=10)
ax2.scatter(np.arange(0, len(Y)), Y[:, 0], s=10, c=idx, cmap=colormap, alpha=0.9, facecolors="none")
ax2.axhline(0, ls="--", c="k")
ax2.yaxis.set_major_locator(MaxNLocator(4))
ax.yaxis.set_major_locator(MaxNLocator(4))
ax.axhline(0, ls="--", c="k")
ax.scatter(np.arange(0, len(Y)), Y[:, 1], s=10, c=idx, cmap=colormap, alpha=0.9, facecolors="none")
ax.set_xlabel("index", fontsize=8)
ax2.set_ylabel("2nd Smallest", fontsize=8)
ax.set_ylabel("3nd Smallest", fontsize=8)
change_tick_fontsize(ax, 8)
change_tick_fontsize(ax2, 8)
for tl in ax2.get_xticklabels():
tl.set_visible(False)
def Scatter_hist(data1,
data2,
data1b=False,
data2b=False,
xbinwidth=0.5,
ybinwidth=0.5,
SSE=None,
SSEb=None,
vmax=1000.,
colormaps=cm.YlOrBr,
cleanstyle=False,
roodlichter=0.5,
*args,
**kwargs):
'''
Three-parts plot with the two parameter sitdirbutions plotted on the sides
Parameters
-----------
data1: ndarray
dataset 1 for x-axis
data2: ndarray
dataset 2 for y-axis
data1b: ndarray
dataset to plot along the data 1 set
data2b: ndarray
dataset to plot along the data 2 set
binwidth: float
defines the width of the bins relative to the data used
cleanstyle: bool True|False
if True, a more minimalistic version of the plot is given
*args, **kwargs: args
arguments given toi the scatter plot
example: s=15, marker='o', edgecolors= 'k',facecolor = 'white'
Returns
---------
fig: matplotlib.figure.Figure object
the resulting figure
axScatter: matplotlib.axes.AxesSubplot object
the scatter plot with the datapoints, can be used to add labels or
change the current ticks settings
axHistx: matplotlib.axes.AxesSubplot object
the x-axis histogram
axHisty: matplotlib.axes.AxesSubplot object
the y-axis histogram
Examples
----------
>>> nMC = 1000
>>> parval1 = np.random.gamma(5.5,size=nMC)
>>> parval2 = np.random.gamma(8.0,size=nMC)
>>> parnames = ['par1','par2']
>>> fig,axScatter,axHistx,axHisty = Scatter_hist(parval1,parval2,
cleanstyle = True, s=48,
marker='o', edgecolors= 'k',
facecolor = 'none',alpha=0.7)
>>> parval1b = np.random.uniform(low=0.0, high=30.0,size=nMC)
>>> parval2b = np.random.uniform(low=0.0, high=30.0,size=nMC)
>>> fig,axScatter,axHistx,axHisty = Scatter_hist(parval1,parval2,parval1b,
parval2b, cleanstyle = True,
s=48, marker='o',
edgecolors= 'k',
facecolor = 'none',
alpha=0.7)
Notes
------
Typical application is to check the dependency of two posterior
parameter distrbutions, eventually compared with their selected posteriors
If a second dataset is added to compare, the style options of the scatter
plot are fixed and the *args, **kwargs have no influence
'''
if not isinstance(data1, np.ndarray):
raise Exception('dataset 1 need to be numpy ndarray')
if not isinstance(data2, np.ndarray):
raise Exception('dataset 2 need to be numpy ndarray')
if isinstance(data1b, np.ndarray):
if not isinstance(data2b, np.ndarray):
raise Exception('Always combine the data of both')
if isinstance(data2b, np.ndarray):
if not isinstance(data1b, np.ndarray):
raise Exception('Always combine the data of both')
fig = plt.figure(figsize=(10, 10))
axScatter = plt.subplot(111)
divider = make_axes_locatable(axScatter)
#axScatter.set_aspect('equal')
axScatter.set_autoscale_on(True)
# create a new axes with above the axScatter
axHistx = divider.new_vertical(1.5, pad=0.0001, sharex=axScatter)
# create a new axes on the right side of the
# axScatter
axHisty = divider.new_horizontal(1.5, pad=0.0001, sharey=axScatter)
fig.add_axes(axHistx)
fig.add_axes(axHisty)
# now determine nice limits by hand:
# binwidth = binwidth
xmin = np.min(data1)
xmax = np.max(data1)
ymin = np.min(data2)
ymax = np.max(data2)
#xymax = np.max( [np.max(np.fabs(data1)), np.max(np.fabs(data2))] )
#lim = (int(xymax/binwidth) + 1) * binwidth
binsx = np.arange(xmin, xmax + xbinwidth, xbinwidth)
binsy = np.arange(ymin, ymax + ybinwidth, ybinwidth)
#bins = np.arange(-lim, lim + binwidth, binwidth)
# the scatter plot:
if isinstance(data1b, np.ndarray): #TWO DATA ENTRIES
if SSE == None:
print '*args, **kwargs do not have any influcence when using two\
options'
axScatter.scatter(data1,
data2,
facecolor='none',
edgecolor='k',
s=25)
axScatter.scatter(data1b,
data2b,
facecolor='none',
edgecolor='grey',
s=25)
xminb = np.min(data1b)
xmaxb = np.max(data1b)
yminb = np.min(data2b)
ymaxb = np.max(data2b)
binsxb = np.arange(xminb, xmaxb + xbinwidth, xbinwidth)
binsyb = np.arange(yminb, ymaxb + ybinwidth, ybinwidth)
axHistx.hist(data1b,
bins=binsxb,
edgecolor='None',
color='grey',
normed=True)
axHisty.hist(data2b,
bins=binsyb,
orientation='horizontal',
edgecolor='None',
color='grey',
normed=True)
axHistx.hist(data1,
bins=binsx,
edgecolor='None',
color='k',
normed=True)
axHisty.hist(data2,
bins=binsy,
orientation='horizontal',
edgecolor='None',
color='k',
normed=True)
else:
print '*args, **kwargs do not have any influcence when using two\
options'
sc1 = axScatter.scatter(data1b,
data2b,
c=SSEb,
vmax=vmax,
alpha=roodlichter,
edgecolors='none',
cmap=colormaps,
*args,
**kwargs)
axScatter.scatter(data1,
data2,
c=SSE,
vmax=vmax,
edgecolors='none',
cmap=colormaps,
*args,
**kwargs)
xminb = np.min(data1b)
xmaxb = np.max(data1b)
yminb = np.min(data2b)
ymaxb = np.max(data2b)
binsxb = np.arange(xminb, xmaxb + xbinwidth, xbinwidth)
binsyb = np.arange(yminb, ymaxb + ybinwidth, ybinwidth)
axHistx.hist(data1b,
bins=binsxb,
edgecolor='None',
color=colormaps(1.),
normed=True)
axHisty.hist(data2b,
bins=binsyb,
orientation='horizontal',
color=colormaps(1.),
edgecolor='None',
normed=True)
axHistx.hist(data1,
bins=binsx,
edgecolor='None',
color=colormaps(0.),
normed=True)
axHisty.hist(data2,
bins=binsy,
orientation='horizontal',
edgecolor='None',
color=colormaps(0.),
normed=True)
else: #ONLY ONE DATA1 and DATA2
if SSE == None:
axScatter.scatter(data1, data2, c='black', *args, **kwargs)
axHistx.hist(data1, bins=binsx, edgecolor='None', color='k')
axHisty.hist(data2,
bins=binsy,
orientation='horizontal',
edgecolor='None',
color='k')
else:
axScatter.scatter(data1,
data2,
c=SSE,
vmax=vmax,
edgecolors='none',
cmap=colormaps,
*args,
**kwargs)
axHistx.hist(data1, bins=binsx, edgecolor='None', color='k')
axHisty.hist(data2,
bins=binsy,
orientation='horizontal',
edgecolor='None',
color='k')
# the xaxis of axHistx and yaxis of axHisty are shared with axScatter,
# thus there is no need to manually adjust the xlim and ylim of these
# axis.
majloc1 = MaxNLocator(nbins=4, prune='lower')
axScatter.yaxis.set_major_locator(majloc1)
majloc2 = MaxNLocator(nbins=4)
axScatter.xaxis.set_major_locator(majloc2)
axScatter.grid(linestyle='dashed', color='0.75', linewidth=1.)
axScatter.set_axisbelow(True)
axHisty.set_axisbelow(True)
axHistx.set_axisbelow(True)
# The 'clean' environment
if cleanstyle == True:
plt.setp(axHistx.get_xticklabels() + axHisty.get_yticklabels(),
visible=False)
plt.setp(axHistx.get_yticklabels() + axHisty.get_xticklabels(),
visible=False)
axHisty.set_xticks([])
axHistx.set_yticks([])
axHistx.xaxis.set_ticks_position('bottom')
axHisty.yaxis.set_ticks_position('left')
axHisty.spines['right'].set_color('none')
axHisty.spines['top'].set_color('none')
axHisty.spines['bottom'].set_color('none')
axHisty.spines['left'].set_color('none')
axHistx.spines['top'].set_color('none')
axHistx.spines['right'].set_color('none')
axHistx.spines['left'].set_color('none')
axHistx.spines['bottom'].set_color('none')
axScatter.spines['top'].set_color('none')
axScatter.spines['right'].set_color('none')
else:
plt.setp(axHistx.get_xticklabels() + axHisty.get_yticklabels(),
visible=False)
for tl in axHisty.get_yticklabels():
tl.set_visible(False)
for tlp in axHisty.get_xticklabels():
tlp.set_rotation(-90)
majloc3 = MaxNLocator(nbins=4, prune='lower')
axHistx.yaxis.set_major_locator(majloc3)
axHistx.yaxis.tick_right()
majloc4 = MaxNLocator(nbins=4, prune='lower')
axHisty.xaxis.set_major_locator(majloc4)
axHisty.xaxis.tick_top()
axHisty.yaxis.grid(linestyle='dashed', color='0.75', linewidth=1.)
axHistx.xaxis.grid(linestyle='dashed', color='0.75', linewidth=1.)
return fig, axScatter, axHistx, axHisty, sc1
def show(self, location='right', width=0.2, pad=0.05, ticks=None,
labels=True, log_format=False, box=None,
box_orientation='vertical', axis_label_text=None,
axis_label_rotation=None, axis_label_pad=5):
'''
Show a colorbar on the side of the image.
Parameters
----------
location : str, optional
Where to place the colorbar. Should be one of 'left', 'right', 'top', 'bottom'.
width : float, optional
The width of the colorbar relative to the canvas size.
pad : float, optional
The spacing between the colorbar and the image relative to the
canvas size.
ticks : list, optional
The position of the ticks on the colorbar.
labels : bool, optional
Whether to show numerical labels.
log_format : bool, optional
Whether to format ticks in exponential notation
box : list, optional
A custom box within which to place the colorbar. This should
be in the form [xmin, ymin, dx, dy] and be in relative figure
units. This overrides the location argument.
box_orientation str, optional
The orientation of the colorbar within the box. Can be
'horizontal' or 'vertical'
axis_label_text str, optional
Optional text label of the colorbar.
'''
self._base_settings['location'] = location
self._base_settings['width'] = width
self._base_settings['pad'] = pad
self._base_settings['ticks'] = ticks
self._base_settings['labels'] = labels
self._base_settings['log_format'] = log_format
self._base_settings['box'] = box
self._base_settings['box_orientation'] = box_orientation
self._base_settings['axis_label_text'] = axis_label_text
self._base_settings['axis_label_rotation'] = axis_label_rotation
self._base_settings['axis_label_pad'] = axis_label_pad
if self._parent.image:
if self._colorbar_axes:
self._parent._figure.delaxes(self._colorbar_axes)
if box is None:
divider = make_axes_locatable(self._parent.ax)
if location == 'right':
self._colorbar_axes = divider.new_horizontal(size=width, pad=pad, axes_class=maxes.Axes)
orientation = 'vertical'
elif location == 'top':
self._colorbar_axes = divider.new_vertical(size=width, pad=pad, axes_class=maxes.Axes)
orientation = 'horizontal'
elif location == 'left':
warnings.warn("Left colorbar not fully implemented")
self._colorbar_axes = divider.new_horizontal(size=width, pad=pad, pack_start=True, axes_class=maxes.Axes)
locator = divider.new_locator(nx=0, ny=0)
self._colorbar_axes.set_axes_locator(locator)
orientation = 'vertical'
elif location == 'bottom':
warnings.warn("Bottom colorbar not fully implemented")
self._colorbar_axes = divider.new_vertical(size=width, pad=pad, pack_start=True, axes_class=maxes.Axes)
locator = divider.new_locator(nx=0, ny=0)
self._colorbar_axes.set_axes_locator(locator)
orientation = 'horizontal'
else:
raise Exception("location should be one of: right/top")
self._parent._figure.add_axes(self._colorbar_axes)
else:
self._colorbar_axes = self._parent._figure.add_axes(box)
orientation = box_orientation
if log_format:
format = LogFormatterMathtext()
else:
format = None
self._colorbar = self._parent._figure.colorbar(self._parent.image, cax=self._colorbar_axes,
orientation=orientation, format=format,
ticks=ticks)
if axis_label_text:
if axis_label_rotation:
self._colorbar.set_label(axis_label_text, rotation=axis_label_rotation)
else:
self._colorbar.set_label(axis_label_text)
if location == 'right':
for tick in self._colorbar_axes.yaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = False
tick.label2On = labels
self._colorbar_axes.yaxis.set_label_position('right')
self._colorbar_axes.yaxis.labelpad = axis_label_pad
elif location == 'top':
for tick in self._colorbar_axes.xaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = False
tick.label2On = labels
self._colorbar_axes.xaxis.set_label_position('top')
self._colorbar_axes.xaxis.labelpad = axis_label_pad
elif location == 'left':
for tick in self._colorbar_axes.yaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = labels
tick.label2On = False
self._colorbar_axes.yaxis.set_label_position('left')
self._colorbar_axes.yaxis.labelpad = axis_label_pad
elif location == 'bottom':
for tick in self._colorbar_axes.xaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = labels
tick.label2On = False
self._colorbar_axes.xaxis.set_label_position('bottom')
self._colorbar_axes.xaxis.labelpad = axis_label_pad
else:
warnings.warn("No image is shown, therefore, no colorbar will be plotted")
def save_spa(odir, filename, stk, anlz, png, pdf, svg):
SaveFileName = os.path.join(odir,filename)
from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
matplotlib.rcParams['pdf.fonttype'] = 42
colors=['#FF0000','#000000','#FFFFFF']
spanclrmap=matplotlib.colors.LinearSegmentedColormap.from_list('spancm',colors)
for i in range (anlz.n_target_spectra):
#Save composition maps
if anlz.pca_calculated == 0:
tsmapimage = anlz.target_svd_maps[:,:,i]
else:
tsmapimage = anlz.target_pcafit_maps[:,:,i]
fig = matplotlib.figure.Figure(figsize =(PlotH, PlotH))
canvas = FigureCanvas(fig)
fig.clf()
axes = fig.gca()
divider = make_axes_locatable(axes)
ax_cb = divider.new_horizontal(size="3%", pad=0.03)
fig.add_axes(ax_cb)
axes.set_position([0.03,0.03,0.8,0.94])
min_val = np.min(tsmapimage)
max_val = np.max(tsmapimage)
bound = np.max((np.abs(min_val), np.abs(max_val)))
im = axes.imshow(tsmapimage, cmap=spanclrmap, vmin = -bound, vmax = bound)
cbar = axes.figure.colorbar(im, orientation='vertical',cax=ax_cb)
axes.axis("off")
if png == 1:
ext = 'png'
fileName_img = SaveFileName+"_TSmap_" +str(i+1)+"."+ext
fig.savefig(fileName_img, bbox_inches='tight', pad_inches = 0.0)
if pdf == 1:
ext = 'pdf'
fileName_img = SaveFileName+"_TSmap_" +str(i+1)+"."+ext
fig.savefig(fileName_img, bbox_inches='tight', pad_inches = 0.0)
if svg == 1:
ext = 'svg'
fileName_img = SaveFileName+"_TSmap_" +str(i+1)+"."+ext
fig.savefig(fileName_img, bbox_inches='tight', pad_inches = 0.0)
#Save spectra
for i in range (anlz.n_target_spectra):
tspectrum = anlz.target_spectra[i, :]
fig = matplotlib.figure.Figure(figsize =(PlotW, PlotH))
canvas = FigureCanvas(fig)
fig.clf()
fig.add_axes((0.15,0.15,0.75,0.75))
axes = fig.gca()
line1 = axes.plot(stk.ev,tspectrum, color='black', label = 'Raw data')
if anlz.pca_calculated == 1:
tspectrumfit = anlz.target_pcafit_spectra[i, :]
diff = np.abs(tspectrum-tspectrumfit)
line2 = axes.plot(stk.ev,tspectrumfit, color='green', label = 'Fit')
line3 = axes.plot(stk.ev,diff, color='grey', label = 'Abs(Raw-Fit)')
fontP = matplotlib.font_manager.FontProperties()
fontP.set_size('small')
axes.legend(loc=4, prop = fontP)
axes.set_xlabel('Photon Energy [eV]')
axes.set_ylabel('Optical Density')
if png == 1:
ext = 'png'
fileName_spec = SaveFileName+"_Tspectrum_" +str(i+1)+"."+ext
fig.savefig(fileName_spec)
if pdf == 1:
ext = 'pdf'
fileName_spec = SaveFileName+"_Tspectrum_" +str(i+1)+"."+ext
fig.savefig(fileName_spec)
if svg == 1:
ext = 'svg'
fileName_spec = SaveFileName+"_Tspectrum_" +str(i+1)+"."+ext
fig.savefig(fileName_spec)
fileName_spec = SaveFileName+"_Tspectrum_" +str(i+1)+".csv"
cname = "Tspectrum_" +str(i+1)
stk.write_csv(fileName_spec, stk.ev, tspectrum, cname = cname)
return
def makePlot(in_m, channel_names=None, fig=None, x_tick_rot=0, size=None,
cmap=plt.cm.RdBu_r, colorbar=True, color_anchor=None, title=None, max_val=None, min_val=None):
N = in_m.shape[0]
ind = np.arange(N) # the evenly spaced plot indices
def channel_formatter(x, pos=None):
thisind = np.clip(int(x), 0, N - 1)
return channel_names[thisind]
if fig is None:
fig=plt.figure()
if size is not None:
fig.set_figwidth(size[0])
fig.set_figheight(size[1])
wid=fig.get_figwidth()
ht=fig.get_figheight()
ax_im = fig.add_subplot(1, 1, 1)
#If you want to draw the colorbar:
# what is make_axes_locatable?
divider = make_axes_locatable(ax_im)
ax_cb = divider.new_vertical(size="20%", pad=0.2, pack_start=True)
fig.add_axes(ax_cb)
#Make a copy of the input, so that you don't make changes to the original
#data provided
m = in_m.copy()
#Null the upper triangle, so that you don't get the redundant and
#the diagonal values:
#idx_null = triu_indices(m.shape[0])
#m[idx_null] = np.nan
#Extract the minimum and maximum values for scaling of the
#colormap/colorbar:
if max_val is None:
max_val = np.nanmax(m)
min_val = np.nanmin(m)
if color_anchor is None:
color_min = min_val
color_max = max_val
elif color_anchor == 0:
bound = max(abs(max_val), abs(min_val))
color_min = -bound
color_max = bound
else:
color_min = color_anchor[0]
color_max = color_anchor[1]
#The call to imshow produces the matrix plot:
im = ax_im.imshow(m, origin='upper', interpolation='nearest',
vmin=color_min, vmax=color_max, cmap=cmap)
#Formatting:
ax = ax_im
ax.grid(True)
#Label each of the cells with the row and the column:
if channel_names is not None:
for i in xrange(0, m.shape[0]):
if i < (m.shape[0] - 1):
ax.text(i - 0.3, i, channel_names[i], rotation=x_tick_rot)
if i > 0:
ax.text(-1, i + 0.3, channel_names[i], horizontalalignment='right')
ax.set_axis_off()
ax.set_xticks(np.arange(N))
ax.xaxis.set_major_formatter(ticker.FuncFormatter(channel_formatter))
fig.autofmt_xdate(rotation=x_tick_rot)
ax.set_yticks(np.arange(N))
ax.set_yticklabels(channel_names)
ax.set_ybound([-0.5, N - 0.5])
ax.set_xbound([-0.5, N - 1.5])
#Make the tick-marks invisible:
for line in ax.xaxis.get_ticklines():
line.set_markeredgewidth(0)
for line in ax.yaxis.get_ticklines():
line.set_markeredgewidth(0)
ax.set_axis_off()
if title is not None:
ax.set_title(title)
#The following produces the colorbar and sets the ticks
if colorbar:
#Set the ticks - if 0 is in the interval of values, set that, as well
#as the maximal and minimal values:
if min_val < 0:
ticks = [min_val, 0, max_val]
#Otherwise - only set the minimal and maximal value:
else:
ticks = [min_val, max_val]
#This makes the colorbar:
cb = fig.colorbar(im, cax=ax_cb, orientation='horizontal',
cmap=cmap,
norm=im.norm,
boundaries=np.linspace(min_val, max_val, 256),
ticks=ticks,
format='%.2f')
# Set the current figure active axis to be the top-one, which is the one
# most likely to be operated on by users later on
fig.sca(ax)
return fig
def draw_graph(G,
labels=None,
node_colors=None,
node_shapes=None,
node_scale=1.0,
edge_style='solid',
edge_cmap=None,
colorbar=False,
vrange=None,
layout=None,
title=None,
font_family='sans-serif',
font_size=9,
stretch_factor=1.0,
edge_alpha=True,
fig_size=None):
"""Draw a weighted graph with options to visualize link weights.
The resulting diagram uses the rank of each node as its size, and the
weight of each link (after discarding thresholded values, see below) as the
link opacity.
It maps edge weight to color as well as line opacity and thickness,
allowing the color part to be hardcoded over a value range (to permit valid
cross-figure comparisons for different graphs, so the same color
corresponds to the same link weight even if each graph has a different
range of weights). The nodes sizes are proportional to their degree,
computed as the sum of the weights of all their links. The layout defaults
to circular, but any nx layout function can be passed in, as well as a
statically precomputed layout.
Parameters
----------
G : weighted graph
The values must be of the form (v1,v2), with all v2 in [0,1]. v1 are
used for colors, v2 for thickness/opacity.
labels : list or dict, optional.
An indexable object that maps nodes to strings. If not given, the
string form of each node is used as a label. If False, no labels are
drawn.
node_colors : list or dict, optional.
An indexable object that maps nodes to valid matplotlib color specs. See
matplotlib's plot() function for details.
node_shapes : list or dict, optional.
An indexable object that maps nodes to valid matplotlib shape specs. See
matplotlib's scatter() function for details. If not given, circles are
used.
node_scale : float, optional
A scale factor to globally stretch or shrink all nodes symbols by.
edge_style : string, optional
Line style for the edges, defaults to 'solid'.
edge_cmap : matplotlib colormap, optional.
A callable that returns valid color specs, like matplotlib colormaps.
If not given, edges are colored black.
colorbar : bool
If true, automatically add a colorbar showing the mapping of graph weight
values to colors.
vrange : pair of floats
If given, this indicates the total range of values that the weights can
in principle occupy, and is used to set the lower/upper range of the
colormap. This allows you to set the range of multiple different figures
to the same values, even if each individual graph has range variations,
so that visual color comparisons across figures are valid.
layout : function or layout dict, optional
A NetworkX-like layout function or the result of a precomputed layout for
the given graph. NetworkX produces layouts as dicts keyed by nodes and
with (x,y) pairs of coordinates as values, any function that produces
this kind of output is acceptable. Defaults to nx.circular_layout.
title : string, optional.
If given, title to put on the main plot.
font_family : string, optional.
Font family used for the node labels and title.
font_size : int, optional.
Font size used for the node labels and title.
stretch_factor : float, optional
A global scaling factor to make the graph larger (or smaller if <1).
This can be used to separate the nodes if they start overlapping.
edge_alpha: bool, optional
Whether to weight the transparency of each edge by a factor equivalent to
its relative weight
fig_size: list of height by width, the size of the figure (in
inches). Defaults to [6,6]
Returns
-------
fig
The matplotlib figure object with the plot.
"""
if fig_size is None:
figsize = [6, 6]
scaler = figsize[0] / 6.
