def plot_convolution(plotname, tsframe, tsframe_max, paperwidth):
colormap=plt.get_cmap('YlOrRd_r')
colormap.set_under([0.2, 0.2, 0.2])
figure_width = paperwidth
figure_height = paperwidth
figure_size = [figure_width, figure_height]
config.load_config_small()
width = tsframe.shape[1]; height = tsframe.shape[0]
xs,ys = np.meshgrid(np.arange(0, width+1), np.arange(0, height+1))
fig, axes = plt.subplots(nrows=1, ncols=1, sharex=False, sharey=False, squeeze=True, figsize=figure_size, dpi=100)
ax = axes
# Workaround inverted y-axis
ax.invert_yaxis()
# pcolormesh aligns cells on their edges, while imshow aligns them on their centers.
ax.pcolormesh(xs-0.5, ys-0.5, tsframe, cmap=colormap, vmin=0.1, vmax=tsframe_max,
shading="faceted", linestyle="dashed", linewidth=0.5, edgecolor=[0.0, 0.0, 0.0])
# Absolute number
#for i,j in ((x,y) for x in np.arange(0, len(tsframe))
# for y in np.arange(0, len(tsframe[0]))):
# if tsframe[i][j] >= 1:
# ax.annotate(str(int(tsframe[i][j])), xy=(j,i), fontsize=3.5, ha='center', va='center')
ax.set_aspect('equal')
ax.set_xlim([-0.5, width-0.5])
ax.set_ylim([height-0.5, -0.5])
ax.xaxis.tick_top()
ax.tick_params(axis='both', which='both', left='off', right='off', bottom='off', top='off', labeltop='on',
pad=1, labelsize=4)
#ax.xaxis.set_major_locator(plt.NullLocator())
#ax.yaxis.set_major_locator(plt.NullLocator())
fig.tight_layout()
fig.savefig(plotname+".pdf", pad_inches=0, bbox_inches='tight', dpi=fig.dpi) # pdf
fig.savefig(plotname+".pgf", pad_inches=0, bbox_inches='tight', dpi=fig.dpi) # pgf
############
# Plotting
###########
text_width = 6.30045 # LaTeX text width in inches
golden_ratio = (1 + np.sqrt(5)) / 2.0
size_factor = 0.75
figure_width = size_factor * text_width
figure_height = (figure_width / golden_ratio)
#figure_height = (text_width / golden_ratio) # height is golden ratio to page width
#figure_height = 1.3 * figure_width
#figure_size = [figure_width, figure_height]
figure_size = [figure_width, figure_width]
config.load_config_small()
# Arbitrary colormap
import matplotlib.colors as colors
import matplotlib.cm as cmx
cdict = {
'red': ((0.0, 0.9, 0.9), (1.0, 0.9, 0.9)),
'green': ((0.0, 0.9, 0.9), (1.0, 0.0, 0.0)),
'blue': ((0.0, 0.0, 0.0), (1.0, 0.0, 0.0))
}
#plt.register_cmap(name='YELLOW_RED', data=cdict)
#colorMap = plt.get_cmap("YELLOW_RED")
colorMap = plt.get_cmap("YlOrRd_r")
fontsize='small', labelspacing=0.5, handletextpad=0.05, columnspacing=0.5, borderpad=0.5,
fancybox=True, shadow=False, framealpha=1.0)
plotname = "scatterplotmatrix_legend"
fig.savefig(plotname+".pdf", pad_inches=0, dpi=fig.dpi) # pdf
fig.savefig(plotname+".pgf", pad_inches=0, dpi=fig.dpi) # pgf
plt.close()
'''
'''
#######################################################
# Bean plot (standardized features)
#######################################################
import numpy as np
import statsmodels.api as sm
import matplotlib.pyplot as plt
config.load_config_small()
figure_size = [1*text_width, 1*text_height]
n_columns = 54
n_rows = 12 # 36
rows = np.array([[0, 12], [12, 24], [24, 36]])
#labels = np.arange(0, 54, 1)
fig.savefig(plotname+".pgf", pad_inches=0, dpi=