# For the size of the node symbols
node_size_base = 1000 * scaler
node_min_size = 200 * scaler
default_node_shape = 'o'
# Default colors if none given
default_node_color = 'r'
default_edge_color = 'k'
# Max edge width
max_width = 13 * scaler
min_width = 2 * scaler
font_family = 'sans-serif'
# We'll use the nodes a lot, let's make a numpy array of them
nodes = np.array(sorted(G.nodes()))
nnod = len(nodes)
# Build a 'weighted degree' array obtained by adding the (absolute value)
# of the weights for all edges pointing to each node:
amat = nx.adj_matrix(G).A # get a normal array out of it
degarr = abs(amat).sum(0) # weights are sums across rows
# Map the degree to the 0-1 range so we can use it for sizing the nodes.
try:
odegree = rescale_arr(degarr, 0, 1)
# Make an array of node sizes based on node degree
node_sizes = odegree * node_size_base + node_min_size
except ZeroDivisionError:
# All nodes same size
node_sizes = np.empty(nnod, float)
node_sizes.fill(0.5 * node_size_base + node_min_size)
# Adjust node size list. We square the scale factor because in mpl, node
# sizes represent area, not linear size, but it's more intuitive for the
# user to think of linear factors (the overall figure scale factor is also
# linear).
node_sizes *= node_scale ** 2
# Set default node properties
if node_colors is None:
node_colors = [default_node_color] * nnod
if node_shapes is None:
node_shapes = [default_node_shape] * nnod
# Set default edge colormap
if edge_cmap is None:
# Make an object with the colormap API, that maps all input values to
# the default color (with proper alhpa)
edge_cmap = (lambda val, alpha:
colors.colorConverter.to_rgba(default_edge_color, alpha))
# if vrange is None, we set the color range from the values, else the user
# can specify it
# e[2] is edge value: edges_iter returns (i,j,data)
gvals = np.array([e[2]['weight'] for e in G.edges(data=True)])
gvmin, gvmax = gvals.min(), gvals.max()
gvrange = gvmax - gvmin
if vrange is None:
vrange = gvmin, gvmax
# Now, construct the normalization for the colormap
cnorm = mpl.colors.Normalize(vmin=vrange[0], vmax=vrange[1])
# Create the actual plot where the graph will be displayed
figsize = np.array(figsize, float)
figsize *= stretch_factor
fig = plt.figure(figsize=figsize)
ax_graph = fig.add_subplot(1, 1, 1)
fig.sca(ax_graph)
if layout is None:
layout = nx.circular_layout
# Compute positions for all nodes - nx has several algorithms
if callable(layout):
pos = layout(G)
else:
# The user can also provide a precomputed layout
pos = layout
# Draw nodes
for nod in nodes:
nx.draw_networkx_nodes(G,
pos,
nodelist=[nod],
node_color=node_colors[nod],
node_shape=node_shapes[nod],
node_size=node_sizes[nod])
# Draw edges
if not isinstance(G, nx.DiGraph):
# Undirected graph, simple lines for edges
# We need the size of the value range to properly scale colors
vsize = vrange[1] - vrange[0]
gvals_normalized = G.metadata['vals_norm']
for (u, v, y) in G.edges(data=True):
# The graph value is the weight, and the normalized values are in
# [0,1], used for thickness/transparency
alpha = gvals_normalized[u, v]
# Scale the color choice to the specified vrange, so that
ecol = (y['weight'] - vrange[0]) / vsize
#print 'u,v:',u,v,'y:',y,'ecol:',ecol # dbg
if edge_alpha:
fade = alpha
else:
fade = 1.0
edge_color = [tuple(edge_cmap(ecol, fade))]
#dbg:
#print u,v,y
draw_networkx_edges(G,
pos,
edgelist=[(u, v)],
width=min_width + alpha * max_width,
edge_color=edge_color,
style=edge_style)
else:
# Directed graph, use arrows.
# XXX - this is currently broken.
raise NotImplementedError("arrow drawing currently broken")
## for (u,v,x) in G.edges(data=True):
## y,w = x
## draw_arrows(G,pos,edgelist=[(u,v)],
## edge_color=[w],
## alpha=w,
## edge_cmap=edge_cmap,
## width=w*max_width)
# Draw labels. If not given, we use the string form of the nodes. If
# labels is False, no labels are drawn.
if labels is None:
labels = map(str, nodes)
if labels:
lab_idx = range(len(labels))
labels_dict = dict(zip(lab_idx, labels))
nx.draw_networkx_labels(G,
pos,
labels_dict,
font_size=font_size,
font_family=font_family)
if title:
plt.title(title, fontsize=font_size)
# Turn off x and y axes labels in pylab
plt.xticks([])
plt.yticks([])
# Add a colorbar if requested
if colorbar:
divider = make_axes_locatable(ax_graph)
ax_cb = divider.new_vertical(size="20%", pad=0.2, pack_start=True)
fig.add_axes(ax_cb)
cb = mpl.colorbar.ColorbarBase(ax_cb,
cmap=edge_cmap,
norm=cnorm,
#boundaries = np.linspace(min((gvmin,0)),
# max((gvmax,0)),
# 256),
orientation='horizontal',
format='%.2f')
# Always return the MPL figure object so the user can further manipulate it
return fig
def plot_tri_matrix(mat,
figure=None,
num='plot_part_of_this_matrix',
size=None,
cmap=pyplot.cm.RdBu_r,
colourbar=True,
color_anchor=None,
node_labels=None,
x_tick_rot=0,
title=None):
r"""Creates a lower-triangle of a square matrix. Very often found to display correlations or coherence.
Parameters
----------
mat : square matrix
node_labels : list of strings with the labels to be applied to
the nodes. Defaults to '0','1','2', etc.
fig : a matplotlib figure
cmap : a matplotlib colormap.
title : figure title (eg '$\alpha$')
color_anchor : determines the clipping for the colormap.
If None, the data min, max are used.
If 0, min and max of colormap correspond to max abs(mat)
If (a,b), min and max are set accordingly (a,b)
Returns
-------
fig: a figure object
"""
def channel_formatter(x, pos=None):
thisidx = numpy.clip(int(x), 0, N - 1)
return node_labels[thisidx]
if figure is not None:
fig = figure
else:
if num is None:
fig = pyplot.figure()
else:
fig = pyplot.figure(num=num)
if size is not None:
fig.set_figwidth(size[0])
fig.set_figheight(size[1])
w = fig.get_figwidth()
h = fig.get_figheight()
ax_im = fig.add_subplot(1, 1, 1)
N = mat.shape[0]
idx = numpy.arange(N)
if colourbar:
if IMPORTED_MPL_TOOLKITS:
divider = make_axes_locatable(ax_im)
ax_cb = divider.new_vertical(size="10%", pad=0.1, pack_start=True)
fig.add_axes(ax_cb)
else:
pass
mat_copy = mat.copy()
#Null the upper triangle, including the main diagonal.
idx_null = numpy.triu_indices(mat_copy.shape[0])
mat_copy[idx_null] = numpy.nan
#Min max values
max_val = numpy.nanmax(mat_copy)
min_val = numpy.nanmin(mat_copy)
if color_anchor is None:
color_min = min_val
color_max = max_val
elif color_anchor == 0:
bound = max(abs(max_val), abs(min_val))
color_min = -bound
color_max = bound
else:
color_min = color_anchor[0]
color_max = color_anchor[1]
#The call to imshow produces the matrix plot:
im = ax_im.imshow(mat_copy,
origin='upper',
interpolation='nearest',
vmin=color_min,
vmax=color_max,
cmap=cmap)
#Formatting:
ax = ax_im
ax.grid(True)
#Label each of the cells with the row and the column:
if node_labels is not None:
for i in range(0, mat_copy.shape[0]):
if i < (mat_copy.shape[0] - 1):
ax.text(i - 0.3, i, node_labels[i], rotation=x_tick_rot)
if i > 0:
ax.text(-1,
i + 0.3,
node_labels[i],
horizontalalignment='right')
ax.set_axis_off()
ax.set_xticks(numpy.arange(N))
ax.xaxis.set_major_formatter(ticker.FuncFormatter(channel_formatter))
fig.autofmt_xdate(rotation=x_tick_rot)
ax.set_yticks(numpy.arange(N))
ax.set_yticklabels(node_labels)
ax.set_ybound([-0.5, N - 0.5])
ax.set_xbound([-0.5, N - 1.5])
#Make the tick-marks invisible:
for line in ax.xaxis.get_ticklines():
line.set_markeredgewidth(0)
for line in ax.yaxis.get_ticklines():
line.set_markeredgewidth(0)
ax.set_axis_off()
if title is not None:
ax.set_title(title)
if colourbar:
#Set the ticks - if 0 is in the interval of values, set that, as well
#as the min, max values:
if min_val < 0:
ticks = [color_min, min_val, 0, max_val, color_max]
#set the min, mid and max values:
else:
ticks = [
color_min, min_val, (color_max - color_min) / 2., max_val,
color_max
]
#colourbar:
if IMPORTED_MPL_TOOLKITS:
cb = fig.colorbar(im,
cax=ax_cb,
orientation='horizontal',
cmap=cmap,
norm=im.norm,
boundaries=numpy.linspace(
color_min, color_max, 256),
ticks=ticks,
format='%.2f')
else:
# the colourbar will be wider than the matrix
cb = fig.colorbar(im,
orientation='horizontal',
cmap=cmap,
norm=im.norm,
boundaries=numpy.linspace(
color_min, color_max, 256),
ticks=ticks,
format='%.2f')
fig.sca(ax)
return fig
def show(self,
location='right',
width=0.2,
pad=0.05,
ticks=None,
labels=True,
box=None,
box_orientation='vertical'):
'''
Show a colorbar on the side of the image.
Optional Keyword Arguments:
*location*: [ string ]
Where to place the colorbar. Should be one of 'left', 'right', 'top', 'bottom'.
*width*: [ float ]
The width of the colorbar relative to the canvas size.
*pad*: [ float ]
The spacing between the colorbar and the image relative to the canvas size.
*ticks*: [ None or list ]
The position of the ticks on the colorbar.
*labels*: [ True or False ]
Whether to show numerical labels.
*box*: [ list ]
A custom box within which to place the colorbar. This should
be in the form [xmin, ymin, dx, dy] and be in relative figure
units. This overrides the location argument.
*box_orientation* [ str ]
The orientation of the colorbar within the box. Can be
'horizontal' or 'vertical'
'''
self._base_settings['location'] = location
self._base_settings['width'] = width
self._base_settings['pad'] = pad
self._base_settings['ticks'] = ticks
self._base_settings['labels'] = labels
self._base_settings['box'] = box
self._base_settings['box_orientation'] = box_orientation
if self._parent.image:
if self._colorbar_axes:
self._parent._figure.delaxes(self._colorbar_axes)
if box is None:
divider = make_axes_locatable(self._parent._ax1)
if location == 'right':
self._colorbar_axes = divider.new_horizontal(
size=width, pad=pad, axes_class=maxes.Axes)
orientation = 'vertical'
elif location == 'top':
self._colorbar_axes = divider.new_vertical(
size=width, pad=pad, axes_class=maxes.Axes)
orientation = 'horizontal'
elif location == 'left':
warnings.warn("Left colorbar not fully implemented")
self._colorbar_axes = divider.new_horizontal(
size=width,
pad=pad,
pack_start=True,
axes_class=maxes.Axes)
locator = divider.new_locator(nx=0, ny=0)
self._colorbar_axes.set_axes_locator(locator)
orientation = 'vertical'
elif location == 'bottom':
warnings.warn("Bottom colorbar not fully implemented")
self._colorbar_axes = divider.new_vertical(
size=width,
pad=pad,
pack_start=True,
axes_class=maxes.Axes)
locator = divider.new_locator(nx=0, ny=0)
self._colorbar_axes.set_axes_locator(locator)
orientation = 'horizontal'
else:
raise Exception("location should be one of: right/top")
self._parent._figure.add_axes(self._colorbar_axes)
else:
self._colorbar_axes = self._parent._figure.add_axes(box)
orientation = box_orientation
self._colorbar = self._parent._figure.colorbar(
self._parent.image,
cax=self._colorbar_axes,
orientation=orientation,
ticks=ticks)
if location == 'right':
for tick in self._colorbar_axes.yaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = False
tick.label2On = labels
elif location == 'top':
for tick in self._colorbar_axes.xaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = False
tick.label2On = labels
elif location == 'left':
for tick in self._colorbar_axes.yaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = labels
tick.label2On = False
elif location == 'bottom':
for tick in self._colorbar_axes.xaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = labels
tick.label2On = False
else:
warnings.warn(
"No image is shown, therefore, no colorbar will be plotted")
def map_spatial_analog(ncfile, variable='dissimilarity', cmap='viridis', title='Spatial analog'):
"""Return a matplotlib Figure instance showing a map of the dissimilarity measure.
"""
import netCDF4 as nc
from flyingpigeon import utils
from mpl_toolkits.axes_grid import make_axes_locatable
import matplotlib.axes as maxes
try:
var = utils.get_values(ncfile, variable)
LOGGER.info('Data loaded')
lats, lons = utils.get_coordinates(ncfile, variable=variable, unrotate=False)
if len(lats.shape) == 1:
cyclic_var, cyclic_lons = add_cyclic_point(var, coord=lons)
lons = cyclic_lons.data
var = cyclic_var
with nc.Dataset(ncfile) as D:
V = D.variables[variable]
lon, lat = map(float, V.target_location.split(','))
LOGGER.info('Lat and lon loaded')
except Exception as e:
msg = 'Failed to get data for plotting: {0}\n{1}'.format(ncfile, e)
LOGGER.exception(msg)
raise Exception(msg)
try:
fig = plt.figure(facecolor='w', edgecolor='k')
fig.subplots_adjust(top=.95, bottom=.05, left=.03, right=.95)
ax = plt.axes(
projection=ccrs.Robinson(central_longitude=int(np.mean(lons))))
divider = make_axes_locatable(ax)
cax = divider.new_horizontal("4%", pad=0.15, axes_class=maxes.Axes)
fig.add_axes(cax)
ax.plot(lon, lat, marker='o', mfc='#292421', ms=13, transform=ccrs.PlateCarree())
ax.plot(lon, lat, marker='o', mfc='#ffffff', ms=7, transform=ccrs.PlateCarree())
cs = ax.contourf(lons, lats, var, 60,
transform=ccrs.PlateCarree(),
cmap=cmap, interpolation='nearest')
ax.coastlines(color='k', linewidth=.8)
ax.set_title(title)
cb = plt.colorbar(cs, cax=cax, orientation='vertical')
cb.set_label(u"– Dissimilarity +") # ha='left', va='center')
cb.set_ticks([])
except:
msg = 'failed to plot graphic'
LOGGER.exception(msg)
LOGGER.info('Plot created and figure saved')
return fig
from mpl_toolkits.axes_grid import make_axes_locatable
import matplotlib.axes as maxes
from matplotlib import cm
if __name__ == '__main__':
fig = P.figure(figsize=(10, 10))
ax1 = fig.add_subplot(221)
ax2 = fig.add_subplot(222)
ax3 = fig.add_subplot(223)
ax4 = fig.add_subplot(224)
s1 = ax1.scatter(numpy.random.rand(10),
numpy.random.rand(10),
c=numpy.random.rand(10))
divider = make_axes_locatable(ax1)
cax1 = divider.new_horizontal('5%', pad=0.0, axes_class=maxes.Axes)
fig.add_axes(cax1)
c1 = fig.colorbar(s1, cax=cax1, orientation='horizontal')
s2 = ax2.scatter(numpy.random.rand(10),
numpy.random.rand(10),
c=numpy.random.rand(10),
cmap=cm.get_cmap('jet'))
divider = make_axes_locatable(ax2)
cax2 = divider.append_axes('right', 0.1, pad=0.1)
c2 = fig.colorbar(s2, cax=cax2)
#p = matplotlib.patches.Patch(color=cm.get_cmap('jet'))
#ax2.legend([p],['Test'])
s3 = ax3.scatter(numpy.random.rand(10),
of.close() f.close() data = np.loadtxt(ofName) os.unlink(ofName) x = data[:, xcol - 1] y = data[:, ycol - 1] fig = plt.figure(1, figsize=(8.5, 8.5)) from mpl_toolkits.axes_grid import make_axes_locatable axScatter = plt.subplot(111) divider = make_axes_locatable(axScatter) # create a new axes with a height of 1.2 inch above the axScatter axHistx = divider.new_vertical(1.2, pad=0.1, sharex=axScatter) # create a new axes with a width of 1.2 inch on the right side of the # axScatter axHisty = divider.new_horizontal(1.2, pad=0.1, sharey=axScatter) fig.add_axes(axHistx) fig.add_axes(axHisty) # make some labels invisible plt.setp(axHistx.get_xticklabels() + axHisty.get_yticklabels(), visible=False) # the scatter plot:
def plot(self, axes):
''' Plot network on given axes
:Parameters:
`axes` : matploltlib Axes
axes where to draw network
'''
classname = self.__class__.__name__
# Plot samples
axes.scatter(self.samples[:,0], self.samples[:,1], s=1.0, color='b', alpha=0.25)
fig = plt.gcf()
divider = make_axes_locatable(axes)
# Plot network
C = self.codebook
Cx,Cy = C[...,0], C[...,1]
if classname != 'NG':
for i in range(C.shape[0]):
axes.plot (Cx[i,:], Cy[i,:], 'k', alpha=0.85, lw=1.5)
for i in range(C.shape[1]):
axes.plot (Cx[:,i], Cy[:,i], 'k', alpha=0.85, lw=1.5)
axes.scatter (Cx.flatten(), Cy.flatten(), s=50, c= 'w', edgecolors='k', zorder=10)
axes.axis([0,1,0,1])
axes.set_xticks([])
axes.set_yticks([])
axes.set_aspect(1)
# Plot distortion
subaxes = divider.new_vertical(1.0, pad=0.4, sharex=axes)
fig.add_axes(subaxes)
subaxes.set_xticks([])
subaxes.yaxis.set_major_locator(matplotlib.ticker.MaxNLocator(2))
subaxes.yaxis.set_ticks_position('right')
subaxes.set_ylabel('Distortion')
subaxes.set_xlabel('Time')
#subaxes.axis([0,1,0,1])
Y = self.distortion[::1]
X = np.arange(len(Y))/float(len(Y)-1)
subaxes.plot(X,Y)
axes.axis([0,1,0,1])
if classname == 'NG':
plt.title('Neural Gas', fontsize=20)
elif classname == 'SOM':
plt.title('Self-Organizing Map', fontsize=20)
elif classname == 'DSOM':
plt.title('Dynamic Self-Organizing Map', fontsize=20)
if classname == 'NG':
axes.text(0.5, -0.01,
r'$\lambda_i = %.3f,\lambda_f = %.3f, \varepsilon_i=%.3f, \varepsilon_f=%.3f$' % (
self.sigma_i, self.sigma_f, self.lrate_i, self.lrate_f),
fontsize=16,
horizontalalignment='center',
verticalalignment='top',
transform = axes.transAxes)
if classname == 'SOM':
axes.text(0.5, -0.01,
r'$\sigma_i = %.3f,\sigma_f = %.3f, \varepsilon_i=%.3f, \varepsilon_f=%.3f$' % (
self.sigma_i, self.sigma_f, self.lrate_i, self.lrate_f),
fontsize=16,
horizontalalignment='center',
verticalalignment='top',
transform = axes.transAxes)
elif classname == 'DSOM':
axes.text(0.5, -0.01,
r'$elasticity = %.2f$, $\varepsilon = %.3f$' % (self.elasticity, self.lrate),
fontsize=16,
horizontalalignment='center',
verticalalignment='top',
transform = axes.transAxes)
def plot(self, statsResults):
'''
Create extended error bar plot.
filename: name of output file
data: matrix of data for each feature of interest; [feature, pValue, effectSize, lowerCI, upperCI, seq1, seq1, power]
oneMinusAlpha: 1-alpha value used to create above data matrix
'''
# *** Check if there is sufficient data to generate the plot
if len(statsResults.activeData) <= 0:
self.emptyAxis()
return
features = statsResults.getColumn('Features')
if len(features) > 200:
QtGui.QApplication.instance().setOverrideCursor(
QtGui.QCursor(QtCore.Qt.ArrowCursor))
reply = QtGui.QMessageBox.question(
self, 'Continue?',
'Profile contains ' + str(len(features)) + ' features. ' +
'It may take several seconds to generate this plot. We recommend filtering your profile first. '
+ 'Do you wish to continue?', QtGui.QMessageBox.Yes,
QtGui.QMessageBox.No)
QtGui.QApplication.instance().restoreOverrideCursor()
if reply == QtGui.QMessageBox.No:
self.emptyAxis()
return
# *** Colour of plot elements
profile1Colour = str(self.preferences['Sample 1 colour'].name())
profile2Colour = str(self.preferences['Sample 2 colour'].name())
# *** Colour of plot elements
highlightColor = (0.9, 0.9, 0.9)
orangeColour = (1, 0.5, 0)
# *** Sort data
if self.sortingField == 'p-values':
statsResults.activeData = TableHelper.SortTable(statsResults.activeData,\
[statsResults.dataHeadings['pValues']], False)
elif self.sortingField == 'Effect sizes':
statsResults.activeData = TableHelper.SortTable(statsResults.activeData,\
[statsResults.dataHeadings['EffectSize']],
True, True, statsResults.confIntervMethod.bRatio)
elif self.sortingField == 'Feature labels':
statsResults.activeData = TableHelper.SortTableStrCol(statsResults.activeData,\
statsResults.dataHeadings['Features'], False)
# *** Create lists for each quantity of interest
features = statsResults.getColumn('Features')
if statsResults.multCompCorrection.method == 'False discovery rate':
pValueTitle = 'q-value'
else:
pValueTitle = 'p-value'
if self.bShowCorrectedPvalues:
pValueLabels = statsResults.getColumnAsStr('pValuesCorrected')
pValueTitle += ' (corrected)'
else:
pValueLabels = statsResults.getColumnAsStr('pValues')
effectSizes = statsResults.getColumn('EffectSize')
lowerCIs = statsResults.getColumn('LowerCI')
upperCIs = statsResults.getColumn('UpperCI')
ciTitle = (
'%.3g' %
(statsResults.oneMinusAlpha() * 100)) + '% confidence intervals'
seqs1 = statsResults.getColumn('Seq1')
seqs2 = statsResults.getColumn('Seq2')
power = statsResults.getColumn('Power')
# *** Truncate feature labels
selectedFeatures = list(
self.preferences['Selected statistical features'])
if self.preferences['Truncate feature names']:
length = self.preferences['Length of truncated feature names']
for i in xrange(0, len(features)):
if len(features[i]) > length + 3:
features[i] = features[i][0:length] + '...'
for i in xrange(0, len(selectedFeatures)):
if len(selectedFeatures[i]) > length + 3:
selectedFeatures[i] = selectedFeatures[i][0:length] + '...'
# *** Check that there is at least one significant feature
if len(features) <= 0:
self.emptyAxis('No significant features')
return
# *** Adjust effect size for axis scale
dominateInSample2 = []
for i in xrange(0, len(effectSizes)):
seqs1[i] = -seqs1[i]
if statsResults.confIntervMethod.bRatio:
if effectSizes[i] < 1:
# mirror CI across y-axis
effectSizes[i] = 1.0 / effectSizes[i]
lowerCI = effectSizes[i] - (1.0 / upperCIs[i])
upperCI = (1.0 / lowerCIs[i]) - effectSizes[i]
lowerCIs[i] = lowerCI
upperCIs[i] = upperCI
dominateInSample2.append(i)
else:
lowerCIs[i] = effectSizes[i] - lowerCIs[i]
upperCIs[i] = upperCIs[i] - effectSizes[i]
else:
lowerCIs[i] = effectSizes[i] - lowerCIs[i]
upperCIs[i] = upperCIs[i] - effectSizes[i]
if effectSizes[i] < 0.0:
dominateInSample2.append(i)
# *** Determine which axes should be created
bShowPowerPlot = self.bShowPowerPlot and len(
power) != 0 and not math.isnan(power[0])
# *** Set figure size
plotHeight = self.figHeightPerRow * len(features)
self.imageWidth = self.figWidth
self.imageHeight = plotHeight + 0.65 # 0.65 inches for bottom and top labels
if self.imageWidth > 256 or self.imageHeight > 256:
QtGui.QApplication.instance().setOverrideCursor(
QtGui.QCursor(QtCore.Qt.ArrowCursor))
self.emptyAxis()
reply = QtGui.QMessageBox.question(
self, 'Excessively large plot',
'The resulting plot is too large to display.')
QtGui.QApplication.instance().restoreOverrideCursor()
return
self.fig.set_size_inches(self.imageWidth, self.imageHeight)
# *** Determine width of y-axis labels
yLabelBounds = self.yLabelExtents(features, 8)
# *** Size plots which comprise the extended errorbar plot
self.fig.clear()
spacingBetweenPlots = 0.3 # inches
heightBottomLabels = 0.4 # inches
totalSpacingBetweenPlots = spacingBetweenPlots * 2 # inches
widthNumSeqPlot = 1.25 # inches
if self.bShowSeqPlot == False:
widthNumSeqPlot = 0.0
totalSpacingBetweenPlots -= spacingBetweenPlots
widthPowerPlot = 0.75 # inches
if bShowPowerPlot == False:
widthPowerPlot = 0.0
totalSpacingBetweenPlots -= spacingBetweenPlots
widthPvalueLabels = 0.75 # inches
if self.bShowPValueLabels == False:
widthPvalueLabels = 0.1
yPlotOffsetFigSpace = heightBottomLabels / self.imageHeight
heightPlotFigSpace = plotHeight / self.imageHeight
xPlotOffsetFigSpace = yLabelBounds.width + 0.1 / self.imageWidth
pValueLabelWidthFigSpace = widthPvalueLabels / self.imageWidth
widthPlotFigSpace = 1.0 - pValueLabelWidthFigSpace - xPlotOffsetFigSpace
widthErrorBarPlot = widthPlotFigSpace * self.imageWidth - widthNumSeqPlot - widthPowerPlot - totalSpacingBetweenPlots
axInitAxis = self.fig.add_axes([
xPlotOffsetFigSpace, yPlotOffsetFigSpace, widthPlotFigSpace,
heightPlotFigSpace
])
divider = make_axes_locatable(axInitAxis)
divider.get_vertical()[0] = Size.Fixed(
len(features) * self.figHeightPerRow)
if self.bShowSeqPlot == True:
divider.get_horizontal()[0] = Size.Fixed(widthNumSeqPlot)
axErrorbar = divider.new_horizontal(widthErrorBarPlot,
pad=spacingBetweenPlots,
sharey=axInitAxis)
self.fig.add_axes(axErrorbar)
else:
divider.get_horizontal()[0] = Size.Fixed(widthErrorBarPlot)
axErrorbar = axInitAxis
if bShowPowerPlot == True:
axPower = divider.new_horizontal(widthPowerPlot,
pad=spacingBetweenPlots,
sharey=axInitAxis)
self.fig.add_axes(axPower)
# *** Plot of sequences for each subsystem
if self.bShowSeqPlot == True:
axNumSeq = axInitAxis
axNumSeq.hlines(np.arange(len(features)), [0],
seqs1,
lw=4,
color=profile1Colour,
zorder=10)
axNumSeq.hlines(np.arange(len(features)), [0],
seqs2,
lw=4,
color=profile2Colour,
zorder=10)
for value in np.arange(-0.5, len(features) - 1, 2):
axNumSeq.axhspan(value,
value + 1,
facecolor=highlightColor,
edgecolor='none',
zorder=1)
axNumSeq.vlines(0, -1, len(features), color='black', zorder=10)
axNumSeq.set_xlabel(self.xLabel, fontsize=8)
axNumSeq.set_xticks([min(seqs1), 0, max(seqs2)])
axNumSeq.set_xticklabels([-min(seqs1), 0, max(seqs2)])
axNumSeq.set_yticks(np.arange(len(features)))
axNumSeq.set_yticklabels(features, size=8)
axNumSeq.set_ylim([-1, len(features)])
for label in axNumSeq.get_xticklabels():
label.set_size(8)
for label in axNumSeq.get_yticklabels():
label.set_size(8)
if label.get_text() in selectedFeatures:
label.set_color('red')
for a in axNumSeq.yaxis.majorTicks:
a.tick1On = False
a.tick2On = False
for a in axNumSeq.xaxis.majorTicks:
a.tick1On = True
a.tick2On = False
for loc, spine in axNumSeq.spines.iteritems():
if loc in ['left', 'right', 'top']:
spine.set_color('none')
# *** Plot confidence intervals for each subsystem
lastAxes = axErrorbar
axErrorbar.errorbar(effectSizes,
np.arange(len(features)),
xerr=[lowerCIs, upperCIs],
fmt='o',
mfc=profile1Colour,
mec='black',
ecolor='black',
zorder=10)
effectSizesSample2 = [
effectSizes[value] for value in dominateInSample2
]
axErrorbar.plot(effectSizesSample2,
dominateInSample2,
ls='',
marker='o',
mfc=profile2Colour,
mec='black',
zorder=100)
if statsResults.confIntervMethod.bRatio:
axErrorbar.vlines(1,
-1,
len(features),
linestyle='dashed',
color=(0.25, 0.25, 0.25))
else:
axErrorbar.vlines(0,
-1,
len(features),
linestyle='dashed',
color=(0.25, 0.25, 0.25))
for value in np.arange(-0.5, len(features) - 1, 2):
axErrorbar.axhspan(value,
value + 1,
facecolor=highlightColor,
edgecolor='none',
zorder=1)
axErrorbar.set_title(ciTitle, fontsize=8)
axErrorbar.set_xlabel(statsResults.confIntervMethod.plotLabel,
fontsize=8)
if self.bCustomLimits:
axErrorbar.set_xlim([self.minX, self.maxX])
else:
self.minX, self.maxX = axErrorbar.get_xlim()
if self.bShowSeqPlot == False:
axErrorbar.set_yticks(np.arange(len(features)))
axErrorbar.set_yticklabels(features, size=8)
axErrorbar.set_ylim([-1, len(features)])
for label in axErrorbar.get_yticklabels():
label.set_size(8)
if label.get_text(
) in self.preferences['Selected statistical features']:
label.set_color('red')
else:
for label in axErrorbar.get_yticklabels():
label.set_visible(False)
for a in axErrorbar.yaxis.majorTicks:
a.set_visible(False)
for label in axErrorbar.get_xticklabels():
label.set_size(8)
for a in axErrorbar.xaxis.majorTicks:
a.tick1On = True
a.tick2On = False
for a in axErrorbar.yaxis.majorTicks:
a.tick1On = False
a.tick2On = False
for loc, spine in axErrorbar.spines.iteritems():
if loc in ['left', 'right', 'top']:
if loc != 'left' or (loc == 'left' and
(self.bShowSeqPlot == True
or bShowPowerPlot == True
or self.bShowPValueLabels == False)):
spine.set_color('none')
# *** Plot results of power test for each subsystem
if bShowPowerPlot == True:
lastAxes = axPower
axPower.scatter(power,
np.arange(len(features)),
s=20,
facecolor=orangeColour,
edgecolor='none',
marker='^',
linewidth=1,
zorder=10)
axPower.vlines(0.5,
-1,
len(features),
linestyle='dashed',
color=(0.25, 0.25, 0.25))
for value in np.arange(-0.5, len(features) - 1, 2):
axPower.axhspan(value,
value + 1,
facecolor=highlightColor,
edgecolor='none',
zorder=1)
axPower.set_xlabel('Power', fontsize=8)
axPower.set_xticks([0, 0.5, 1.0])
axPower.set_xlim(-0.1, 1.1)
axPower.set_ylim(-1, len(features))
for label in axPower.get_xticklabels():
label.set_size(8)
for label in axPower.get_yticklabels():
label.set_visible(False)
for a in axPower.yaxis.majorTicks:
a.set_visible(False)
for a in axPower.xaxis.majorTicks:
a.tick1On = True
a.tick2On = False
for loc, spine in axPower.spines.iteritems():
if loc in ['left', 'right', 'top']:
spine.set_color('none')
# *** Show p-values on right of last plot
if self.bShowPValueLabels == True:
axRight = lastAxes.twinx()
axRight.set_yticks(np.arange(len(pValueLabels)))
axRight.set_yticklabels(pValueLabels, size=8)
axRight.set_ylim([-1, len(pValueLabels)])
axRight.set_ylabel(pValueTitle, fontsize=8)
if bShowPowerPlot == True:
axRight.set_xticks([0, 0.5, 1.0])
for a in axRight.yaxis.majorTicks:
a.tick1On = False
a.tick2On = False
for label in axRight.get_yticklabels():
label.set_size(8)
for loc, spine in axRight.spines.iteritems():
if loc in ['left', 'right', 'top']:
spine.set_color('none')
self.updateGeometry()
self.draw()
def create_scatterhist(all_parameters,
parameter1,
parameter2,
measured,
modelled,
pars_names,
objective_function='SSE',
colormaps="red_yellow_green",
threshold=0.005,
*args,
**kwargs):
'''
Function to create the interactive scatterplot. Behavioural is always taken
as lower as the given threshold.
Parameters
----------
all_parameters : Nxp np.array
A number of parameter (p) value combinations used as model input. Length N
corresponds to the number of simulations
parameter1, parameter 2: {parametername : value}
Parametername of the parameters that the user wants to use for
model evaluation
modelled: Nxk np.array
simulation outputs for all parameterscombinations (N). Length k is the length
of the timeserie
measured: 1xk np.array
observations. Length k is the length
of the timeserie
parsnames: {parametername : value}
dict with all parameters (p) used the model
objective_function : Nx1 np.array
The user is allowed to choose between different objective functions
(SSE, RMSE, RRMSE)
colormaps :
Colormap of the scatterplot
treshold :
Value between 0 and 1. Parametersets for which the scaled value of the
objective function < threshold are retained (these parametersets are behavioral).
*args, **kwargs: args
arguments given to the scatter plot
example: s=15, marker='o', edgecolors= 'k', facecolor = 'white'
...
'''
# Extract values from pd dataframes
all_parameters = all_parameters.values
measured = measured.values
modelled = modelled.values
selected_of = _define_thresholdrange(measured, modelled,
objective_function)
#Translate color into a colormap
colormaps = translate_cmap(colormaps)
#Selected parameters & objective function
parameters = np.vstack(
(all_parameters[:, parameter1], all_parameters[:, parameter2])).T
scaled_of = np.array((selected_of - selected_of.min()) /
(selected_of.max() - selected_of.min()))
#calculation of the real threshold value from the relative one
real_threshold = threshold * (selected_of.max() -
selected_of.min()) + selected_of.min()
print("Current threshold = " + str(real_threshold))
#check if selected parameter are not the same
if parameter1 == parameter2:
raise Exception('Select two different parameters')
#check if parameter and of are really arrays
if not isinstance(parameters, np.ndarray):
raise Exception('parameters need to be numpy ndarray')
if not isinstance(scaled_of, np.ndarray):
raise Exception('objective function need to be numpy ndarray')
# check if objective function is of size 1xN
scaled_of = np.atleast_2d(scaled_of).T
if (len(scaled_of.shape) != 2):
raise Exception(
"Objective function need to be of size (1, N) got %s instead" %
(scaled_of.shape))
# check that SSE row length is equal to parameters
if not parameters.shape[0] == scaled_of.shape[0]:
raise Exception("None corresponding size of parameters and OF!")
# Check if threshold is in range of SSE values
if threshold < 0 or threshold > 1:
raise Exception("Threshold outside objective function ranges")
# Select behavioural parameter sets with of lower as threshold
search = np.where(scaled_of < threshold)
behav_par = parameters[search[0]]
behav_obj = selected_of[search[0]].T
print("Number of behavioural parametersets = " + str(behav_obj.shape[0]) +
" out of " + str(parameters.shape[0]))
if not behav_par.size > 0:
raise Exception('Threshold to severe, no behavioural sets.')
fig, ax_scatter = plt.subplots(figsize=(8, 6))
divider = make_axes_locatable(ax_scatter)
ax_scatter.set_autoscale_on(True)
# create a new axes with above the axScatter
ax_histx = divider.new_vertical(1.5, pad=0.0001, sharex=ax_scatter)
# create a new axes on the right side of the axScatter
ax_histy = divider.new_horizontal(1.5, pad=0.0001, sharey=ax_scatter)
fig.add_axes(ax_histx)
fig.add_axes(ax_histy)
# now determine nice limits by hand:
xmin = np.min(all_parameters[:, parameter1])
xmax = np.max(all_parameters[:, parameter1])
ymin = np.min(all_parameters[:, parameter2])
ymax = np.max(all_parameters[:, parameter2])
ax_histx.set_xlim((xmin, xmax))
ax_histy.set_ylim((ymin, ymax))
#determine binwidth (pylab examples:scatter_hist.py )
binwidth = 0.05
xymax = np.max([np.max(behav_par[:, 0]), np.max(behav_par[:, 1])])
lim = (int(xymax / binwidth) + 1) * binwidth
bins = np.arange(-lim, lim + binwidth, binwidth)
# create scatter & histogram
sc1 = ax_scatter.scatter(behav_par[:, 0],
behav_par[:, 1],
c=behav_obj,
edgecolors='none',
cmap=colormaps,
*args,
**kwargs)
ax_histx.hist(behav_par[:, 0], color='0.6', edgecolor='None', bins=bins)
ax_histy.hist(behav_par[:, 1],
orientation='horizontal',
edgecolor='None',
color='0.6',
bins=bins)
#determine number of bins scatter
majloc1 = MaxNLocator(nbins=5, prune='lower')
ax_scatter.yaxis.set_major_locator(majloc1)
majloc2 = MaxNLocator(nbins=5)
ax_scatter.xaxis.set_major_locator(majloc2)
ax_scatter.grid(linestyle='dashed', color='0.75', linewidth=1.)
ax_scatter.set_axisbelow(True)
#ax_histx.set_axisbelow(True)
plt.setp(ax_histx.get_xticklabels() + ax_histy.get_yticklabels(),
visible=False)
plt.setp(ax_histx.get_yticklabels() + ax_histy.get_xticklabels(),
visible=False)
ax_histy.set_xticks([])
ax_histx.set_yticks([])
ax_histx.xaxis.set_ticks_position('bottom')
ax_histy.yaxis.set_ticks_position('left')
ax_histy.spines['right'].set_color('none')
ax_histy.spines['top'].set_color('none')
ax_histy.spines['bottom'].set_color('none')
ax_histy.spines['left'].set_color('none')
ax_histx.spines['top'].set_color('none')
ax_histx.spines['right'].set_color('none')
ax_histx.spines['left'].set_color('none')
ax_histx.spines['bottom'].set_color('none')
ax_scatter.spines['top'].set_color('none')
ax_scatter.spines['right'].set_color('none')
# x and y label of parameter names
ax_scatter.set_xlabel(pars_names[parameter1],
horizontalalignment='center',
verticalalignment='center')
ax_scatter.set_ylabel(pars_names[parameter2],
horizontalalignment='center',
verticalalignment='center')
# Colorbar
cbar = fig.colorbar(sc1,
ax=ax_scatter,
cmap=colormaps,
orientation='vertical')
cbar.ax.set_ylabel('Value objective function')
behav_outputs = modelled[search[0]]
return fig, ax_scatter, ax_histx, ax_histy, sc1, cbar, behav_outputs
def plot(self, profile, statsResults):
# *** Check if there is sufficient data to generate the plot
if len(statsResults.activeData) <= 0:
self.emptyAxis()
return
features = statsResults.getColumn('Features')
if len(features) > 200:
QtGui.QApplication.instance().setOverrideCursor(
QtGui.QCursor(QtCore.Qt.ArrowCursor))
reply = QtGui.QMessageBox.question(
self, 'Continue?',
'Profile contains ' + str(len(features)) + ' features. ' +
'It may take several seconds to generate this plot. We recommend filtering your profile first. '
+ 'Do you wish to continue?', QtGui.QMessageBox.Yes,
QtGui.QMessageBox.No)
QtGui.QApplication.instance().restoreOverrideCursor()
if reply == QtGui.QMessageBox.No:
self.emptyAxis()
return
# *** Colour of plot elements
axesColour = str(self.preferences['Axes colour'].name())
group1Colour = str(
self.preferences['Group colours'][profile.groupName1].name())
group2Colour = str(
self.preferences['Group colours'][profile.groupName2].name())
# *** Colour of plot elements
highlightColor = (0.9, 0.9, 0.9)
# *** Sort data
if self.sortingField == 'p-values':
statsResults.activeData = TableHelper.SortTable(statsResults.activeData,\
[statsResults.dataHeadings['pValues']], False)
elif self.sortingField == 'Effect sizes':
statsResults.activeData = TableHelper.SortTable(statsResults.activeData,\
[statsResults.dataHeadings['EffectSize']],
True, True, False)
elif self.sortingField == 'Feature labels':
statsResults.activeData = TableHelper.SortTableStrCol(statsResults.activeData,\
statsResults.dataHeadings['Features'], False)
features = statsResults.getColumn(
'Features') # get sorted feature labels
# *** Create lists for each quantity of interest
if statsResults.multCompCorrection.method == 'False discovery rate':
pValueTitle = 'q-value'
else:
pValueTitle = 'p-value'
if self.bShowCorrectedPvalues:
pValueLabels = statsResults.getColumnAsStr('pValuesCorrected')
pValueTitle += ' (corrected)'
else:
pValueLabels = statsResults.getColumnAsStr('pValues')
effectSizes = statsResults.getColumn('EffectSize')
lowerCIs = statsResults.getColumn('LowerCI')
upperCIs = statsResults.getColumn('UpperCI')
ciTitle = (
'%.3g' %
(statsResults.oneMinusAlpha() * 100)) + '% confidence intervals'
# *** Truncate feature labels
highlightedFeatures = list(
self.preferences['Highlighted group features'])
if self.preferences['Truncate feature names']:
length = self.preferences['Length of truncated feature names']
for i in xrange(0, len(features)):
if len(features[i]) > length + 3:
features[i] = features[i][0:length] + '...'
for i in xrange(0, len(highlightedFeatures)):
if len(highlightedFeatures[i]) > length + 3:
highlightedFeatures[
i] = highlightedFeatures[i][0:length] + '...'
# *** Check that there is at least one significant feature
if len(features) <= 0:
self.emptyAxis('No significant features')
return
# *** Adjust effect size for axis scale
dominateInSample2 = []
percentage1 = []
percentage2 = []
for i in xrange(0, len(effectSizes)):
if statsResults.bConfIntervRatio:
if effectSizes[i] < 1:
# mirror CI across y-axis
effectSizes[i] = 1.0 / effectSizes[i]
lowerCI = effectSizes[i] - (1.0 / upperCIs[i])
upperCI = (1.0 / lowerCIs[i]) - effectSizes[i]
lowerCIs[i] = lowerCI
upperCIs[i] = upperCI
dominateInSample2.append(i)
else:
lowerCIs[i] = effectSizes[i] - lowerCIs[i]
upperCIs[i] = upperCIs[i] - effectSizes[i]
else:
lowerCIs[i] = effectSizes[i] - lowerCIs[i]
upperCIs[i] = upperCIs[i] - effectSizes[i]
if effectSizes[i] < 0.0:
dominateInSample2.append(i)
# *** Set figure size
if self.legendPos == 3 or self.legendPos == 4 or self.legendPos == 8: # bottom legend
heightBottomLabels = 0.56 # inches
else:
heightBottomLabels = 0.4 # inches
heightTopLabels = 0.25
plotHeight = self.figHeightPerRow * len(features)
self.imageWidth = self.figWidth
self.imageHeight = plotHeight + heightBottomLabels + heightTopLabels
if self.imageWidth > 256 or self.imageHeight > 256:
QtGui.QApplication.instance().setOverrideCursor(
QtGui.QCursor(QtCore.Qt.ArrowCursor))
self.emptyAxis()
reply = QtGui.QMessageBox.question(
self, 'Excessively large plot',
'The resulting plot is too large to display.')
QtGui.QApplication.instance().restoreOverrideCursor()
return
self.fig.set_size_inches(self.imageWidth, self.imageHeight)
# *** Determine width of y-axis labels
yLabelBounds = self.yLabelExtents(features, 8)
# *** Size plots which comprise the extended errorbar plot
self.fig.clear()
spacingBetweenPlots = 0.25 # inches
widthNumSeqPlot = 1.25 # inches
if self.bShowBarPlot == False:
widthNumSeqPlot = 0.0
spacingBetweenPlots = 0.0
widthPvalueLabels = 0.75 # inches
if self.bShowPValueLabels == False:
widthPvalueLabels = 0.1
yPlotOffsetFigSpace = heightBottomLabels / self.imageHeight
heightPlotFigSpace = plotHeight / self.imageHeight
xPlotOffsetFigSpace = yLabelBounds.width + 0.1 / self.imageWidth
pValueLabelWidthFigSpace = widthPvalueLabels / self.imageWidth
widthPlotFigSpace = 1.0 - pValueLabelWidthFigSpace - xPlotOffsetFigSpace
widthErrorBarPlot = widthPlotFigSpace * self.imageWidth - widthNumSeqPlot - spacingBetweenPlots
axInitAxis = self.fig.add_axes([
xPlotOffsetFigSpace, yPlotOffsetFigSpace, widthPlotFigSpace,
heightPlotFigSpace
])
divider = make_axes_locatable(axInitAxis)
divider.get_vertical()[0] = Size.Fixed(
len(features) * self.figHeightPerRow)
if self.bShowBarPlot == True:
divider.get_horizontal()[0] = Size.Fixed(widthNumSeqPlot)
axErrorbar = divider.new_horizontal(widthErrorBarPlot,
pad=spacingBetweenPlots,
sharey=axInitAxis)
self.fig.add_axes(axErrorbar)
else:
divider.get_horizontal()[0] = Size.Fixed(widthErrorBarPlot)
axErrorbar = axInitAxis
# *** Plot of sequences for each subsystem
if self.bShowBarPlot == True:
axNumSeq = axInitAxis
meanRelFreqSeqs1 = statsResults.getColumn('MeanRelFreq1')
meanRelFreqSeqs2 = statsResults.getColumn('MeanRelFreq2')
if self.bShowStdDev:
stdDev1 = statsResults.getColumn('StdDevRelFreq1')
stdDev2 = statsResults.getColumn('StdDevRelFreq2')
endCapSize = self.endCapSize
else:
stdDev1 = [0] * len(meanRelFreqSeqs1)
stdDev2 = [0] * len(meanRelFreqSeqs2)
endCapSize = 0
axNumSeq.barh(np.arange(len(features)) + 0.0,
meanRelFreqSeqs1,
height=0.3,
xerr=stdDev1,
color=group1Colour,
ecolor='black',
capsize=endCapSize)
axNumSeq.barh(np.arange(len(features)) - 0.3,
meanRelFreqSeqs2,
height=0.3,
xerr=stdDev2,
color=group2Colour,
ecolor='black',
capsize=endCapSize)
for value in np.arange(-0.5, len(features) - 1, 2):
axNumSeq.axhspan(value,
value + 1,
facecolor=highlightColor,
edgecolor='none',
zorder=-1)
axNumSeq.set_xlabel('Mean proportion (%)')
maxPercentage = max(max(meanRelFreqSeqs1), max(meanRelFreqSeqs2))
axNumSeq.set_xticks([0, maxPercentage])
axNumSeq.set_xlim([0, maxPercentage * 1.05])
maxPercentageStr = '%.1f' % maxPercentage
axNumSeq.set_xticklabels(['0.0', maxPercentageStr])
axNumSeq.set_yticks(np.arange(len(features)))
axNumSeq.set_yticklabels(features)
axNumSeq.set_ylim([-1, len(features)])
for label in axNumSeq.get_yticklabels():
if label.get_text() in highlightedFeatures:
label.set_color('red')
for a in axNumSeq.yaxis.majorTicks:
a.tick1On = False
a.tick2On = False
for a in axNumSeq.xaxis.majorTicks:
a.tick1On = True
a.tick2On = False
for line in axNumSeq.yaxis.get_ticklines():
line.set_color(axesColour)
for line in axNumSeq.xaxis.get_ticklines():
line.set_color(axesColour)
for loc, spine in axNumSeq.spines.iteritems():
if loc in ['left', 'right', 'top']:
spine.set_color('none')
else:
spine.set_color(axesColour)
# *** Plot confidence intervals for each subsystem
lastAxes = axErrorbar
markerSize = math.sqrt(float(self.markerSize))
axErrorbar.errorbar(effectSizes,
np.arange(len(features)),
xerr=[lowerCIs, upperCIs],
fmt='o',
ms=markerSize,
mfc=group1Colour,
mec='black',
ecolor='black',
zorder=10)
effectSizesSample2 = [
effectSizes[value] for value in dominateInSample2
]
axErrorbar.plot(effectSizesSample2,
dominateInSample2,
ls='',
marker='o',
ms=markerSize,
mfc=group2Colour,
mec='black',
zorder=100)
if statsResults.bConfIntervRatio:
axErrorbar.vlines(1,
-1,
len(features),
linestyle='dashed',
color=axesColour)
else:
axErrorbar.vlines(0,
-1,
len(features),
linestyle='dashed',
color=axesColour)
for value in np.arange(-0.5, len(features) - 1, 2):
axErrorbar.axhspan(value,
value + 1,
facecolor=highlightColor,
edgecolor='none',
zorder=1)
axErrorbar.set_title(ciTitle)
axErrorbar.set_xlabel('Difference in mean proportions (%)')
if self.bCustomLimits:
axErrorbar.set_xlim([self.minX, self.maxX])
else:
self.minX, self.maxX = axErrorbar.get_xlim()
if self.bShowBarPlot == False:
axErrorbar.set_yticks(np.arange(len(features)))
axErrorbar.set_yticklabels(features)
axErrorbar.set_ylim([-1, len(features)])
for label in axErrorbar.get_yticklabels():
if label.get_text(
) in self.preferences['Highlighted group features']:
label.set_color('red')
else:
for label in axErrorbar.get_yticklabels():
label.set_visible(False)
for a in axErrorbar.yaxis.majorTicks:
a.set_visible(False)
for a in axErrorbar.xaxis.majorTicks:
a.tick1On = True
a.tick2On = False
for a in axErrorbar.yaxis.majorTicks:
a.tick1On = False
a.tick2On = False
for line in axErrorbar.yaxis.get_ticklines():
line.set_visible(False)
for line in axErrorbar.xaxis.get_ticklines():
line.set_color(axesColour)
for loc, spine in axErrorbar.spines.iteritems():
if loc in ['left', 'right', 'top']:
spine.set_color('none')
else:
spine.set_color(axesColour)
# *** Show p-values on right of last plot
if self.bShowPValueLabels == True:
axRight = lastAxes.twinx()
axRight.set_yticks(np.arange(len(pValueLabels)))
axRight.set_yticklabels(pValueLabels)
axRight.set_ylim([-1, len(pValueLabels)])
axRight.set_ylabel(pValueTitle)
for a in axRight.yaxis.majorTicks:
a.tick1On = False
a.tick2On = False
# *** Legend
if self.legendPos != -1:
legend1 = Rectangle((0, 0), 1, 1, fc=group1Colour)
legend2 = Rectangle((0, 0), 1, 1, fc=group2Colour)
legend = self.fig.legend([legend1, legend2],
(profile.groupName1, profile.groupName2),
loc=self.legendPos,
ncol=2)
legend.get_frame().set_linewidth(0)
self.updateGeometry()
self.draw()
from mpl_toolkits.axes_grid import make_axes_locatable
import matplotlib.axes as maxes
from matplotlib import cm
if __name__ == '__main__':
fig = P.figure(figsize=(10,10))
ax1 = fig.add_subplot(221)
ax2 = fig.add_subplot(222)
ax3 = fig.add_subplot(223)
ax4 = fig.add_subplot(224)
s1 = ax1.scatter(numpy.random.rand(10),
numpy.random.rand(10),
c=numpy.random.rand(10))
divider = make_axes_locatable(ax1)
cax1 = divider.new_horizontal('5%', pad=0.0, axes_class=maxes.Axes)
fig.add_axes(cax1)
c1 = fig.colorbar(s1, cax = cax1,orientation = 'horizontal')
s2 = ax2.scatter(numpy.random.rand(10),
numpy.random.rand(10),
c=numpy.random.rand(10),
cmap = cm.get_cmap('jet'))
divider = make_axes_locatable(ax2)
cax2 = divider.append_axes('right', 0.1, pad=0.1)
c2 = fig.colorbar(s2, cax = cax2)
#p = matplotlib.patches.Patch(color=cm.get_cmap('jet'))
#ax2.legend([p],['Test'])
s3 = ax3.scatter(numpy.random.rand(10),
def histogram(values, label_x=None, label_y=None, color="#00ff00", size=(650,650)):
"""
Values will come as a list of tuples representing (x,y) values.
values = [(0,100),(15,2500),(20,3000),(25,2700),(30,2800),(35,3600),(40,4200),(45,3500),(50,2100)]
label_x = "Age"
label_y = "Population"
"""
# unpack the tuple value list into x and y lists
x_vals = []; y_vals = []
for tuple in values:
first, second = tuple;
x_vals.append(first)
y_vals.append(second)
tick_val = max(min(x_vals, y_vals))
# convert the data to a numpy list
x = np.array(x_vals)
y = np.array(y_vals)
# the width and height of the figure
fig_w = int(size[0])/100
fig_h = int(size[1])/100
# set up the graph descriptor object and prefrences for it
fig = Figure(figsize=(fig_w,fig_h), dpi=100, facecolor='#ffffff', frameon=False)
axScatter = fig.add_subplot(111)
axScatter.yaxis.grid(True, linestyle='-', which='major', color='lightgrey',alpha=0.5)
axScatter.xaxis.grid(True, linestyle='-', which='major', color='lightgrey',alpha=0.5)
divider = make_axes_locatable(axScatter)
# create a new axes with a height of 1.2 inch above the axScatter
axHistx = divider.new_vertical(1.2, pad=0.1, sharex=axScatter)
axHistx.yaxis.grid(True, linestyle='-', which='major', color='lightgrey',alpha=0.5)
axHistx.xaxis.grid(True, linestyle='-', which='major', color='lightgrey',alpha=0.5)
# create a new axes with a width of 1.2 inch on the right side of the axScatter
axHisty = divider.new_horizontal(1.2, pad=0.1, sharey=axScatter)
axHisty.yaxis.grid(True, linestyle='-', which='major', color='lightgrey',alpha=0.5)
axHisty.xaxis.grid(True, linestyle='-', which='major', color='lightgrey',alpha=0.5)
fig.add_axes(axHistx)
fig.add_axes(axHisty)
# plot the scatterplot and set aspect ratio, grid lines
axScatter.scatter(x, y, facecolors=color)
axScatter.set_aspect("auto")
axScatter.grid(True)
# now determine limits manually and set the bin number
binwidth = 0.25
#xymax = np.max( [np.max(np.fabs(x)), np.max(np.fabs(y))] )
#lim = ( int(xymax/binwidth) + 1) * binwidth
#bins = np.arange(-lim, lim + binwidth, binwidth)
# plot the histograms for x and y
axHistx.hist(x, bins=9, facecolor=color, alpha=0.75)
axHisty.hist(y, bins=9, orientation='horizontal', facecolor=color, alpha=0.75)
# the xaxis of axHistx and yaxis of axHisty are shared with axScatter,
# thus there is no need to manually adjust the xlim and ylim of these
# axis.
# axHistx.axis["bottom"].major_ticklabels.set_visible(False)
for tl in axHistx.get_xticklabels():
tl.set_visible(False)
axHistx.set_yticks([0, int(tick_val/8), int(tick_val/4)])
axHistx.set_title(label_x)
# axHisty.axis["left"].major_ticklabels.set_visible(False)
for tl in axHisty.get_yticklabels():
tl.set_visible(False)
axHisty.set_xticks([0, int(tick_val/8), int(tick_val/4)])
axHisty.set_title(label_y)
return fig
def create_scatterhist(all_parameters, parameter1, parameter2, objective,
pars_names, xbinwidth = 0.05, ybinwidth = 0.05,
objective_function='SSE',
colormaps = "red_yellow_green",
threshold=0.02,
season = 'Winter',
*args, **kwargs):
'''
Function to create the interactive scatterplot. Behavioural is always taken
as lower as the given threshold.
Parameters
----------
all_parameters : Nxp np.array
A number of parameter (p) value combinations used as model input. Length N
corresponds to the number of simulations
parameter1, parameter 2: {parametername : value}
Parametername of the parameters that the user wants to use for
model evaluation
objective: nxk pd.Dataframe
Values of objective function for all simulations (n) and for all possible combination (k)
of seasons and criteria (SSE, RMSE, RRMSE). Each column contains the
the objective function values for all simulations calculated
for a selected criteria (SSE, RMSE, RRMSE) based on the whole dateset
or on a subdataset (specific season).
parsnames: {parametername : value}
dict with all parameters (p) used the model
xbinwidth: float
defines the width of the xbin to the data used
ybinwidth: float
defines the width of the ybin to the data used
objective_function : Nx1 np.array
The user is allowed to choose between different objective functions
(SSE, RMSE, RRMSE)
colormaps :
Colormap of the scatterplot
treshold :
Value between 0 and 1. Parametersets for which the scaled value of the
objective function < threshold are retained (these parametersets are behavioral).
*args, **kwargs: args
arguments given to the scatter plot
example: s=15, marker='o', edgecolors= 'k', facecolor = 'white'
Returns
---------
fig: matplotlib.figure.Figure object
the resulting figure
axScatter: matplotlib.axes.AxesSubplot object
the scatter plot with the datapoints, can be used to add labels or
change the current ticks settings
axHistx: matplotlib.axes.AxesSubplot object
the x-axis histogram
axHisty: matplotlib.axes.AxesSubplot object
the y-axis histogram
...
'''
# Selected season & objective_function
all_parameters=all_parameters.values
selected_of= _select_season_OF(objective,season,objective_function)
#Translate color into a colormap
colormaps = translate_cmap(colormaps)
#Selected parameters & objective function
parameters = np.vstack((all_parameters[:,parameter1], all_parameters[:,parameter2])).T
scaled_of = np.array((selected_of-selected_of.min())/(selected_of.max()-selected_of.min()))
#calculation of the real threshold value from the relative one
real_threshold = threshold*(selected_of.max() - selected_of.min())+selected_of.min()
print("Current threshold = " + str(real_threshold))
#check if selected parameter are not the same
if parameter1 == parameter2:
raise Exception('Select two different parameters')
#check if parameter and of are really arrays
if not isinstance(parameters, np.ndarray):
raise Exception('parameters need to be numpy ndarray')
if not isinstance(scaled_of, np.ndarray):
raise Exception('objective function need to be numpy ndarray')
# check if objective function is of size 1xN
scaled_of = np.atleast_2d(scaled_of).T
if (len(scaled_of.shape) != 2 ):
raise Exception("Objective function need to be of size (1, N) got %s instead" %(scaled_of.shape))
# check that SSE row length is equal to parameters
if not parameters.shape[0] == scaled_of.shape[0]:
raise Exception("None corresponding size of parameters and OF!")
# Check if threshold is in range of SSE values
if threshold < 0 or threshold > 1:
raise Exception("Threshold outside objective function ranges")
# Select behavioural parameter sets with of lower as threshold
search=np.where(scaled_of < threshold)
behav_par = parameters[search[0]]
behav_obj = selected_of[search[0]].T
print("Number of behavioural parametersets = " + str(behav_obj.shape[0]) + " out of " + str(parameters.shape[0]))
if not behav_par.size > 0:
raise Exception('Threshold to severe, no behavioural sets.')
fig, ax_scatter = plt.subplots(figsize=(8,6))
divider = make_axes_locatable(ax_scatter)
ax_scatter.set_autoscale_on(True)
# create a new axes with above the axScatter
ax_histx = divider.new_vertical(1.5, pad=0.0001, sharex=ax_scatter)
# create a new axes on the right side of the axScatter
ax_histy = divider.new_horizontal(1.5, pad=0.0001, sharey=ax_scatter)
fig.add_axes(ax_histx)
fig.add_axes(ax_histy)
# now determine nice limits by hand:
xmin = np.min(all_parameters[:,parameter1])
xmax = np.max(all_parameters[:,parameter1])
ymin = np.min(all_parameters[:,parameter2])
ymax = np.max(all_parameters[:,parameter2])
ax_histx.set_xlim( (xmin, xmax) )
ax_histy.set_ylim( (ymin, ymax) )
#determine binwidth (pylab examples:scatter_hist.py )
xbinwidth = (xmax-xmin)*xbinwidth
binsx = np.arange(xmin,xmax+xbinwidth,xbinwidth)
ybinwidth = (ymax-ymin)*ybinwidth
binsy = np.arange(ymin,ymax+ybinwidth,ybinwidth)
# create scatter & histogram
sc1 = ax_scatter.scatter(behav_par[:,0], behav_par[:,1], c=behav_obj,
edgecolors= 'none', cmap=colormaps, *args, **kwargs)
ax_histx.hist(behav_par[:,0], color='0.6', edgecolor='None', bins = binsx)
ax_histy.hist(behav_par[:,1], orientation='horizontal', edgecolor='None',
color='0.6', bins=binsy)
#determine number of bins scatter
majloc1 = MaxNLocator(nbins=5, prune='lower')
ax_scatter.yaxis.set_major_locator(majloc1)
majloc2 = MaxNLocator(nbins=5)
ax_scatter.xaxis.set_major_locator(majloc2)
ax_scatter.grid(linestyle = 'dashed', color = '0.75',linewidth = 1.)
ax_scatter.set_axisbelow(True)
#ax_histx.set_axisbelow(True)
plt.setp(ax_histx.get_xticklabels() + ax_histy.get_yticklabels(),
visible=False)
plt.setp(ax_histx.get_yticklabels() + ax_histy.get_xticklabels(),
visible=False)
ax_histy.set_xticks([])
ax_histx.set_yticks([])
ax_histx.xaxis.set_ticks_position('bottom')
ax_histy.yaxis.set_ticks_position('left')
ax_histy.spines['right'].set_color('none')
ax_histy.spines['top'].set_color('none')
ax_histy.spines['bottom'].set_color('none')
ax_histy.spines['left'].set_color('none')
ax_histx.spines['top'].set_color('none')
ax_histx.spines['right'].set_color('none')
ax_histx.spines['left'].set_color('none')
ax_histx.spines['bottom'].set_color('none')
ax_scatter.spines['top'].set_color('none')
ax_scatter.spines['right'].set_color('none')
# x and y label of parameter names
ax_scatter.set_xlabel(pars_names[parameter1], horizontalalignment ='center', verticalalignment ='center')
ax_scatter.set_ylabel(pars_names[parameter2], horizontalalignment ='center', verticalalignment ='center')
# Colorbar
cbar = fig.colorbar(sc1, ax=ax_scatter, cmap=colormaps,
orientation='vertical')
cbar.ax.set_ylabel('Value objective function')
return fig, ax_scatter, ax_histx, ax_histy, sc1, cbar
def plot(axes, net):
classname = net.__class__.__name__
axes.set_xticks([])
axes.set_yticks([])
divider = make_axes_locatable(axes)
subaxes = divider.new_vertical(1.0, pad=0.4, sharex=axes)
fig.add_axes(subaxes)
subaxes.set_xticks([])
subaxes.yaxis.set_major_locator(matplotlib.ticker.MaxNLocator(2))
subaxes.yaxis.set_ticks_position('right')
subaxes.set_ylabel('Distortion')
subaxes.set_xlabel('Time')
Y = net.distortion[::1]
X = np.arange(len(Y)) / float(len(Y) - 1)
subaxes.plot(X, Y)
if classname == 'NG':
plt.title('Neural Gas', fontsize=20)
elif classname == 'SOM':
plt.title('Self-Organizing Map', fontsize=20)
elif classname == 'DSOM':
plt.title('Dynamic Self-Organizing Map', fontsize=20)
axes.axis([0, 1, 0, 1])
axes.set_aspect(1)
codebook = net.codebook
axes.imshow(codebook, interpolation='nearest')
#interpolation='bicubic')
if classname == 'NG':
axes.text(
0.5,
-0.01,
r'$\lambda_i = %.3f,\lambda_f = %.3f, \varepsilon_i=%.3f, \varepsilon_f=%.3f$'
% (net.sigma_i, net.sigma_f, net.lrate_i, net.lrate_f),
fontsize=16,
horizontalalignment='center',
verticalalignment='top',
transform=axes.transAxes)
if classname == 'SOM':
axes.text(
0.5,
-0.01,
r'$\sigma_i = %.3f,\sigma_f = %.3f, \varepsilon_i=%.3f, \varepsilon_f=%.3f$'
% (net.sigma_i, net.sigma_f, net.lrate_i, net.lrate_f),
fontsize=16,
horizontalalignment='center',
verticalalignment='top',
transform=axes.transAxes)
elif classname == 'DSOM':
axes.text(0.5,
-0.01,
r'$elasticity = %.2f$' % (net.elasticity),
fontsize=16,
horizontalalignment='center',
verticalalignment='top',
transform=axes.transAxes)
def plot(self, profile, statsResults):
# *** Check if there is sufficient data to generate the plot
if len(statsResults.activeData) <= 0:
self.emptyAxis()
return
features = statsResults.getColumn('Features')
if len(features) > 200:
QtGui.QApplication.instance().setOverrideCursor(QtGui.QCursor(QtCore.Qt.ArrowCursor))
reply = QtGui.QMessageBox.question(self, 'Continue?', 'Profile contains ' + str(len(features)) + ' features. ' +
'It may take several seconds to generate this plot. We recommend filtering your profile first. ' +
'Do you wish to continue?', QtGui.QMessageBox.Yes, QtGui.QMessageBox.No)
QtGui.QApplication.instance().restoreOverrideCursor()
if reply == QtGui.QMessageBox.No:
self.emptyAxis()
return
# *** Colour of plot elements
axesColour = str(self.preferences['Axes colour'].name())
profile1Colour = str(self.preferences['Sample 1 colour'].name())
profile2Colour = str(self.preferences['Sample 2 colour'].name())
# *** Colour of plot elements
highlightColor = (0.9, 0.9, 0.9)
# *** Sort data
if self.sortingField == 'p-values':
statsResults.activeData = TableHelper.SortTable(statsResults.activeData,\
[statsResults.dataHeadings['pValues']], False)
elif self.sortingField == 'Effect sizes':
statsResults.activeData = TableHelper.SortTable(statsResults.activeData,\
[statsResults.dataHeadings['EffectSize']],
True, True, statsResults.confIntervMethod.bRatio)
elif self.sortingField == 'Feature labels':
statsResults.activeData = TableHelper.SortTableStrCol(statsResults.activeData,\
statsResults.dataHeadings['Features'], False)
features = statsResults.getColumn('Features') # get sorted feature labels
# *** Create lists for each quantity of interest
if statsResults.multCompCorrection.method == 'False discovery rate':
pValueTitle = 'q-value'
else:
pValueTitle = 'p-value'
if self.bShowCorrectedPvalues:
pValueLabels = statsResults.getColumnAsStr('pValuesCorrected')
if statsResults.multCompCorrection.method != 'No correction':
pValueTitle += ' (corrected)'
else:
pValueLabels = statsResults.getColumnAsStr('pValues')
effectSizes = statsResults.getColumn('EffectSize')
lowerCIs = statsResults.getColumn('LowerCI')
upperCIs = statsResults.getColumn('UpperCI')
ciTitle = ('%.3g' % (statsResults.oneMinusAlpha()*100)) + '% confidence intervals'
seqs1 = statsResults.getColumn('Seq1')
seqs2 = statsResults.getColumn('Seq2')
parentSeqs1 = statsResults.getColumn('ParentalSeq1')
parentSeqs2 = statsResults.getColumn('ParentalSeq2')
# *** Truncate feature labels
highlightedFeatures = list(self.preferences['Highlighted sample features'])
if self.preferences['Truncate feature names']:
length = self.preferences['Length of truncated feature names']
for i in xrange(0, len(features)):
if len(features[i]) > length+3:
features[i] = features[i][0:length] + '...'
for i in xrange(0, len(highlightedFeatures)):
if len(highlightedFeatures[i]) > length+3:
highlightedFeatures[i] = highlightedFeatures[i][0:length] + '...'
# *** Adjust effect size for axis scale
dominateInSample2 = []
percentage1 = []
percentage2 = []
for i in xrange(0, len(effectSizes)):
percentage1.append(float(seqs1[i])*100 / parentSeqs1[i])
percentage2.append(float(seqs2[i])*100 / parentSeqs2[i])
if statsResults.confIntervMethod.bRatio:
if effectSizes[i] < 1:
# mirror CI across y-axis
effectSizes[i] = 1.0 / effectSizes[i]
lowerCI = effectSizes[i] - (1.0 / upperCIs[i])
upperCI = (1.0 / lowerCIs[i]) - effectSizes[i]
lowerCIs[i] = lowerCI
upperCIs[i] = upperCI
dominateInSample2.append(i)
else:
lowerCIs[i] = effectSizes[i] - lowerCIs[i]
upperCIs[i] = upperCIs[i] - effectSizes[i]
else:
lowerCIs[i] = effectSizes[i] - lowerCIs[i]
upperCIs[i] = upperCIs[i] - effectSizes[i]
if effectSizes[i] < 0.0:
dominateInSample2.append(i)
# *** Set figure size
if self.legendPos == 3 or self.legendPos == 4 or self.legendPos == 8: # bottom legend
heightBottomLabels = 0.56 # inches
else:
heightBottomLabels = 0.4 # inches
heightTopLabels = 0.25
plotHeight = self.figHeightPerRow*len(features)
self.imageWidth = self.figWidth
self.imageHeight = plotHeight + heightBottomLabels + heightTopLabels
if self.imageWidth > 256 or self.imageHeight > 256:
QtGui.QApplication.instance().setOverrideCursor(QtGui.QCursor(QtCore.Qt.ArrowCursor))
self.emptyAxis()
reply = QtGui.QMessageBox.question(self, 'Excessively large plot', 'The resulting plot is too large to display.')
QtGui.QApplication.instance().restoreOverrideCursor()
return
self.fig.set_size_inches(self.imageWidth, self.imageHeight)
# *** Determine width of y-axis labels
yLabelBounds = self.yLabelExtents(features, 8)
# *** Size plots which comprise the extended errorbar plot
self.fig.clear()
spacingBetweenPlots = 0.25 # inches
widthNumSeqPlot = 1.25 # inches
if self.bShowBarPlot == False:
widthNumSeqPlot = 0.0
spacingBetweenPlots = 0.0
widthPvalueLabels = 0.75 # inches
if self.bShowPValueLabels == False:
widthPvalueLabels = 0.1
yPlotOffsetFigSpace = heightBottomLabels / self.imageHeight
heightPlotFigSpace = plotHeight / self.imageHeight
xPlotOffsetFigSpace = yLabelBounds.width + 0.1 / self.imageWidth
pValueLabelWidthFigSpace = widthPvalueLabels / self.imageWidth
widthPlotFigSpace = 1.0 - pValueLabelWidthFigSpace - xPlotOffsetFigSpace
widthErrorBarPlot = widthPlotFigSpace*self.imageWidth - widthNumSeqPlot - spacingBetweenPlots
axInitAxis = self.fig.add_axes([xPlotOffsetFigSpace,yPlotOffsetFigSpace,widthPlotFigSpace,heightPlotFigSpace])
divider = make_axes_locatable(axInitAxis)
divider.get_vertical()[0] = Size.Fixed(len(features)*self.figHeightPerRow)
if self.bShowBarPlot == True:
divider.get_horizontal()[0] = Size.Fixed(widthNumSeqPlot)
axErrorbar = divider.new_horizontal(widthErrorBarPlot, pad=spacingBetweenPlots, sharey=axInitAxis)
self.fig.add_axes(axErrorbar)
else:
divider.get_horizontal()[0] = Size.Fixed(widthErrorBarPlot)
axErrorbar = axInitAxis
# *** Plot of sequences for each subsystem
if self.bShowBarPlot == True:
axNumSeq = axInitAxis
if self.percentageOrSeqCount == 'Proportion (%)':
# plot percentage
axNumSeq.barh(np.arange(len(features))+0.0, percentage1, height = 0.3, color=profile1Colour, zorder=10, ecolor='black')
axNumSeq.barh(np.arange(len(features))-0.3, percentage2, height = 0.3, color=profile2Colour, zorder=10, ecolor='black')
for value in np.arange(-0.5, len(features)-1, 2):
axNumSeq.axhspan(value, value+1, facecolor=highlightColor,edgecolor='none',zorder=1)
axNumSeq.set_xlabel(self.percentageOrSeqCount)
maxPercentage = max(max(percentage1), max(percentage2))
axNumSeq.set_xticks([0, maxPercentage])
axNumSeq.set_xlim([0, maxPercentage*1.05])
maxPercentageStr = '%.1f' % maxPercentage
axNumSeq.set_xticklabels(['0.0', maxPercentageStr])
else:
# plot sequence count
axNumSeq.barh(np.arange(len(features))+0.0, seqs1, height = 0.3, color=profile1Colour, zorder=10, ecolor='black')
axNumSeq.barh(np.arange(len(features))-0.3, seqs2, height = 0.3, color=profile2Colour, zorder=10, ecolor='black')
for value in np.arange(-0.5, len(features)-1, 2):
axNumSeq.axhspan(value, value+1, facecolor=highlightColor,edgecolor='none',zorder=1)
axNumSeq.set_xlabel(self.percentageOrSeqCount)
maxSeqs = max(max(seqs1), max(seqs2))
axNumSeq.set_xticks([0, maxSeqs])
axNumSeq.set_xlim([0, maxSeqs*1.05])
axNumSeq.set_xticklabels([0, str(maxSeqs)])
axNumSeq.set_yticks(np.arange(len(features)))
axNumSeq.set_yticklabels(features)
axNumSeq.set_ylim([-1, len(features)])
for label in axNumSeq.get_yticklabels():
if label.get_text() in highlightedFeatures:
label.set_color('red')
for a in axNumSeq.yaxis.majorTicks:
a.tick1On=False
a.tick2On=False
for a in axNumSeq.xaxis.majorTicks:
a.tick1On=True
a.tick2On=False
for line in axNumSeq.yaxis.get_ticklines():
line.set_color(axesColour)
for line in axNumSeq.xaxis.get_ticklines():
line.set_color(axesColour)
for loc, spine in axNumSeq.spines.iteritems():
if loc in ['left', 'right','top']:
spine.set_color('none')
else:
spine.set_color(axesColour)
# *** Plot confidence intervals for each subsystem
lastAxes = axErrorbar
markerSize = math.sqrt(float(self.markerSize))
axErrorbar.errorbar(effectSizes, np.arange(len(features)), xerr=[lowerCIs,upperCIs], fmt='o', ms=markerSize, mfc=profile1Colour, mec='black', ecolor='black', zorder=10)
effectSizesSample2 = [effectSizes[value] for value in dominateInSample2]
axErrorbar.plot(effectSizesSample2, dominateInSample2, ls='', marker='o', ms=markerSize, mfc=profile2Colour, mec='black', zorder=100)
if statsResults.confIntervMethod.bRatio:
axErrorbar.vlines(1, -1, len(features), linestyle='dashed', color=axesColour)
else:
axErrorbar.vlines(0, -1, len(features), linestyle='dashed', color=axesColour)
for value in np.arange(-0.5, len(features)-1, 2):
axErrorbar.axhspan(value, value+1, facecolor=highlightColor,edgecolor='none',zorder=1)
axErrorbar.set_title(ciTitle)
axErrorbar.set_xlabel(statsResults.confIntervMethod.plotLabel)
if self.bCustomLimits:
axErrorbar.set_xlim([self.minX, self.maxX])
else:
self.minX, self.maxX = axErrorbar.get_xlim()
if self.bShowBarPlot == False:
axErrorbar.set_yticks(np.arange(len(features)))
axErrorbar.set_yticklabels(features)
axErrorbar.set_ylim([-1, len(features)])
for label in axErrorbar.get_yticklabels():
if label.get_text() in highlightedFeatures:
label.set_color('red')
else:
for label in axErrorbar.get_yticklabels():
label.set_visible(False)
for a in axErrorbar.yaxis.majorTicks:
a.set_visible(False)
for a in axErrorbar.xaxis.majorTicks:
a.tick1On=True
a.tick2On=False
for a in axErrorbar.yaxis.majorTicks:
a.tick1On=False
a.tick2On=False
for line in axErrorbar.yaxis.get_ticklines():
line.set_visible(False)
for line in axErrorbar.xaxis.get_ticklines():
line.set_color(axesColour)
for loc, spine in axErrorbar.spines.iteritems():
if loc in ['left','right','top']:
spine.set_color('none')
else:
spine.set_color(axesColour)
# *** Show p-values on right of last plot
if self.bShowPValueLabels == True:
axRight = lastAxes.twinx()
axRight.set_yticks(np.arange(len(pValueLabels)))
axRight.set_yticklabels(pValueLabels)
axRight.set_ylim([-1, len(pValueLabels)])
axRight.set_ylabel(pValueTitle)
for a in axRight.yaxis.majorTicks:
a.tick1On=False
a.tick2On=False
for loc, spine in axRight.spines.iteritems():
spine.set_color('none')
# *** Legend
# *** Legend
if self.legendPos != -1:
legend1 = Rectangle((0, 0), 1, 1, fc=profile1Colour)
legend2 = Rectangle((0, 0), 1, 1, fc=profile2Colour)
legend = self.fig.legend([legend1, legend2], (statsResults.profile.sampleNames[0], statsResults.profile.sampleNames[1]), loc=self.legendPos, ncol=2)
legend.get_frame().set_linewidth(0)
self.updateGeometry()
self.draw()
def show(self,
location='right',
width=0.2,
pad=0.05,
ticks=None,
labels=True,
format=None,
sig_digits=0,
box=None,
box_orientation='vertical',
axis_label_text=None,
axis_label_rotation=None,
axis_label_pad=5):
'''
Show a colorbar on the side of the image.
Parameters
----------
location : str, optional
Where to place the colorbar. Should be one of 'left', 'right', 'top', 'bottom'.
width : float, optional
The width of the colorbar relative to the canvas size.
pad : float, optional
The spacing between the colorbar and the image relative to the canvas size.
ticks : list, optional
The position of the ticks on the colorbar.
labels : bool, optional
Whether to show numerical labels.
format : str, optional
Format of the colorbar tick labels. Can be 'exp' (exponential
notation) or 'sci' (scientific notation). Default is None.
sig_digits: int, optional
Number of significant digits if format is set to 'sci'.
box : list, optional
A custom box within which to place the colorbar. This should
be in the form [xmin, ymin, dx, dy] and be in relative figure
units. This overrides the location argument.
box_orientation str, optional
The orientation of the colorbar within the box. Can be
'horizontal' or 'vertical'
axis_label_text str, optional
Optional text label of the colorbar.
'''
self._base_settings['location'] = location
self._base_settings['width'] = width
self._base_settings['pad'] = pad
self._base_settings['ticks'] = ticks
self._base_settings['labels'] = labels
self._base_settings['format'] = format
self._base_settings['sig_digits'] = sig_digits
self._base_settings['box'] = box
self._base_settings['box_orientation'] = box_orientation
self._base_settings['axis_label_text'] = axis_label_text
self._base_settings['axis_label_rotation'] = axis_label_rotation
self._base_settings['axis_label_pad'] = axis_label_pad
if self._parent.image:
if self._colorbar_axes:
self._parent._figure.delaxes(self._colorbar_axes)
if box is None:
divider = make_axes_locatable(self._parent._ax1)
if location == 'right':
self._colorbar_axes = divider.new_horizontal(
size=width, pad=pad, axes_class=maxes.Axes)
orientation = 'vertical'
elif location == 'top':
self._colorbar_axes = divider.new_vertical(
size=width, pad=pad, axes_class=maxes.Axes)
orientation = 'horizontal'
elif location == 'left':
warnings.warn("Left colorbar not fully implemented")
self._colorbar_axes = divider.new_horizontal(
size=width,
pad=pad,
pack_start=True,
axes_class=maxes.Axes)
locator = divider.new_locator(nx=0, ny=0)
self._colorbar_axes.set_axes_locator(locator)
orientation = 'vertical'
elif location == 'bottom':
warnings.warn("Bottom colorbar not fully implemented")
self._colorbar_axes = divider.new_vertical(
size=width,
pad=pad,
pack_start=True,
axes_class=maxes.Axes)
locator = divider.new_locator(nx=0, ny=0)
self._colorbar_axes.set_axes_locator(locator)
orientation = 'horizontal'
else:
raise Exception("location should be one of: right/top")
self._parent._figure.add_axes(self._colorbar_axes)
else:
self._colorbar_axes = self._parent._figure.add_axes(box)
orientation = box_orientation
if format == 'exp':
format = LogFormatterMathtext()
elif format == 'sci':
format = '%.' + str(sig_digits) + 'e'
else:
format = None
self._colorbar = self._parent._figure.colorbar(
self._parent.image,
cax=self._colorbar_axes,
orientation=orientation,
format=format,
ticks=ticks)
if axis_label_text:
if axis_label_rotation:
self._colorbar.set_label(axis_label_text,
rotation=axis_label_rotation)
else:
self._colorbar.set_label(axis_label_text)
if location == 'right':
for tick in self._colorbar_axes.yaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = False
tick.label2On = labels
self._colorbar_axes.yaxis.set_label_position('right')
self._colorbar_axes.yaxis.labelpad = axis_label_pad
elif location == 'top':
for tick in self._colorbar_axes.xaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = False
tick.label2On = labels
self._colorbar_axes.xaxis.set_label_position('top')
self._colorbar_axes.xaxis.labelpad = axis_label_pad
elif location == 'left':
for tick in self._colorbar_axes.yaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = labels
tick.label2On = False
self._colorbar_axes.yaxis.set_label_position('left')
self._colorbar_axes.yaxis.labelpad = axis_label_pad
elif location == 'bottom':
for tick in self._colorbar_axes.xaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = labels
tick.label2On = False
self._colorbar_axes.xaxis.set_label_position('bottom')
self._colorbar_axes.xaxis.labelpad = axis_label_pad
else:
warnings.warn(
"No image is shown, therefore, no colorbar will be plotted")
def show(self, location='right', width=0.2, pad=0.05, ticks=None, labels=True, box=None, box_orientation='vertical'):
'''
Show a colorbar on the side of the image.
Optional Keyword Arguments:
*location*: [ string ]
Where to place the colorbar. Should be one of 'left', 'right', 'top', 'bottom'.
*width*: [ float ]
The width of the colorbar relative to the canvas size.
*pad*: [ float ]
The spacing between the colorbar and the image relative to the canvas size.
*ticks*: [ None or list ]
The position of the ticks on the colorbar.
*labels*: [ True or False ]
Whether to show numerical labels.
*box*: [ list ]
A custom box within which to place the colorbar. This should
be in the form [xmin, ymin, dx, dy] and be in relative figure
units. This overrides the location argument.
*box_orientation* [ str ]
The orientation of the colorbar within the box. Can be
'horizontal' or 'vertical'
'''
self._base_settings['location'] = location
self._base_settings['width'] = width
self._base_settings['pad'] = pad
self._base_settings['ticks'] = ticks
self._base_settings['labels'] = labels
self._base_settings['box'] = box
self._base_settings['box_orientation'] = box_orientation
if self._parent.image:
if self._colorbar_axes:
self._parent._figure.delaxes(self._colorbar_axes)
if box is None:
divider = make_axes_locatable(self._parent._ax1)
if location == 'right':
self._colorbar_axes = divider.new_horizontal(size=width, pad=pad, axes_class=maxes.Axes)
orientation = 'vertical'
elif location == 'top':
self._colorbar_axes = divider.new_vertical(size=width, pad=pad, axes_class=maxes.Axes)
orientation = 'horizontal'
elif location == 'left':
warnings.warn("Left colorbar not fully implemented")
self._colorbar_axes = divider.new_horizontal(size=width, pad=pad, pack_start=True, axes_class=maxes.Axes)
locator = divider.new_locator(nx=0, ny=0)
self._colorbar_axes.set_axes_locator(locator)
orientation = 'vertical'
elif location == 'bottom':
warnings.warn("Bottom colorbar not fully implemented")
self._colorbar_axes = divider.new_vertical(size=width, pad=pad, pack_start=True, axes_class=maxes.Axes)
locator = divider.new_locator(nx=0, ny=0)
self._colorbar_axes.set_axes_locator(locator)
orientation = 'horizontal'
else:
raise Exception("location should be one of: right/top")
self._parent._figure.add_axes(self._colorbar_axes)
else:
self._colorbar_axes = self._parent._figure.add_axes(box)
orientation = box_orientation
self._colorbar = self._parent._figure.colorbar(self._parent.image, cax=self._colorbar_axes, orientation=orientation, ticks=ticks)
if location == 'right':
for tick in self._colorbar_axes.yaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = False
tick.label2On = labels
elif location == 'top':
for tick in self._colorbar_axes.xaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = False
tick.label2On = labels
elif location == 'left':
for tick in self._colorbar_axes.yaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = labels
tick.label2On = False
elif location == 'bottom':
for tick in self._colorbar_axes.xaxis.get_major_ticks():
tick.tick1On = True
tick.tick2On = True
tick.label1On = labels
tick.label2On = False
else:
warnings.warn("No image is shown, therefore, no colorbar will be plotted")
def save_pca(odir, filename, stk, anlz, png, pdf, svg):
SaveFileName = os.path.join(odir,filename)
from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
matplotlib.rcParams['pdf.fonttype'] = 42
#Save evals
evalmax = np.min([stk.n_ev, 40])
pcaevals = anlz.eigenvals[0:evalmax]
fig = matplotlib.figure.Figure(figsize =(PlotW, PlotH))
canvas = FigureCanvas(fig)
fig.clf()
fig.add_axes((0.15,0.15,0.75,0.75))
axes = fig.gca()
evalsplot = axes.semilogy(np.arange(1,evalmax+1), pcaevals,'b.')
axes.set_xlabel('Principal Component')
axes.set_ylabel('Log(Eigenvalue)')
if png == 1:
ext = 'png'
fileName_evals = SaveFileName+"_PCAevals."+ext
fig.savefig(fileName_evals, bbox_inches='tight', pad_inches = 0.0)
if pdf == 1:
ext = 'pdf'
fileName_evals = SaveFileName+"_PCAevals."+ext
fig.savefig(fileName_evals, bbox_inches='tight', pad_inches = 0.0)
if svg == 1:
ext = 'svg'
fileName_evals = SaveFileName+"_PCAevals."+ext
fig.savefig(fileName_evals, bbox_inches='tight', pad_inches = 0.0)
for i in range(10):
pcaimage = anlz.pcaimages[:,:,i]
fig = matplotlib.figure.Figure(figsize =(PlotH*1.15, PlotH))
canvas = FigureCanvas(fig)
axes = fig.gca()
divider = make_axes_locatable(axes)
ax_cb = divider.new_horizontal(size="3%", pad=0.03)
fig.add_axes(ax_cb)
axes.set_position([0.03,0.03,0.8,0.94])
bound = anlz.pcaimagebounds[i]
im = axes.imshow(pcaimage, cmap=matplotlib.cm.get_cmap("seismic_r"), vmin = -bound, vmax = bound)
cbar = axes.figure.colorbar(im, orientation='vertical',cax=ax_cb)
axes.axis("off")
if png == 1:
ext = 'png'
fileName_img = SaveFileName+"_PCA_" +str(i+1)+"."+ext
fig.savefig(fileName_img, bbox_inches='tight', pad_inches = 0.0)
if pdf == 1:
ext = 'pdf'
fileName_img = SaveFileName+"_PCA_" +str(i+1)+"."+ext
fig.savefig(fileName_img, bbox_inches='tight', pad_inches = 0.0)
if svg == 1:
ext = 'svg'
fileName_img = SaveFileName+"_PCA_" +str(i+1)+"."+ext
fig.savefig(fileName_img, bbox_inches='tight', pad_inches = 0.0)
for i in range(10):
pcaspectrum = anlz.eigenvecs[:,i]
fig = matplotlib.figure.Figure(figsize =(PlotW, PlotH))
canvas = FigureCanvas(fig)
fig.add_axes((0.15,0.15,0.75,0.75))
axes = fig.gca()
specplot = axes.plot(stk.ev, pcaspectrum)
axes.set_xlabel('Photon Energy [eV]')
axes.set_ylabel('Optical Density')
if png == 1:
ext = 'png'
fileName_spec = SaveFileName+"_PCAspectrum_" +str(i+1)+"."+ext
fig.savefig(fileName_spec)
if pdf == 1:
ext = 'pdf'
fileName_spec = SaveFileName+"_PCAspectrum_" +str(i+1)+"."+ext
fig.savefig(fileName_spec)
if svg == 1:
ext = 'svg'
fileName_spec = SaveFileName+"_PCAspectrum_" +str(i+1)+"."+ext
fig.savefig(fileName_spec)
for i in range (10):
pcaspectrum = anlz.eigenvecs[:,i]
fileName_spec = SaveFileName+"_PCAspectrum_" +str(i+1)+".csv"
cname = "PCAspectrum_" +str(i+1)
stk.write_csv(fileName_spec, stk.ev, pcaspectrum, cname=cname)
return
def drawmatrix_channels(in_m, channel_names=None, fig=None, x_tick_rot=0,
size=None, cmap=plt.cm.RdBu_r, colorbar=True,
color_anchor=None, title=None):
r"""Creates a lower-triangle of the matrix of an nxn set of values. This is
the typical format to show a symmetrical bivariate quantity (such as
correlation or coherence between two different ROIs).
Parameters
----------
in_m: nxn array with values of relationships between two sets of rois or
channels
channel_names (optional): list of strings with the labels to be applied to
the channels in the input. Defaults to '0','1','2', etc.
fig (optional): a matplotlib figure
cmap (optional): a matplotlib colormap to be used for displaying the values
of the connections on the graph
title (optional): string to title the figure (can be like '$\alpha$')
color_anchor (optional): determine the mapping from values to colormap
if None, min and max of colormap correspond to min and max of in_m
if 0, min and max of colormap correspond to max of abs(in_m)
if (a,b), min and max of colormap correspond to (a,b)
Returns
-------
fig: a figure object
"""
N = in_m.shape[0]
ind = np.arange(N) # the evenly spaced plot indices
def channel_formatter(x, pos=None):
thisind = np.clip(int(x), 0, N - 1)
return channel_names[thisind]
if fig is None:
fig = plt.figure()
if size is not None:
fig.set_figwidth(size[0])
fig.set_figheight(size[1])
w = fig.get_figwidth()
h = fig.get_figheight()
ax_im = fig.add_subplot(1, 1, 1)
# If you want to draw the colorbar:
if colorbar:
divider = make_axes_locatable(ax_im)
ax_cb = divider.new_vertical(size="10%", pad=0.1, pack_start=True)
fig.add_axes(ax_cb)
# Make a copy of the input, so that you don't make changes to the original
# data provided
m = in_m.copy()
# Null the upper triangle, so that you don't get the redundant and the
# diagonal values:
idx_null = triu_indices(m.shape[0])
m[idx_null] = np.nan
# Extract the minimum and maximum values for scaling of the
# colormap/colorbar:
max_val = np.nanmax(m)
min_val = np.nanmin(m)
if color_anchor is None:
color_min = min_val
color_max = max_val
elif color_anchor == 0:
bound = max(abs(max_val), abs(min_val))
color_min = -bound
color_max = bound
else:
color_min = color_anchor[0]
color_max = color_anchor[1]
# The call to imshow produces the matrix plot:
im = ax_im.imshow(m, origin='upper', interpolation='nearest',
vmin=color_min, vmax=color_max, cmap=cmap)
# Formatting:
ax = ax_im
ax.grid(True)
# Label each of the cells with the row and the column:
if channel_names is not None:
for i in range(0, m.shape[0]):
if i < (m.shape[0] - 1):
ax.text(i - 0.3, i, channel_names[i], rotation=x_tick_rot)
if i > 0:
ax.text(-1, i + 0.3, channel_names[i],
horizontalalignment='right')
ax.set_axis_off()
ax.set_xticks(np.arange(N))
ax.xaxis.set_major_formatter(ticker.FuncFormatter(channel_formatter))
fig.autofmt_xdate(rotation=x_tick_rot)
ax.set_yticks(np.arange(N))
ax.set_yticklabels(channel_names)
ax.set_ybound([-0.5, N - 0.5])
ax.set_xbound([-0.5, N - 1.5])
# Make the tick-marks invisible:
for line in ax.xaxis.get_ticklines():
line.set_markeredgewidth(0)
for line in ax.yaxis.get_ticklines():
line.set_markeredgewidth(0)
ax.set_axis_off()
if title is not None:
ax.set_title(title)
# The following produces the colorbar and sets the ticks
if colorbar:
# Set the ticks - if 0 is in the interval of values, set that, as well
# as the maximal and minimal values:
if min_val < 0:
ticks = [color_min, min_val, 0, max_val, color_max]
# Otherwise - only set the minimal and maximal value:
else:
ticks = [color_min, min_val, max_val, color_max]
# This makes the colorbar:
cb = fig.colorbar(im, cax=ax_cb, orientation='horizontal',
cmap=cmap,
norm=im.norm,
boundaries=np.linspace(color_min, color_max, 256),
ticks=ticks,
format='%.2f')
# Set the current figure active axis to be the top-one, which is the one
# most likely to be operated on by users later on
fig.sca(ax)
return fig
def drawmatrix_channels(in_m,
channel_names=None,
fig=None,
x_tick_rot=0,
size=None,
cmap=pl.cm.RdBu_r,
colorbar=True,
color_anchor=None,
title=None):
r"""Creates a lower-triangle of the matrix of an nxn set of values. This is
the typical format to show a symmetrical bivariate quantity (such as
correlation or coherence between two different ROIs).
Parameters
----------
in_m: nxn array with values of relationships between two sets of rois or
channels
channel_names (optional): list of strings with the labels to be applied to
the channels in the input. Defaults to '0','1','2', etc.
fig (optional): a matplotlib figure
cmap (optional): a matplotlib colormap to be used for displaying the values
of the connections on the graph
title (optional): string to title the figure (can be like '$\alpha$')
color_anchor (optional): determine the mapping from values to colormap
if None, min and max of colormap correspond to min and max of in_m
if 0, min and max of colormap correspond to max of abs(in_m)
if (a,b), min and max of colormap correspond to (a,b)
Returns
-------
fig: a figure object
"""
N = in_m.shape[0]
ind = np.arange(N) # the evenly spaced plot indices
def channel_formatter(x, pos=None):
thisind = np.clip(int(x), 0, N - 1)
return channel_names[thisind]
if fig is None:
fig = pl.figure()
if size is not None:
fig.set_figwidth(size[0])
fig.set_figheight(size[1])
w = fig.get_figwidth()
h = fig.get_figheight()
ax_im = fig.add_subplot(1, 1, 1)
#If you want to draw the colorbar:
if colorbar:
divider = make_axes_locatable(ax_im)
ax_cb = divider.new_vertical(size="10%", pad=0.1, pack_start=True)
fig.add_axes(ax_cb)
#Make a copy of the input, so that you don't make changes to the original
#data provided
m = in_m.copy()
#Null the upper triangle, so that you don't get the redundant and the
#diagonal values:
idx_null = triu_indices(m.shape[0])
m[idx_null] = np.nan
#Extract the minimum and maximum values for scaling of the
#colormap/colorbar:
max_val = np.nanmax(m)
min_val = np.nanmin(m)
if color_anchor is None:
color_min = min_val
color_max = max_val
elif color_anchor == 0:
bound = max(abs(max_val), abs(min_val))
color_min = -bound
color_max = bound
else:
color_min = color_anchor[0]
color_max = color_anchor[1]
#The call to imshow produces the matrix plot:
im = ax_im.imshow(m,
origin='upper',
interpolation='nearest',
vmin=color_min,
vmax=color_max,
cmap=cmap)
#Formatting:
ax = ax_im
ax.grid(True)
#Label each of the cells with the row and the column:
if channel_names is not None:
for i in range(0, m.shape[0]):
if i < (m.shape[0] - 1):
ax.text(i - 0.3, i, channel_names[i], rotation=x_tick_rot)
if i > 0:
ax.text(-1,
i + 0.3,
channel_names[i],
horizontalalignment='right')
ax.set_axis_off()
ax.set_xticks(np.arange(N))
ax.xaxis.set_major_formatter(ticker.FuncFormatter(channel_formatter))
fig.autofmt_xdate(rotation=x_tick_rot)
ax.set_yticks(np.arange(N))
ax.set_yticklabels(channel_names)
ax.set_ybound([-0.5, N - 0.5])
ax.set_xbound([-0.5, N - 1.5])
#Make the tick-marks invisible:
for line in ax.xaxis.get_ticklines():
line.set_markeredgewidth(0)
for line in ax.yaxis.get_ticklines():
line.set_markeredgewidth(0)
ax.set_axis_off()
if title is not None:
ax.set_title(title)
#The following produces the colorbar and sets the ticks
if colorbar:
#Set the ticks - if 0 is in the interval of values, set that, as well
#as the maximal and minimal values:
if min_val < 0:
ticks = [color_min, min_val, 0, max_val, color_max]
#Otherwise - only set the minimal and maximal value:
else:
ticks = [color_min, min_val, max_val, color_max]
#This makes the colorbar:
cb = fig.colorbar(im,
cax=ax_cb,
orientation='horizontal',
cmap=cmap,
norm=im.norm,
boundaries=np.linspace(color_min, color_max, 256),
ticks=ticks,
format='%.2f')
# Set the current figure active axis to be the top-one, which is the one
# most likely to be operated on by users later on
fig.sca(ax)
return fig
def plot_ica_components(ica, picks=None, ch_type='mag', res=64,
layout=None, vmin=None, vmax=None, cmap='RdBu_r',
sensors=True, colorbar=False, title=None,
show=True, outlines='head', contours=6,
image_interp='bilinear'):
"""Project unmixing matrix on interpolated sensor topogrpahy.
Parameters
----------
ica : instance of mne.preprocessing.ICA
The ICA solution.
picks : int | array-like | None
The indices of the sources to be plotted.
If None all are plotted in batches of 20.
ch_type : 'mag' | 'grad' | 'planar1' | 'planar2' | 'eeg'
The channel type to plot. For 'grad', the gradiometers are
collected in pairs and the RMS for each pair is plotted.
layout : None | Layout
Layout instance specifying sensor positions (does not need to
be specified for Neuromag data). If possible, the correct layout is
inferred from the data.
vmin : float | callable
The value specfying the lower bound of the color range.
If None, and vmax is None, -vmax is used. Else np.min(data).
If callable, the output equals vmin(data).
vmax : float | callable
The value specfying the upper bound of the color range.
If None, the maximum absolute value is used. If vmin is None,
but vmax is not, defaults to np.min(data).
If callable, the output equals vmax(data).
cmap : matplotlib colormap
Colormap.
sensors : bool | str
Add markers for sensor locations to the plot. Accepts matplotlib
plot format string (e.g., 'r+' for red plusses). If True, a circle
will be used (via .add_artist). Defaults to True.
colorbar : bool
Plot a colorbar.
res : int
The resolution of the topomap image (n pixels along each side).
show : bool
Call pyplot.show() at the end.
outlines : 'head' | dict | None
The outlines to be drawn. If 'head', a head scheme will be drawn.
If dict, each key refers to a tuple of x and y positions. The
values in 'mask_pos' will serve as image mask. If None,
nothing will be drawn. defaults to 'head'.
contours : int | False | None
The number of contour lines to draw. If 0, no contours will be drawn.
image_interp : str
The image interpolation to be used. All matplotlib options are
accepted.
Returns
-------
fig : instance of matplotlib.pyplot.Figure or list
The figure object(s).
"""
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid import make_axes_locatable
if picks is None: # plot components by sets of 20
n_components = ica.mixing_matrix_.shape[1]
p = 20
figs = []
for k in range(0, n_components, p):
picks = range(k, min(k + p, n_components))
fig = plot_ica_components(ica, picks=picks,
ch_type=ch_type, res=res, layout=layout,
vmax=vmax, cmap=cmap, sensors=sensors,
colorbar=colorbar, title=title,
show=show, outlines=outlines,
contours=contours,
image_interp=image_interp)
figs.append(fig)
return figs
elif np.isscalar(picks):
picks = [picks]
data = np.dot(ica.mixing_matrix_[:, picks].T,
ica.pca_components_[:ica.n_components_])
if ica.info is None:
raise RuntimeError('The ICA\'s measurement info is missing. Please '
'fit the ICA or add the corresponding info object.')
data_picks, pos, merge_grads, names, _ = _prepare_topo_plot(ica, ch_type,
layout)
pos, outlines = _check_outlines(pos, outlines)
if outlines not in (None, 'head'):
image_mask, pos = _make_image_mask(outlines, pos, res)
else:
image_mask = None
data = np.atleast_2d(data)
data = data[:, data_picks]
# prepare data for iteration
fig, axes = _prepare_trellis(len(data), max_col=5)
if title is None:
title = 'ICA components'
fig.suptitle(title)
if merge_grads:
from ..channels.layout import _merge_grad_data
for ii, data_, ax in zip(picks, data, axes):
ax.set_title('IC #%03d' % ii, fontsize=12)
data_ = _merge_grad_data(data_) if merge_grads else data_
vmin_, vmax_ = _setup_vmin_vmax(data_, vmin, vmax)
im = plot_topomap(data_.flatten(), pos, vmin=vmin_, vmax=vmax_,
res=res, axis=ax, cmap=cmap, outlines=outlines,
image_mask=image_mask, contours=contours,
image_interp=image_interp)[0]
if colorbar:
divider = make_axes_locatable(ax)
cax = divider.append_axes("right", size="5%", pad=0.05)
cbar = plt.colorbar(im, cax=cax, format='%3.2f', cmap=cmap)
cbar.ax.tick_params(labelsize=12)
cbar.set_ticks((vmin_, vmax_))
cbar.ax.set_title('AU', fontsize=10)
ax.set_yticks([])
ax.set_xticks([])
ax.set_frame_on(False)
tight_layout(fig=fig)
fig.subplots_adjust(top=0.95)
fig.canvas.draw()
if show is True:
plt.show()
return fig
def plot(net, n, p):
classname = net.__class__.__name__
axes.set_xticks([])
axes.set_yticks([])
divider = make_axes_locatable(axes)
subaxes = divider.new_vertical(1.0, pad=0.4, sharex=axes)
fig.add_axes(subaxes)
subaxes.set_xticks([])
subaxes.yaxis.set_major_locator(matplotlib.ticker.MaxNLocator(2))
subaxes.yaxis.set_ticks_position('right')
subaxes.set_ylabel('Distortion')
subaxes.set_xlabel('Time')
Y = net.distortion[::1]
X = np.arange(len(Y))/float(len(Y)-1)
subaxes.plot(X,Y)
if classname == 'NG':
plt.title('Neural Gas', fontsize=20)
elif classname == 'SOM':
plt.title('Self-Organizing Map', fontsize=20)
elif classname == 'DSOM':
plt.title('Dynamic Self-Organizing Map', fontsize=20)
axes.axis([0,1,0,1])
axes.set_aspect(1)
bounds = divider.locate(0,0).bounds
grid = AxesGrid(fig, bounds, nrows_ncols = (n,n), axes_pad = 0.05, label_mode = "1")
for row in range(n):
for col in range(n):
index = row*n+col
Z = net.codebook[row,col].reshape(p,p)
im = grid[index].imshow(Z, interpolation = 'nearest', vmin=0, vmax=1, cmap=plt.cm.gray)
grid[index].set_yticks([])
grid[index].set_xticks([])
classname = net.__class__.__name__
if classname == 'NG':
axes.text(0.5, -0.01,
r'$\lambda_i = %.3f,\lambda_f = %.3f, \varepsilon_i=%.3f, \varepsilon_f=%.3f$' % (
net.sigma_i, net.sigma_f, net.lrate_i, net.lrate_f),
fontsize=16,
horizontalalignment='center',
verticalalignment='top',
transform = axes.transAxes)
if classname == 'SOM':
axes.text(0.5, -0.01,
r'$\sigma_i = %.3f,\sigma_f = %.3f, \varepsilon_i=%.3f, \varepsilon_f=%.3f$' % (
net.sigma_i, net.sigma_f, net.lrate_i, net.lrate_f),
fontsize=16,
horizontalalignment='center',
verticalalignment='top',
transform = axes.transAxes)
elif classname == 'DSOM':
axes.text(0.5, -0.01,
r'$elasticity = %.2f$, $\varepsilon = %.3f$' % (net.elasticity, net.lrate),
fontsize=16,
horizontalalignment='center',
verticalalignment='top',
transform = axes.transAxes)
def Scatter_hist(data1, data2, data1b=False, data2b=False, xbinwidth = 0.5,
ybinwidth=0.5, SSE=None, SSEb=None, vmax=1000., colormaps=cm.YlOrBr,
cleanstyle = False, roodlichter=0.5, *args, **kwargs):
'''
Three-parts plot with the two parameter sitdirbutions plotted on the sides
Parameters
-----------
data1: ndarray
dataset 1 for x-axis
data2: ndarray
dataset 2 for y-axis
data1b: ndarray
dataset to plot along the data 1 set
data2b: ndarray
dataset to plot along the data 2 set
binwidth: float
defines the width of the bins relative to the data used
cleanstyle: bool True|False
if True, a more minimalistic version of the plot is given
*args, **kwargs: args
arguments given toi the scatter plot
example: s=15, marker='o', edgecolors= 'k',facecolor = 'white'
Returns
---------
fig: matplotlib.figure.Figure object
the resulting figure
axScatter: matplotlib.axes.AxesSubplot object
the scatter plot with the datapoints, can be used to add labels or
change the current ticks settings
axHistx: matplotlib.axes.AxesSubplot object
the x-axis histogram
axHisty: matplotlib.axes.AxesSubplot object
the y-axis histogram
Examples
----------
>>> nMC = 1000
>>> parval1 = np.random.gamma(5.5,size=nMC)
>>> parval2 = np.random.gamma(8.0,size=nMC)
>>> parnames = ['par1','par2']
>>> fig,axScatter,axHistx,axHisty = Scatter_hist(parval1,parval2,
cleanstyle = True, s=48,
marker='o', edgecolors= 'k',
facecolor = 'none',alpha=0.7)
>>> parval1b = np.random.uniform(low=0.0, high=30.0,size=nMC)
>>> parval2b = np.random.uniform(low=0.0, high=30.0,size=nMC)
>>> fig,axScatter,axHistx,axHisty = Scatter_hist(parval1,parval2,parval1b,
parval2b, cleanstyle = True,
s=48, marker='o',
edgecolors= 'k',
facecolor = 'none',
alpha=0.7)
Notes
------
Typical application is to check the dependency of two posterior
parameter distrbutions, eventually compared with their selected posteriors
If a second dataset is added to compare, the style options of the scatter
plot are fixed and the *args, **kwargs have no influence
'''
if not isinstance(data1, np.ndarray):
raise Exception('dataset 1 need to be numpy ndarray')
if not isinstance(data2, np.ndarray):
raise Exception('dataset 2 need to be numpy ndarray')
if isinstance(data1b, np.ndarray):
if not isinstance(data2b, np.ndarray):
raise Exception('Always combine the data of both')
if isinstance(data2b, np.ndarray):
if not isinstance(data1b, np.ndarray):
raise Exception('Always combine the data of both')
fig = plt.figure(figsize=(10,10))
axScatter = plt.subplot(111)
divider = make_axes_locatable(axScatter)
#axScatter.set_aspect('equal')
axScatter.set_autoscale_on(True)
# create a new axes with above the axScatter
axHistx = divider.new_vertical(1.5, pad=0.0001, sharex=axScatter)
# create a new axes on the right side of the
# axScatter
axHisty = divider.new_horizontal(1.5, pad=0.0001, sharey=axScatter)
fig.add_axes(axHistx)
fig.add_axes(axHisty)
# now determine nice limits by hand:
# binwidth = binwidth
xmin = np.min(data1)
xmax = np.max(data1)
ymin = np.min(data2)
ymax = np.max(data2)
#xymax = np.max( [np.max(np.fabs(data1)), np.max(np.fabs(data2))] )
#lim = (int(xymax/binwidth) + 1) * binwidth
binsx = np.arange(xmin, xmax + xbinwidth, xbinwidth)
binsy = np.arange(ymin, ymax + ybinwidth, ybinwidth)
#bins = np.arange(-lim, lim + binwidth, binwidth)
# the scatter plot:
if isinstance(data1b, np.ndarray): #TWO DATA ENTRIES
if SSE == None:
print '*args, **kwargs do not have any influcence when using two\
options'
axScatter.scatter(data1, data2, facecolor = 'none',
edgecolor='k',s=25)
axScatter.scatter(data1b, data2b, facecolor='none',
edgecolor='grey',s=25)
xminb = np.min(data1b)
xmaxb = np.max(data1b)
yminb = np.min(data2b)
ymaxb = np.max(data2b)
binsxb = np.arange(xminb, xmaxb + xbinwidth, xbinwidth)
binsyb = np.arange(yminb, ymaxb + ybinwidth, ybinwidth)
axHistx.hist(data1b, bins=binsxb, edgecolor='None',
color='grey',normed=True)
axHisty.hist(data2b, bins=binsyb, orientation='horizontal',
edgecolor='None', color='grey', normed=True)
axHistx.hist(data1, bins=binsx, edgecolor='None',
color='k', normed=True)
axHisty.hist(data2, bins=binsy, orientation='horizontal',
edgecolor='None', color='k', normed=True)
else:
print '*args, **kwargs do not have any influcence when using two\
options'
sc1 = axScatter.scatter(data1b, data2b, c=SSEb, vmax=vmax,alpha=roodlichter,
edgecolors= 'none', cmap = colormaps, *args, **kwargs)
axScatter.scatter(data1, data2, c=SSE, vmax=vmax,
edgecolors= 'none', cmap = colormaps, *args, **kwargs)
xminb = np.min(data1b)
xmaxb = np.max(data1b)
yminb = np.min(data2b)
ymaxb = np.max(data2b)
binsxb = np.arange(xminb, xmaxb + xbinwidth, xbinwidth)
binsyb = np.arange(yminb, ymaxb + ybinwidth, ybinwidth)
axHistx.hist(data1b, bins=binsxb, edgecolor='None',
color=colormaps(1.),normed=True)
axHisty.hist(data2b, bins=binsyb, orientation='horizontal', color=colormaps(1.),
edgecolor='None', normed=True)
axHistx.hist(data1, bins=binsx, edgecolor='None',
color=colormaps(0.), normed=True)
axHisty.hist(data2, bins=binsy, orientation='horizontal',
edgecolor='None', color=colormaps(0.), normed=True)
else: #ONLY ONE DATA1 and DATA2
if SSE == None:
axScatter.scatter(data1, data2, c= 'black', *args, **kwargs)
axHistx.hist(data1, bins=binsx, edgecolor='None', color='k')
axHisty.hist(data2, bins=binsy, orientation='horizontal',
edgecolor='None', color='k')
else:
axScatter.scatter(data1, data2, c=SSE, vmax=vmax,
edgecolors= 'none', cmap = colormaps, *args, **kwargs)
axHistx.hist(data1, bins=binsx, edgecolor='None', color='k')
axHisty.hist(data2, bins=binsy, orientation='horizontal',
edgecolor='None', color='k')
# the xaxis of axHistx and yaxis of axHisty are shared with axScatter,
# thus there is no need to manually adjust the xlim and ylim of these
# axis.
majloc1 = MaxNLocator(nbins=4, prune='lower')
axScatter.yaxis.set_major_locator(majloc1)
majloc2 = MaxNLocator(nbins=4)
axScatter.xaxis.set_major_locator(majloc2)
axScatter.grid(linestyle = 'dashed', color = '0.75',linewidth = 1.)
axScatter.set_axisbelow(True)
axHisty.set_axisbelow(True)
axHistx.set_axisbelow(True)
# The 'clean' environment
if cleanstyle == True:
plt.setp(axHistx.get_xticklabels() + axHisty.get_yticklabels(),
visible=False)
plt.setp(axHistx.get_yticklabels() + axHisty.get_xticklabels(),
visible=False)
axHisty.set_xticks([])
axHistx.set_yticks([])
axHistx.xaxis.set_ticks_position('bottom')
axHisty.yaxis.set_ticks_position('left')
axHisty.spines['right'].set_color('none')
axHisty.spines['top'].set_color('none')
axHisty.spines['bottom'].set_color('none')
axHisty.spines['left'].set_color('none')
axHistx.spines['top'].set_color('none')
axHistx.spines['right'].set_color('none')
axHistx.spines['left'].set_color('none')
axHistx.spines['bottom'].set_color('none')
axScatter.spines['top'].set_color('none')
axScatter.spines['right'].set_color('none')
else:
plt.setp(axHistx.get_xticklabels() + axHisty.get_yticklabels(),
visible=False)
for tl in axHisty.get_yticklabels():
tl.set_visible(False)
for tlp in axHisty.get_xticklabels():
tlp.set_rotation(-90)
majloc3 = MaxNLocator(nbins=4, prune='lower')
axHistx.yaxis.set_major_locator(majloc3)
axHistx.yaxis.tick_right()
majloc4 = MaxNLocator(nbins=4, prune='lower')
axHisty.xaxis.set_major_locator(majloc4)
axHisty.xaxis.tick_top()
axHisty.yaxis.grid(linestyle = 'dashed', color = '0.75',linewidth = 1.)
axHistx.xaxis.grid(linestyle = 'dashed', color = '0.75',linewidth = 1.)
return fig,axScatter,axHistx,axHisty
def plot_ica_components(ica,
picks=None,
ch_type='mag',
res=64,
layout=None,
vmin=None,
vmax=None,
cmap='RdBu_r',
sensors=True,
colorbar=False,
title=None,
show=True,
outlines='head',
contours=6,
image_interp='bilinear'):
"""Project unmixing matrix on interpolated sensor topogrpahy.
Parameters
----------
ica : instance of mne.preprocessing.ICA
The ICA solution.
picks : int | array-like | None
The indices of the sources to be plotted.
If None all are plotted in batches of 20.
ch_type : 'mag' | 'grad' | 'planar1' | 'planar2' | 'eeg'
The channel type to plot. For 'grad', the gradiometers are
collected in pairs and the RMS for each pair is plotted.
layout : None | Layout
Layout instance specifying sensor positions (does not need to
be specified for Neuromag data). If possible, the correct layout is
inferred from the data.
vmin : float | callable
The value specfying the lower bound of the color range.
If None, and vmax is None, -vmax is used. Else np.min(data).
If callable, the output equals vmin(data).
vmax : float | callable
The value specfying the upper bound of the color range.
If None, the maximum absolute value is used. If vmin is None,
but vmax is not, defaults to np.min(data).
If callable, the output equals vmax(data).
cmap : matplotlib colormap
Colormap.
sensors : bool | str
Add markers for sensor locations to the plot. Accepts matplotlib
plot format string (e.g., 'r+' for red plusses). If True, a circle
will be used (via .add_artist). Defaults to True.
colorbar : bool
Plot a colorbar.
res : int
The resolution of the topomap image (n pixels along each side).
show : bool
Call pyplot.show() at the end.
outlines : 'head' | dict | None
The outlines to be drawn. If 'head', a head scheme will be drawn.
If dict, each key refers to a tuple of x and y positions. The
values in 'mask_pos' will serve as image mask. If None,
nothing will be drawn. defaults to 'head'.
contours : int | False | None
The number of contour lines to draw. If 0, no contours will be drawn.
image_interp : str
The image interpolation to be used. All matplotlib options are
accepted.
Returns
-------
fig : instance of matplotlib.pyplot.Figure or list
The figure object(s).
"""
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid import make_axes_locatable
if picks is None: # plot components by sets of 20
n_components = ica.mixing_matrix_.shape[1]
p = 20
figs = []
for k in range(0, n_components, p):
picks = range(k, min(k + p, n_components))
fig = plot_ica_components(ica,
picks=picks,
ch_type=ch_type,
res=res,
layout=layout,
vmax=vmax,
cmap=cmap,
sensors=sensors,
colorbar=colorbar,
title=title,
show=show,
outlines=outlines,
contours=contours,
image_interp=image_interp)
figs.append(fig)
return figs
elif np.isscalar(picks):
picks = [picks]
data = np.dot(ica.mixing_matrix_[:, picks].T,
ica.pca_components_[:ica.n_components_])
if ica.info is None:
raise RuntimeError('The ICA\'s measurement info is missing. Please '
'fit the ICA or add the corresponding info object.')
data_picks, pos, merge_grads, names, _ = _prepare_topo_plot(
ica, ch_type, layout)
pos, outlines = _check_outlines(pos, outlines)
if outlines not in (None, 'head'):
image_mask, pos = _make_image_mask(outlines, pos, res)
else:
image_mask = None
data = np.atleast_2d(data)
data = data[:, data_picks]
# prepare data for iteration
fig, axes = _prepare_trellis(len(data), max_col=5)
if title is None:
title = 'ICA components'
fig.suptitle(title)
if merge_grads:
from ..channels.layout import _merge_grad_data
for ii, data_, ax in zip(picks, data, axes):
ax.set_title('IC #%03d' % ii, fontsize=12)
data_ = _merge_grad_data(data_) if merge_grads else data_
vmin_, vmax_ = _setup_vmin_vmax(data_, vmin, vmax)
im = plot_topomap(data_.flatten(),
pos,
vmin=vmin_,
vmax=vmax_,
res=res,
axis=ax,
cmap=cmap,
outlines=outlines,
image_mask=image_mask,
contours=contours,
image_interp=image_interp)[0]
if colorbar:
divider = make_axes_locatable(ax)
cax = divider.append_axes("right", size="5%", pad=0.05)
cbar = plt.colorbar(im, cax=cax, format='%3.2f', cmap=cmap)
cbar.ax.tick_params(labelsize=12)
cbar.set_ticks((vmin_, vmax_))
cbar.ax.set_title('AU', fontsize=10)
ax.set_yticks([])
ax.set_xticks([])
ax.set_frame_on(False)
tight_layout(fig=fig)
fig.subplots_adjust(top=0.95)
fig.canvas.draw()
if show is True:
plt.show()
return fig
def plot_tfr_topomap(tfr,
tmin=None,
tmax=None,
fmin=None,
fmax=None,
ch_type='mag',
baseline=None,
mode='mean',
layout=None,
vmax=None,
vmin=None,
cmap='RdBu_r',
sensors=True,
colorbar=True,
unit=None,
res=64,
size=2,
format='%1.1e',
show_names=False,
title=None,
axes=None,
show=True):
"""Plot topographic maps of specific time-frequency intervals of TFR data
Parameters
----------
tfr : AvereageTFR
The AvereageTFR object.
tmin : None | float
The first time instant to display. If None the first time point
available is used.
tmax : None | float
The last time instant to display. If None the last time point
available is used.
fmin : None | float
The first frequency to display. If None the first frequency
available is used.
fmax : None | float
The last frequency to display. If None the last frequency
available is used.
ch_type : 'mag' | 'grad' | 'planar1' | 'planar2' | 'eeg'
The channel type to plot. For 'grad', the gradiometers are
collected in pairs and the RMS for each pair is plotted.
baseline : tuple or list of length 2
The time interval to apply rescaling / baseline correction.
If None do not apply it. If baseline is (a, b)
the interval is between "a (s)" and "b (s)".
If a is None the beginning of the data is used
and if b is None then b is set to the end of the interval.
If baseline is equal to (None, None) all the time
interval is used.
mode : 'logratio' | 'ratio' | 'zscore' | 'mean' | 'percent'
Do baseline correction with ratio (power is divided by mean
power during baseline) or z-score (power is divided by standard
deviation of power during baseline after subtracting the mean,
power = [power - mean(power_baseline)] / std(power_baseline))
If None, baseline no correction will be performed.
layout : None | Layout
Layout instance specifying sensor positions (does not need to
be specified for Neuromag data). If possible, the correct layout
file is inferred from the data; if no appropriate layout file
was found, the layout is automatically generated from the sensor
locations.
vmin : float | callable
The value specfying the lower bound of the color range.
If None, and vmax is None, -vmax is used. Else np.min(data).
If callable, the output equals vmin(data).
vmax : float | callable
The value specfying the upper bound of the color range.
If None, the maximum absolute value is used. If vmin is None,
but vmax is not, defaults to np.min(data).
If callable, the output equals vmax(data).
cmap : matplotlib colormap
Colormap. For magnetometers and eeg defaults to 'RdBu_r', else
'Reds'.
sensors : bool | str
Add markers for sensor locations to the plot. Accepts matplotlib
plot format string (e.g., 'r+' for red plusses). If True, a circle will
be used (via .add_artist). Defaults to True.
colorbar : bool
Plot a colorbar.
unit : str | None
The unit of the channel type used for colorbar labels.
res : int
The resolution of the topomap image (n pixels along each side).
size : float
Side length per topomap in inches.
format : str
String format for colorbar values.
show_names : bool | callable
If True, show channel names on top of the map. If a callable is
passed, channel names will be formatted using the callable; e.g., to
delete the prefix 'MEG ' from all channel names, pass the function
lambda x: x.replace('MEG ', ''). If `mask` is not None, only
significant sensors will be shown.
title : str | None
Title. If None (default), no title is displayed.
axes : instance of Axis | None
The axes to plot to. If None the axes is defined automatically.
show : bool
Call pyplot.show() at the end.
Returns
-------
fig : matplotlib.figure.Figure
The figure containing the topography.
"""
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import make_axes_locatable
picks, pos, merge_grads, names, _ = _prepare_topo_plot(
tfr, ch_type, layout)
if not show_names:
names = None
data = tfr.data
if mode is not None and baseline is not None:
data = rescale(data, tfr.times, baseline, mode, copy=True)
# crop time
itmin, itmax = None, None
idx = np.where(_time_mask(tfr.times, tmin, tmax))[0]
if tmin is not None:
itmin = idx[0]
if tmax is not None:
itmax = idx[-1] + 1
# crop freqs
ifmin, ifmax = None, None
idx = np.where(_time_mask(tfr.freqs, fmin, fmax))[0]
if fmin is not None:
ifmin = idx[0]
if fmax is not None:
ifmax = idx[-1] + 1
data = data[picks, ifmin:ifmax, itmin:itmax]
data = np.mean(np.mean(data, axis=2), axis=1)[:, np.newaxis]
if merge_grads:
from ..channels.layout import _merge_grad_data
data = _merge_grad_data(data)
vmin, vmax = _setup_vmin_vmax(data, vmin, vmax)
if axes is None:
fig = plt.figure()
ax = fig.gca()
else:
fig = axes.figure
ax = axes
ax.set_yticks([])
ax.set_xticks([])
ax.set_frame_on(False)
if title is not None:
ax.set_title(title)
im, _ = plot_topomap(data[:, 0],
pos,
vmin=vmin,
vmax=vmax,
axis=ax,
cmap=cmap,
image_interp='bilinear',
contours=False,
names=names)
if colorbar:
divider = make_axes_locatable(ax)
cax = divider.append_axes("right", size="5%", pad=0.05)
cbar = plt.colorbar(im, cax=cax, format='%3.2f', cmap=cmap)
cbar.set_ticks((vmin, vmax))
cbar.ax.tick_params(labelsize=12)
cbar.ax.set_title('AU')
if show:
plt.show()
return fig
for str in flist:
str=str.rstrip('\n')
if str.find('Real') != -1:
flag=1
elif flag==1 and str.find('Imag') != -1:
flag=0
elif flag==1 and isNum(str):
x.append(float(str))
elif flag==0 and isNum(str):
y.append(float(str))
fig = plt.figure(1, figsize=(10,10), dpi=50)
axScatter = plt.subplot(111)
divider = make_axes_locatable(axScatter)
axScatter.scatter(x, y)
axScatter.set_aspect(1.)
axes = fig.get_axes()[0]
axes.set_xlim((-3, 3))
axes.set_ylim((-3, 3))
axes.set_aspect('auto', adjustable='box')
plt.draw()
plt.show()
os.remove('/tmp/symbols.txt')
os.remove('/tmp/symbols_noise.txt')
def plot(self, profile, statsResults):
# *** Check if there is sufficient data to generate the plot
if len(statsResults.postHocResults.pValues) <= 0:
self.emptyAxis('No post-hoc test results')
return
if len(statsResults.postHocResults.pValues) > 200:
QtGui.QApplication.instance().setOverrideCursor(QtGui.QCursor(QtCore.Qt.ArrowCursor))
reply = QtGui.QMessageBox.question(self, 'Continue?', 'Plots contains ' + str(len(statsResults.postHocResults.pValues)) + ' rows. ' +
'It may take several seconds to generate this plot. We recommend filtering the results first.' +
'Do you wish to continue?', QtGui.QMessageBox.Yes, QtGui.QMessageBox.No)
QtGui.QApplication.instance().restoreOverrideCursor()
if reply == QtGui.QMessageBox.No:
self.emptyAxis('Too many rows.')
return
# *** Set plot properties
axesColour = str(self.preferences['Axes colour'].name())
highlightColor = (0.9, 0.9, 0.9)
# Apply p-value filter
labels = []
pValues = []
effectSizes = []
lowerCIs = []
upperCIs = []
if self.bPvalueFilter:
for i in xrange(0, len(statsResults.postHocResults.labels)):
# get numeric p-value
if isinstance(statsResults.postHocResults.pValues[i], str):
pValueSplit = statsResults.postHocResults.pValues[i].split(' ')
if pValueSplit[0][0] == '<':
pValue = float(pValueSplit[1]) - 1e-6
else:
pValue = 1.0
else:
pValue = statsResults.postHocResults.pValues[i]
# check if p-value should be filtered
if pValue <= statsResults.postHocResults.alpha:
labels.append(statsResults.postHocResults.labels[i])
pValues.append(statsResults.postHocResults.pValues[i])
effectSizes.append(statsResults.postHocResults.effectSizes[i])
lowerCIs.append(statsResults.postHocResults.lowerCIs[i])
upperCIs.append(statsResults.postHocResults.upperCIs[i])
else:
labels = list(statsResults.postHocResults.labels)
pValues = list(statsResults.postHocResults.pValues)
effectSizes = list(statsResults.postHocResults.effectSizes)
lowerCIs = list(statsResults.postHocResults.lowerCIs)
upperCIs = list(statsResults.postHocResults.upperCIs)
if len(labels) == 0:
self.emptyAxis('No rows above nominal level.')
return
# *** Determine dominant group for each contrast (i.e., row).
# Adjust labels and effect sizes to reflect the dominant group.
for i in xrange(0, len(effectSizes)):
labelSplit = labels[i].split(':')
if effectSizes[i] > 0.0:
lowerCIs[i] = effectSizes[i] - lowerCIs[i]
upperCIs[i] = upperCIs[i] - effectSizes[i]
else:
labels[i] = labelSplit[1].strip() + ' : ' + labelSplit[0].strip()
lowerCIs[i] = effectSizes[i] - lowerCIs[i]
upperCIs[i] = upperCIs[i] - effectSizes[i]
effectSizes[i] = -effectSizes[i]
# *** Sort data
data = zip(labels, pValues, effectSizes, lowerCIs, upperCIs)
if self.sortingField == 'p-values':
data = sorted(data, key = operator.itemgetter(1), reverse = True)
elif self.sortingField == 'Effect sizes':
data = sorted(data, key = operator.itemgetter(2))
elif self.sortingField == 'Group labels':
data = sorted(data, key = lambda row: row[0].lower(), reverse = True)
labels, pValues, effectSizes, lowerCIs, upperCIs = zip(*data)
labels = list(labels)
pValues = list(pValues)
effectSizes = list(effectSizes)
lowerCIs = list(lowerCIs)
upperCIs = list(upperCIs)
# *** Make list of which group is dominant in each contrast.
dominantGroup = {}
for i in xrange(0, len(effectSizes)):
labelSplit = labels[i].split(':')
groupName = labelSplit[0].strip()
if groupName in dominantGroup:
dominantGroup[groupName][0].append(effectSizes[i])
dominantGroup[groupName][1].append(i)
else:
dominantGroup[groupName] = [[effectSizes[i]],[i]]
# *** Create p-value labels
pValueTitle = 'p-value'
pValueLabels = []
for pValue in pValues:
if isinstance(pValue, str):
pValueSplit = pValue.split(' ')
if pValue[0] == '<':
pValueLabels.append(r'$<$' + pValueSplit[1])
else:
pValueLabels.append(r'$\geq$' + pValueSplit[1])
else:
pValueLabels.append(statsResults.getPValueStr(pValue))
# *** Truncate labels
adjustedLabels = list(labels)
if self.preferences['Truncate feature names']:
length = self.preferences['Length of truncated feature names']
for i in xrange(0, len(labels)):
if len(labels[i]) > length+3:
adjustedLabels[i] = labels[i][0:length] + '...'
# *** Set figure size
plotHeight = self.figHeightPerRow*len(adjustedLabels)
self.imageWidth = self.figWidth
self.imageHeight = plotHeight + 0.65 # 0.65 inches for bottom and top labels
if self.imageWidth > 256 or self.imageHeight > 256:
QtGui.QApplication.instance().setOverrideCursor(QtGui.QCursor(QtCore.Qt.ArrowCursor))
self.emptyAxis()
reply = QtGui.QMessageBox.question(self, 'Excessively large plot', 'The resulting plot is too large to display.')
QtGui.QApplication.instance().restoreOverrideCursor()
return
self.fig.set_size_inches(self.imageWidth, self.imageHeight)
# *** Determine width of y-axis labels
yLabelBounds = self.yLabelExtents(adjustedLabels, 8)
# *** Size plots which comprise the extended errorbar plot
self.fig.clear()
heightBottomLabels = 0.4 # inches
spacingBetweenPlots = 0.25 # inches
widthNumSeqPlot = 1.25 # inches
if self.bShowBarPlot == False:
widthNumSeqPlot = 0.0
spacingBetweenPlots = 0.0
widthPvalueLabels = 0.75 # inches
if self.bShowPValueLabels == False:
widthPvalueLabels = 0.1
yPlotOffsetFigSpace = heightBottomLabels / self.imageHeight
heightPlotFigSpace = plotHeight / self.imageHeight
xPlotOffsetFigSpace = yLabelBounds.width + 0.1 / self.imageWidth
pValueLabelWidthFigSpace = widthPvalueLabels / self.imageWidth
widthPlotFigSpace = 1.0 - pValueLabelWidthFigSpace - xPlotOffsetFigSpace
widthErrorBarPlot = widthPlotFigSpace*self.imageWidth - widthNumSeqPlot - spacingBetweenPlots
axInitAxis = self.fig.add_axes([xPlotOffsetFigSpace,yPlotOffsetFigSpace,widthPlotFigSpace,heightPlotFigSpace])
divider = make_axes_locatable(axInitAxis)
divider.get_vertical()[0] = Size.Fixed(len(labels)*self.figHeightPerRow)
self.fig.text(0.0,1.0,self.preferences['Selected multiple group feature'], va='top', ha='left')
if self.bShowBarPlot == True:
divider.get_horizontal()[0] = Size.Fixed(widthNumSeqPlot)
axErrorbar = divider.new_horizontal(widthErrorBarPlot, pad=spacingBetweenPlots, sharey=axInitAxis)
self.fig.add_axes(axErrorbar)
else:
divider.get_horizontal()[0] = Size.Fixed(widthErrorBarPlot)
axErrorbar = axInitAxis
# *** Plot of sequences for each subsystem
if self.bShowBarPlot == True:
axNumSeq = axInitAxis
# get relative frequency and standard deviation of each contrast
maxPercentage = 0
for i in xrange(0, len(labels)):
splitLabel = labels[i].split(':')
groupName1 = splitLabel[0].strip()
groupName2 = splitLabel[1].strip()
colour1 = str(self.preferences['Group colours'][groupName1].name())
colour2 = str(self.preferences['Group colours'][groupName2].name())
meanRelFreq1 = statsResults.getDataFromTable(statsResults.postHocResults.feature, groupName1 + ': mean rel. freq. (%)')
meanRelFreq2 = statsResults.getDataFromTable(statsResults.postHocResults.feature, groupName2 + ': mean rel. freq. (%)')
if self.bShowStdDev:
stdDev1 = statsResults.getDataFromTable(statsResults.postHocResults.feature, groupName1 + ': std. dev. (%)')
stdDev2 = statsResults.getDataFromTable(statsResults.postHocResults.feature, groupName2 + ': std. dev. (%)')
endCapSize = self.endCapSize
else:
stdDev1 = 0
stdDev2 = 0
endCapSize = 0
if meanRelFreq1 + stdDev1 > maxPercentage:
maxPercentage = meanRelFreq1 + stdDev1
if meanRelFreq2 + stdDev2 > maxPercentage:
maxPercentage = meanRelFreq2 + stdDev2
axNumSeq.barh(i+0.0, meanRelFreq1, height = 0.3, xerr=stdDev1, color=colour1, ecolor='black', capsize=endCapSize)
axNumSeq.barh(i-0.3, meanRelFreq2, height = 0.3, xerr=stdDev2, color=colour2, ecolor='black', capsize=endCapSize)
for value in np.arange(-0.5, len(labels)-1, 2):
axNumSeq.axhspan(value, value+1, facecolor=highlightColor, edgecolor='none', zorder=-1)
axNumSeq.set_xlabel('Mean proportion (%)')
axNumSeq.set_xticks([0, maxPercentage])
axNumSeq.set_xlim([0, maxPercentage*1.05])
maxPercentageStr = '%.1f' % maxPercentage
axNumSeq.set_xticklabels(['0.0', maxPercentageStr])
axNumSeq.set_yticks(np.arange(len(labels)))
axNumSeq.set_yticklabels(adjustedLabels)
axNumSeq.set_ylim([-1, len(labels)])
for a in axNumSeq.yaxis.majorTicks:
a.tick1On=False
a.tick2On=False
for a in axNumSeq.xaxis.majorTicks:
a.tick1On=True
a.tick2On=False
for line in axNumSeq.yaxis.get_ticklines():
line.set_color(axesColour)
for line in axNumSeq.xaxis.get_ticklines():
line.set_color(axesColour)
for loc, spine in axNumSeq.spines.iteritems():
if loc in ['left', 'right','top']:
spine.set_color('none')
else:
spine.set_color(axesColour)
# *** Plot confidence intervals for each subsystem
lastAxes = axErrorbar
markerSize = math.sqrt(float(self.markerSize))
axErrorbar.errorbar(effectSizes, np.arange(len(labels)), xerr=[lowerCIs,upperCIs], fmt='o', ms=markerSize, mfc='black', mec='black', ecolor='black', zorder=10)
for groupName in dominantGroup:
colour = str(self.preferences['Group colours'][groupName].name())
effectSizes = dominantGroup[groupName][0]
indices = dominantGroup[groupName][1]
axErrorbar.plot(effectSizes, indices, ls='', marker='o', ms=markerSize, mfc=colour, mec='black', zorder=100)
axErrorbar.vlines(0, -1, len(labels), linestyle='dashed', color=axesColour)
for value in np.arange(-0.5, len(labels)-1, 2):
axErrorbar.axhspan(value, value+1, facecolor=highlightColor,edgecolor='none',zorder=1)
ciTitle = ('%.3g' % ((1.0-statsResults.postHocResults.alpha)*100)) + '% confidence intervals'
axErrorbar.set_title(ciTitle)
axErrorbar.set_xlabel('Difference in mean proportions (%)')
if self.bCustomLimits:
axErrorbar.set_xlim([self.minX, self.maxX])
else:
self.minX, self.maxX = axErrorbar.get_xlim()
if self.bShowBarPlot == False:
axErrorbar.set_yticks(np.arange(len(labels)))
axErrorbar.set_yticklabels(labels)
axErrorbar.set_ylim([-1, len(labels)])
else:
for label in axErrorbar.get_yticklabels():
label.set_visible(False)
for a in axErrorbar.yaxis.majorTicks:
a.set_visible(False)
for a in axErrorbar.xaxis.majorTicks:
a.tick1On=True
a.tick2On=False
for a in axErrorbar.yaxis.majorTicks:
a.tick1On=False
a.tick2On=False
for line in axErrorbar.yaxis.get_ticklines():
line.set_visible(False)
for line in axErrorbar.xaxis.get_ticklines():
line.set_color(axesColour)
for loc, spine in axErrorbar.spines.iteritems():
if loc in ['left','right','top']:
spine.set_color('none')
else:
spine.set_color(axesColour)
# *** Show p-values on right of last plot
if self.bShowPValueLabels == True:
axRight = lastAxes.twinx()
axRight.set_yticks(np.arange(len(pValueLabels)))
axRight.set_yticklabels(pValueLabels)
axRight.set_ylim([-1, len(pValueLabels)])
axRight.set_ylabel(pValueTitle)
for a in axRight.yaxis.majorTicks:
a.tick1On=False
a.tick2On=False
for loc, spine in axRight.spines.iteritems():
spine.set_color('none')
self.updateGeometry()
self.draw()
if (num == 0):
Vortensity0 = Vortensity
print("Doing geometric transformation to R-theta plane...")
# create polar-coordinate array
Vortensity_polar = geometric_transform(Vortensity.T, cartesian2polar, output_shape=(Vortensity.T.shape[0], Vortensity.T.shape[0]),
extra_keywords={'inputshape': Vortensity.T.shape, 'origin': (Vortensity.T.shape[0]/2, Vortensity.T.shape[1]/2)})
if (num == 0):
Vortensity0_polar = Vortensity_polar
####################################################################
# plot
if not (skip_cartesian):
ax = fig.add_subplot(1, len(dir_array), count_dir)
divider = make_axes_locatable(ax)
cmap = cm.get_cmap('jet')
im = ax.imshow(np.log10(Vortensity/Vortensity0), origin='lower',
vmin=min_scale, vmax=max_scale,
extent=[rangeX[0], rangeX[1], rangeY[0], rangeY[1]], cmap=cmap)
xlabel("$x$", fontsize=16)
ylabel("$y$", fontsize=16)
ax.set_xlim(rangeX[0], rangeX[1])
ax.set_ylim(rangeY[0], rangeY[1])
xticks, yticks = ax.xaxis.get_majorticklocs(), ax.yaxis.get_majorticklocs()
ax.xaxis.set_ticklabels(['%d' % (xticks[n] - 0.5*BoxX)
for n in range(len(xticks))])
ax.yaxis.set_ticklabels(['%d' % (yticks[n] - 0.5 * BoxY)
if run_data_for_tutorials:
# Make moment 0 maps for the tutorial data
fid_mom0 = fits.open(
"../../testingdata/Fiducial0_flatrho_0021_00_radmc_moment0.fits")[0]
des_mom0 = fits.open(
"../../testingdata/Design4_flatrho_0021_00_radmc_moment0.fits")[0]
from mpl_toolkits.axes_grid import make_axes_locatable
ax = plt.subplot(121)
im1 = plt.imshow(fid_mom0.data / 1000., origin='lower')
ax.set_title("Fiducial")
divider = make_axes_locatable(ax)
cax = divider.append_axes("right", "5%", pad="3%")
cb = plt.colorbar(im1, cax=cax)
ax2 = plt.subplot(122)
im2 = plt.imshow(des_mom0.data / 1000., origin='lower')
ax2.set_title("Design")
divider = make_axes_locatable(ax2)
cax2 = divider.append_axes("right", "5%", pad="3%")
cb2 = plt.colorbar(im2, cax=cax2)
cb2.set_label(r"K km s$^{-1}$")
plt.tight_layout()
plt.savefig(osjoin(fig_path, "design_fiducial_moment0.png"))
plt.close()
def draw_graph(G,
labels=None,
node_colors=None,
node_shapes=None,
node_scale=1.0,
edge_style='solid',
edge_cmap=None,
colorbar=False,
vrange=None,
layout=None,
title=None,
font_family='sans-serif',
font_size=9,
stretch_factor=1.0,
edge_alpha=True,
fig_size=None):
"""Draw a weighted graph with options to visualize link weights.
The resulting diagram uses the rank of each node as its size, and the
weight of each link (after discarding thresholded values, see below) as the
link opacity.
It maps edge weight to color as well as line opacity and thickness,
allowing the color part to be hardcoded over a value range (to permit valid
cross-figure comparisons for different graphs, so the same color
corresponds to the same link weight even if each graph has a different
range of weights). The nodes sizes are proportional to their degree,
computed as the sum of the weights of all their links. The layout defaults
to circular, but any nx layout function can be passed in, as well as a
statically precomputed layout.
Parameters
----------
G : weighted graph
The values must be of the form (v1,v2), with all v2 in [0,1]. v1 are
used for colors, v2 for thickness/opacity.
labels : list or dict, optional.
An indexable object that maps nodes to strings. If not given, the
string form of each node is used as a label. If False, no labels are
drawn.
node_colors : list or dict, optional.
An indexable object that maps nodes to valid matplotlib color specs. See
matplotlib's plot() function for details.
node_shapes : list or dict, optional.
An indexable object that maps nodes to valid matplotlib shape specs. See
matplotlib's scatter() function for details. If not given, circles are
used.
node_scale : float, optional
A scale factor to globally stretch or shrink all nodes symbols by.
edge_style : string, optional
Line style for the edges, defaults to 'solid'.
edge_cmap : matplotlib colormap, optional.
A callable that returns valid color specs, like matplotlib colormaps.
If not given, edges are colored black.
colorbar : bool
If true, automatically add a colorbar showing the mapping of graph weight
values to colors.
vrange : pair of floats
If given, this indicates the total range of values that the weights can
in principle occupy, and is used to set the lower/upper range of the
colormap. This allows you to set the range of multiple different figures
to the same values, even if each individual graph has range variations,
so that visual color comparisons across figures are valid.
layout : function or layout dict, optional
A NetworkX-like layout function or the result of a precomputed layout for
the given graph. NetworkX produces layouts as dicts keyed by nodes and
with (x,y) pairs of coordinates as values, any function that produces
this kind of output is acceptable. Defaults to nx.circular_layout.
title : string, optional.
If given, title to put on the main plot.
font_family : string, optional.
Font family used for the node labels and title.
font_size : int, optional.
Font size used for the node labels and title.
stretch_factor : float, optional
A global scaling factor to make the graph larger (or smaller if <1).
This can be used to separate the nodes if they start overlapping.
edge_alpha: bool, optional
Whether to weight the transparency of each edge by a factor equivalent to
its relative weight
fig_size: list of height by width, the size of the figure (in
inches). Defaults to [6,6]
Returns
-------
fig
The matplotlib figure object with the plot.
"""
if fig_size is None:
figsize = [6, 6]
scaler = figsize[0] / 6.
# For the size of the node symbols
node_size_base = 1000 * scaler
node_min_size = 200 * scaler
default_node_shape = 'o'
# Default colors if none given
default_node_color = 'r'
default_edge_color = 'k'
# Max edge width
max_width = 13 * scaler
min_width = 2 * scaler
font_family = 'sans-serif'
# We'll use the nodes a lot, let's make a numpy array of them
nodes = np.array(sorted(G.nodes()))
nnod = len(nodes)
# Build a 'weighted degree' array obtained by adding the (absolute value)
# of the weights for all edges pointing to each node:
amat = nx.adj_matrix(G).A # get a normal array out of it
degarr = abs(amat).sum(0) # weights are sums across rows
# Map the degree to the 0-1 range so we can use it for sizing the nodes.
try:
odegree = rescale_arr(degarr, 0, 1)
# Make an array of node sizes based on node degree
node_sizes = odegree * node_size_base + node_min_size
except ZeroDivisionError:
# All nodes same size
node_sizes = np.empty(nnod, float)
node_sizes.fill(0.5 * node_size_base + node_min_size)
# Adjust node size list. We square the scale factor because in mpl, node
# sizes represent area, not linear size, but it's more intuitive for the
# user to think of linear factors (the overall figure scale factor is also
# linear).
node_sizes *= node_scale ** 2
# Set default node properties
if node_colors is None:
node_colors = [default_node_color] * nnod
if node_shapes is None:
node_shapes = [default_node_shape] * nnod
# Set default edge colormap
if edge_cmap is None:
# Make an object with the colormap API, that maps all input values to
# the default color (with proper alhpa)
edge_cmap = (lambda val, alpha:
colors.colorConverter.to_rgba(default_edge_color, alpha))
# if vrange is None, we set the color range from the values, else the user
# can specify it
# e[2] is edge value: edges_iter returns (i,j,data)
gvals = np.array([e[2]['weight'] for e in G.edges(data=True)])
gvmin, gvmax = gvals.min(), gvals.max()
gvrange = gvmax - gvmin
if vrange is None:
vrange = gvmin, gvmax
# Now, construct the normalization for the colormap
cnorm = mpl.colors.Normalize(vmin=vrange[0], vmax=vrange[1])
# Create the actual plot where the graph will be displayed
figsize = np.array(figsize, float)
figsize *= stretch_factor
fig = plt.figure(figsize=figsize)
ax_graph = fig.add_subplot(1, 1, 1)
fig.sca(ax_graph)
if layout is None:
layout = nx.circular_layout
# Compute positions for all nodes - nx has several algorithms
if callable(layout):
pos = layout(G)
else:
# The user can also provide a precomputed layout
pos = layout
# Draw nodes
for nod in nodes:
nx.draw_networkx_nodes(G,
pos,
nodelist=[nod],
node_color=node_colors[nod],
node_shape=node_shapes[nod],
node_size=node_sizes[nod],
edgecolors='k')
# Draw edges
if not isinstance(G, nx.DiGraph):
# Undirected graph, simple lines for edges
# We need the size of the value range to properly scale colors
vsize = vrange[1] - vrange[0]
gvals_normalized = G.metadata['vals_norm']
for (u, v, y) in G.edges(data=True):
# The graph value is the weight, and the normalized values are in
# [0,1], used for thickness/transparency
alpha = gvals_normalized[u, v]
# Scale the color choice to the specified vrange, so that
ecol = (y['weight'] - vrange[0]) / vsize
#print 'u,v:',u,v,'y:',y,'ecol:',ecol # dbg
if edge_alpha:
fade = alpha
else:
fade = 1.0
edge_color = [tuple(edge_cmap(ecol, fade))]
#dbg:
#print u,v,y
nx.draw_networkx_edges(G,
pos,
edgelist=[(u, v)],
width=min_width + alpha * max_width,
edge_color=edge_color,
style=edge_style)
else:
# Directed graph, use arrows.
# XXX - this is currently broken.
raise NotImplementedError("arrow drawing currently broken")
## for (u,v,x) in G.edges(data=True):
## y,w = x
## draw_arrows(G,pos,edgelist=[(u,v)],
## edge_color=[w],
## alpha=w,
## edge_cmap=edge_cmap,
## width=w*max_width)
# Draw labels. If not given, we use the string form of the nodes. If
# labels is False, no labels are drawn.
if labels is None:
labels = map(str, nodes)
if labels:
lab_idx = list(range(len(labels)))
labels_dict = dict(zip(lab_idx, labels))
nx.draw_networkx_labels(G,
pos,
labels_dict,
font_size=font_size,
font_family=font_family)
if title:
plt.title(title, fontsize=font_size)
# Turn off x and y axes labels in pylab
plt.xticks([])
plt.yticks([])
# Add a colorbar if requested
if colorbar:
divider = make_axes_locatable(ax_graph)
ax_cb = divider.new_vertical(size="20%", pad=0.2, pack_start=True)
fig.add_axes(ax_cb)
cb = mpl.colorbar.ColorbarBase(ax_cb,
cmap=edge_cmap,
norm=cnorm,
#boundaries = np.linspace(min((gvmin,0)),
# max((gvmax,0)),
# 256),
orientation='horizontal',
format='%.2f')
# Always return the MPL figure object so the user can further manipulate it
return fig
def plot(self, profile, statsResults):
# *** Check if there is sufficient data to generate the plot
if len(statsResults.postHocResults.pValues) <= 0:
self.emptyAxis('No post-hoc test results')
return
if len(statsResults.postHocResults.pValues) > 200:
QtGui.QApplication.instance().setOverrideCursor(QtGui.QCursor(QtCore.Qt.ArrowCursor))
reply = QtGui.QMessageBox.question(self, 'Continue?', 'Plots contains ' + str(len(statsResults.postHocResults.pValues)) + ' rows. ' +
'It may take several seconds to generate this plot. We recommend filtering the results first.' +
'Do you wish to continue?', QtGui.QMessageBox.Yes, QtGui.QMessageBox.No)
QtGui.QApplication.instance().restoreOverrideCursor()
if reply == QtGui.QMessageBox.No:
self.emptyAxis('Too many rows.')
return
# *** Set plot properties
axesColour = str(self.preferences['Axes colour'].name())
highlightColor = (0.9, 0.9, 0.9)
# Apply p-value filter
labels = []
pValues = []
effectSizes = []
lowerCIs = []
upperCIs = []
if self.bPvalueFilter:
for i in xrange(0, len(statsResults.postHocResults.labels)):
# get numeric p-value
if isinstance(statsResults.postHocResults.pValues[i], str):
pValueSplit = statsResults.postHocResults.pValues[i].split(' ')
if pValueSplit[0][0] == '<':
pValue = float(pValueSplit[1]) - 1e-6
else:
pValue = 1.0
else:
pValue = statsResults.postHocResults.pValues[i]
# check if p-value should be filtered
if pValue <= statsResults.postHocResults.alpha:
labels.append(statsResults.postHocResults.labels[i])
pValues.append(statsResults.postHocResults.pValues[i])
effectSizes.append(statsResults.postHocResults.effectSizes[i])
lowerCIs.append(statsResults.postHocResults.lowerCIs[i])
upperCIs.append(statsResults.postHocResults.upperCIs[i])
else:
labels = list(statsResults.postHocResults.labels)
pValues = list(statsResults.postHocResults.pValues)
effectSizes = list(statsResults.postHocResults.effectSizes)
lowerCIs = list(statsResults.postHocResults.lowerCIs)
upperCIs = list(statsResults.postHocResults.upperCIs)
if len(labels) == 0:
self.emptyAxis('No rows above nominal level.')
return
# *** Determine dominant group for each contrast (i.e., row).
# Adjust labels and effect sizes to reflect the dominant group.
for i in xrange(0, len(effectSizes)):
labelSplit = labels[i].split(':')
if effectSizes[i] > 0.0:
lowerCIs[i] = effectSizes[i] - lowerCIs[i]
upperCIs[i] = upperCIs[i] - effectSizes[i]
else:
labels[i] = labelSplit[1].strip() + ' : ' + labelSplit[0].strip()
lowerCIs[i] = effectSizes[i] - lowerCIs[i]
upperCIs[i] = upperCIs[i] - effectSizes[i]
effectSizes[i] = -effectSizes[i]
# *** Sort data
data = zip(labels, pValues, effectSizes, lowerCIs, upperCIs)
if self.sortingField == 'p-values':
data = sorted(data, key = operator.itemgetter(1), reverse = True)
elif self.sortingField == 'Effect sizes':
data = sorted(data, key = operator.itemgetter(2))
elif self.sortingField == 'Group labels':
data = sorted(data, key = lambda row: row[0].lower(), reverse = True)
labels, pValues, effectSizes, lowerCIs, upperCIs = zip(*data)
labels = list(labels)
pValues = list(pValues)
effectSizes = list(effectSizes)
lowerCIs = list(lowerCIs)
upperCIs = list(upperCIs)
# *** Make list of which group is dominant in each contrast.
dominantGroup = {}
for i in xrange(0, len(effectSizes)):
labelSplit = labels[i].split(':')
groupName = labelSplit[0].strip()
if groupName in dominantGroup:
dominantGroup[groupName][0].append(effectSizes[i])
dominantGroup[groupName][1].append(i)
else:
dominantGroup[groupName] = [[effectSizes[i]],[i]]
# *** Create p-value labels
pValueTitle = 'p-value'
pValueLabels = []
for pValue in pValues:
if isinstance(pValue, str):
pValueSplit = pValue.split(' ')
if pValue[0] == '<':
pValueLabels.append(r'$<$' + pValueSplit[1])
else:
pValueLabels.append(r'$\geq$' + pValueSplit[1])
else:
pValueLabels.append(statsResults.getPValueStr(pValue))
# *** Truncate labels
adjustedLabels = list(labels)
if self.preferences['Truncate feature names']:
length = self.preferences['Length of truncated feature names']
for i in xrange(0, len(labels)):
if len(labels[i]) > length+3:
adjustedLabels[i] = labels[i][0:length] + '...'
# *** Set figure size
plotHeight = self.figHeightPerRow*len(adjustedLabels)
self.imageWidth = self.figWidth
self.imageHeight = plotHeight + 0.65 # 0.65 inches for bottom and top labels
if self.imageWidth > 256 or self.imageHeight > 256:
QtGui.QApplication.instance().setOverrideCursor(QtGui.QCursor(QtCore.Qt.ArrowCursor))
self.emptyAxis()
reply = QtGui.QMessageBox.question(self, 'Excessively large plot', 'The resulting plot is too large to display.')
QtGui.QApplication.instance().restoreOverrideCursor()
return
self.fig.set_size_inches(self.imageWidth, self.imageHeight)
# *** Determine width of y-axis labels
yLabelBounds = self.yLabelExtents(adjustedLabels, 8)
# *** Size plots which comprise the extended errorbar plot
self.fig.clear()
heightBottomLabels = 0.4 # inches
spacingBetweenPlots = 0.25 # inches
widthNumSeqPlot = 1.25 # inches
if self.bShowBarPlot == False:
widthNumSeqPlot = 0.0
spacingBetweenPlots = 0.0
widthPvalueLabels = 0.75 # inches
if self.bShowPValueLabels == False:
widthPvalueLabels = 0.1
yPlotOffsetFigSpace = heightBottomLabels / self.imageHeight
heightPlotFigSpace = plotHeight / self.imageHeight
xPlotOffsetFigSpace = yLabelBounds.width + 0.1 / self.imageWidth
pValueLabelWidthFigSpace = widthPvalueLabels / self.imageWidth
widthPlotFigSpace = 1.0 - pValueLabelWidthFigSpace - xPlotOffsetFigSpace
widthErrorBarPlot = widthPlotFigSpace*self.imageWidth - widthNumSeqPlot - spacingBetweenPlots
axInitAxis = self.fig.add_axes([xPlotOffsetFigSpace,yPlotOffsetFigSpace,widthPlotFigSpace,heightPlotFigSpace])
divider = make_axes_locatable(axInitAxis)
divider.get_vertical()[0] = Size.Fixed(len(labels)*self.figHeightPerRow)
self.fig.text(0.0,1.0,self.preferences['Selected multiple group feature'], va='top', ha='left')
if self.bShowBarPlot == True:
divider.get_horizontal()[0] = Size.Fixed(widthNumSeqPlot)
axErrorbar = divider.new_horizontal(widthErrorBarPlot, pad=spacingBetweenPlots, sharey=axInitAxis)
self.fig.add_axes(axErrorbar)
else:
divider.get_horizontal()[0] = Size.Fixed(widthErrorBarPlot)
axErrorbar = axInitAxis
# *** Plot of sequences for each subsystem
if self.bShowBarPlot == True:
axNumSeq = axInitAxis
# get relative frequency and standard deviation of each contrast
maxPercentage = 0
for i in xrange(0, len(labels)):
splitLabel = labels[i].split(':')
groupName1 = splitLabel[0].strip()
groupName2 = splitLabel[1].strip()
colour1 = str(self.preferences['Group colours'][groupName1].name())
colour2 = str(self.preferences['Group colours'][groupName2].name())
meanRelFreq1 = statsResults.getDataFromTable(statsResults.postHocResults.feature, groupName1 + ': mean rel. freq. (%)')
meanRelFreq2 = statsResults.getDataFromTable(statsResults.postHocResults.feature, groupName2 + ': mean rel. freq. (%)')
if self.bShowStdDev:
stdDev1 = statsResults.getDataFromTable(statsResults.postHocResults.feature, groupName1 + ': std. dev. (%)')
stdDev2 = statsResults.getDataFromTable(statsResults.postHocResults.feature, groupName2 + ': std. dev. (%)')
endCapSize = self.endCapSize
else:
stdDev1 = 0
stdDev2 = 0
endCapSize = 0
if meanRelFreq1 + stdDev1 > maxPercentage:
maxPercentage = meanRelFreq1 + stdDev1
if meanRelFreq2 + stdDev2 > maxPercentage:
maxPercentage = meanRelFreq2 + stdDev2
axNumSeq.barh(i+0.0, meanRelFreq1, height = 0.3, xerr=stdDev1, color=colour1, ecolor='black', capsize=endCapSize)
axNumSeq.barh(i-0.3, meanRelFreq2, height = 0.3, xerr=stdDev2, color=colour2, ecolor='black', capsize=endCapSize)
for value in np.arange(-0.5, len(labels)-1, 2):
axNumSeq.axhspan(value, value+1, facecolor=highlightColor, edgecolor='none', zorder=-1)
axNumSeq.set_xlabel('Mean proportion (%)')
axNumSeq.set_xticks([0, maxPercentage])
axNumSeq.set_xlim([0, maxPercentage*1.05])
maxPercentageStr = '%.1f' % maxPercentage
axNumSeq.set_xticklabels(['0.0', maxPercentageStr])
axNumSeq.set_yticks(np.arange(len(labels)))
axNumSeq.set_yticklabels(adjustedLabels)
axNumSeq.set_ylim([-1, len(labels)])
for a in axNumSeq.yaxis.majorTicks:
a.tick1On=False
a.tick2On=False
for a in axNumSeq.xaxis.majorTicks:
a.tick1On=True
a.tick2On=False
for line in axNumSeq.yaxis.get_ticklines():
line.set_color(axesColour)
for line in axNumSeq.xaxis.get_ticklines():
line.set_color(axesColour)
for loc, spine in axNumSeq.spines.iteritems():
if loc in ['left', 'right','top']:
spine.set_color('none')
else:
spine.set_color(axesColour)
# *** Plot confidence intervals for each subsystem
lastAxes = axErrorbar
markerSize = math.sqrt(float(self.markerSize))
axErrorbar.errorbar(effectSizes, np.arange(len(labels)), xerr=[lowerCIs,upperCIs], fmt='o', ms=markerSize, mfc='black', mec='black', ecolor='black', zorder=10)
for groupName in dominantGroup:
colour = str(self.preferences['Group colours'][groupName].name())
effectSizes = dominantGroup[groupName][0]
indices = dominantGroup[groupName][1]
axErrorbar.plot(effectSizes, indices, ls='', marker='o', ms=markerSize, mfc=colour, mec='black', zorder=100)
axErrorbar.vlines(0, -1, len(labels), linestyle='dashed', color=axesColour)
for value in np.arange(-0.5, len(labels)-1, 2):
axErrorbar.axhspan(value, value+1, facecolor=highlightColor,edgecolor='none',zorder=1)
ciTitle = ('%.3g' % ((1.0-statsResults.postHocResults.alpha)*100)) + '% confidence intervals'
axErrorbar.set_title(ciTitle)
axErrorbar.set_xlabel('Difference in mean proportions (%)')
if self.bCustomLimits:
axErrorbar.set_xlim([self.minX, self.maxX])
else:
self.minX, self.maxX = axErrorbar.get_xlim()
if self.bShowBarPlot == False:
axErrorbar.set_yticks(np.arange(len(labels)))
axErrorbar.set_yticklabels(labels)
axErrorbar.set_ylim([-1, len(labels)])
else:
for label in axErrorbar.get_yticklabels():
label.set_visible(False)
for a in axErrorbar.yaxis.majorTicks:
a.set_visible(False)
for a in axErrorbar.xaxis.majorTicks:
a.tick1On=True
a.tick2On=False
for a in axErrorbar.yaxis.majorTicks:
a.tick1On=False
a.tick2On=False
for line in axErrorbar.yaxis.get_ticklines():
line.set_visible(False)
for line in axErrorbar.xaxis.get_ticklines():
line.set_color(axesColour)
for loc, spine in axErrorbar.spines.iteritems():
if loc in ['left','right','top']:
spine.set_color('none')
else:
spine.set_color(axesColour)
# *** Show p-values on right of last plot
if self.bShowPValueLabels == True:
axRight = lastAxes.twinx()
axRight.set_yticks(np.arange(len(pValueLabels)))
axRight.set_yticklabels(pValueLabels)
axRight.set_ylim([-1, len(pValueLabels)])
axRight.set_ylabel(pValueTitle)
for a in axRight.yaxis.majorTicks:
a.tick1On=False
a.tick2On=False
self.updateGeometry()
self.draw()
def plot_tfr_topomap(tfr, tmin=None, tmax=None, fmin=None, fmax=None,
ch_type='mag', baseline=None, mode='mean', layout=None,
vmax=None, vmin=None, cmap='RdBu_r', sensors=True,
colorbar=True, unit=None, res=64, size=2, format='%1.1e',
show_names=False, title=None, axes=None, show=True):
"""Plot topographic maps of specific time-frequency intervals of TFR data
Parameters
----------
tfr : AvereageTFR
The AvereageTFR object.
tmin : None | float
The first time instant to display. If None the first time point
available is used.
tmax : None | float
The last time instant to display. If None the last time point
available is used.
fmin : None | float
The first frequency to display. If None the first frequency
available is used.
fmax : None | float
The last frequency to display. If None the last frequency
available is used.
ch_type : 'mag' | 'grad' | 'planar1' | 'planar2' | 'eeg'
The channel type to plot. For 'grad', the gradiometers are
collected in pairs and the RMS for each pair is plotted.
baseline : tuple or list of length 2
The time interval to apply rescaling / baseline correction.
If None do not apply it. If baseline is (a, b)
the interval is between "a (s)" and "b (s)".
If a is None the beginning of the data is used
and if b is None then b is set to the end of the interval.
If baseline is equal to (None, None) all the time
interval is used.
mode : 'logratio' | 'ratio' | 'zscore' | 'mean' | 'percent'
Do baseline correction with ratio (power is divided by mean
power during baseline) or z-score (power is divided by standard
deviation of power during baseline after subtracting the mean,
power = [power - mean(power_baseline)] / std(power_baseline))
If None, baseline no correction will be performed.
layout : None | Layout
Layout instance specifying sensor positions (does not need to
be specified for Neuromag data). If possible, the correct layout
file is inferred from the data; if no appropriate layout file
was found, the layout is automatically generated from the sensor
locations.
vmin : float | callable
The value specfying the lower bound of the color range.
If None, and vmax is None, -vmax is used. Else np.min(data).
If callable, the output equals vmin(data).
vmax : float | callable
The value specfying the upper bound of the color range.
If None, the maximum absolute value is used. If vmin is None,
but vmax is not, defaults to np.min(data).
If callable, the output equals vmax(data).
cmap : matplotlib colormap
Colormap. For magnetometers and eeg defaults to 'RdBu_r', else
'Reds'.
sensors : bool | str
Add markers for sensor locations to the plot. Accepts matplotlib
plot format string (e.g., 'r+' for red plusses). If True, a circle will
be used (via .add_artist). Defaults to True.
colorbar : bool
Plot a colorbar.
unit : str | None
The unit of the channel type used for colorbar labels.
res : int
The resolution of the topomap image (n pixels along each side).
size : float
Side length per topomap in inches.
format : str
String format for colorbar values.
show_names : bool | callable
If True, show channel names on top of the map. If a callable is
passed, channel names will be formatted using the callable; e.g., to
delete the prefix 'MEG ' from all channel names, pass the function
lambda x: x.replace('MEG ', ''). If `mask` is not None, only
significant sensors will be shown.
title : str | None
Title. If None (default), no title is displayed.
axes : instance of Axis | None
The axes to plot to. If None the axes is defined automatically.
show : bool
Call pyplot.show() at the end.
Returns
-------
fig : matplotlib.figure.Figure
The figure containing the topography.
"""
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import make_axes_locatable
picks, pos, merge_grads, names, _ = _prepare_topo_plot(tfr, ch_type,
layout)
if not show_names:
names = None
data = tfr.data
if mode is not None and baseline is not None:
data = rescale(data, tfr.times, baseline, mode, copy=True)
# crop time
itmin, itmax = None, None
if tmin is not None:
itmin = np.where(tfr.times >= tmin)[0][0]
if tmax is not None:
itmax = np.where(tfr.times <= tmax)[0][-1]
# crop freqs
ifmin, ifmax = None, None
if fmin is not None:
ifmin = np.where(tfr.freqs >= fmin)[0][0]
if fmax is not None:
ifmax = np.where(tfr.freqs <= fmax)[0][-1]
data = data[picks, ifmin:ifmax, itmin:itmax]
data = np.mean(np.mean(data, axis=2), axis=1)[:, np.newaxis]
if merge_grads:
from ..channels.layout import _merge_grad_data
data = _merge_grad_data(data)
vmin, vmax = _setup_vmin_vmax(data, vmin, vmax)
if axes is None:
fig = plt.figure()
ax = fig.gca()
else:
fig = axes.figure
ax = axes
ax.set_yticks([])
ax.set_xticks([])
ax.set_frame_on(False)
if title is not None:
ax.set_title(title)
im, _ = plot_topomap(data[:, 0], pos, vmin=vmin, vmax=vmax,
axis=ax, cmap=cmap, image_interp='bilinear',
contours=False, names=names)
if colorbar:
divider = make_axes_locatable(ax)
cax = divider.append_axes("right", size="5%", pad=0.05)
cbar = plt.colorbar(im, cax=cax, format='%3.2f', cmap=cmap)
cbar.set_ticks((vmin, vmax))
cbar.ax.tick_params(labelsize=12)
cbar.ax.set_title('AU')
if show:
plt.show()
return fig
def coloured_spy(A, cmap_name="coolwarm", log=False, symmetric_colorbar=False, **kwargs):
"""
Convenience function for using matplotlib to
generate a spy plot for inspecting e.g. a jacobian
matrix or its LU decomposition.
Parameters
----------
A: 2D array
Array to inspect, populated e.g. by jacobian callback.
cmap_name: string (default: coolwarm)
name of matplotlib colormap to use, kwargs["cmap"] overrides this.
log: bool or int
when True: LogNorm/SymLogNorm is used,
when integer: SymlogNorm(10**log). Note that "norm" in kwargs
override this.
symmetric_colorbar: bool or float
to make divergent colormaps pass through zero as intended.
if float: max abolute value of colormap (linear)
Returns
-------
Pair (tuple) of axes plotted to (spy, colorbar)
.. note:: colorbar does not play nicely with
SymLogNorm why a custom colorbar axes is drawn.
"""
from matplotlib.ticker import MaxNLocator
from matplotlib.cm import get_cmap
from matplotlib.colors import LogNorm, SymLogNorm
from mpl_toolkits.axes_grid import make_axes_locatable
A = np.asarray(A)
if "cmap" not in kwargs:
kwargs["cmap"] = get_cmap(cmap_name)
plt.figure()
ax_imshow = plt.subplot(111)
if log is not False and "norm" not in kwargs:
if isinstance(symmetric_colorbar, (float, int)) and symmetric_colorbar is not False:
Amin = -symmetric_colorbar
Amax = symmetric_colorbar
else:
Amin = np.min(A[np.where(A != 0)])
Amax = np.max(A[np.where(A != 0)])
if symmetric_colorbar is True:
Amin = -max(-Amin, Amax)
Amax = -Amin
if log is True:
if np.any(A < 0):
log = int(np.round(np.log10(Amax) - 13))
else:
minlog = int(floor(np.log10(Amin)))
maxlog = int(ceil(np.log10(Amax)))
tick_locations = [10 ** x for x in range(minlog, maxlog + 1)]
kwargs["norm"] = LogNorm()
elif isinstance(log, int):
tick_locations = []
if Amin < 0:
minlog = int(ceil(np.log10(-Amin)))
tick_locations.extend([-(10 ** x) for x in range(minlog, log - 1, -1)])
tick_locations.extend([0])
else:
tick_locations.extend([0])
minlog = int(floor(np.log10(Amin)))
tick_locations.extend([10 ** x for x in range(minlog, log + 1)])
if Amax < 0:
pass # Ticks already reach 0
else:
maxlog = int(ceil(np.log10(Amax)))
tick_locations.extend([10 ** x for x in range(log, maxlog + 1)])
kwargs["norm"] = SymLogNorm(10 ** log)
else:
raise TypeError("log kwarg not understood: {}".format(log))
else:
tick_locations = np.linspace(np.min(A), np.max(A), 10)
ax_imshow.imshow(A, interpolation="none", **kwargs)
ya = ax_imshow.get_yaxis()
ya.set_major_locator(MaxNLocator(integer=True))
xa = ax_imshow.get_xaxis()
xa.set_major_locator(MaxNLocator(integer=True))
divider = make_axes_locatable(ax_imshow)
ax_colorbar = divider.append_axes("right", size="5%", pad=0.05)
def colorbar(ticks=None, norm=None, log10threshy=None):
if isinstance(norm, (LogNorm, SymLogNorm)):
levels = np.concatenate(
[np.linspace(ticks[i], ticks[i + 1], 9, endpoint=False) for i in range(len(ticks) - 1)]
+ [np.array([ticks[-1]])]
).flatten()
else:
levels = ticks
xc = [np.zeros_like(levels), np.ones_like(levels)]
yc = [levels, levels]
ax_colorbar.contourf(xc, yc, yc, levels=levels, norm=norm, cmap=kwargs["cmap"])
if isinstance(norm, LogNorm):
ax_colorbar.set_yscale("log")
elif isinstance(norm, SymLogNorm):
ax_colorbar.set_yscale("symlog", linthreshy=10 ** log)
ax_colorbar.yaxis.tick_right()
ax_colorbar.set_xticks([])
ax_colorbar.set_yticks(ticks)
colorbar(ticks=tick_locations, norm=kwargs.get("norm", None))
plt.tight_layout()
return ax_imshow, ax_colorbar
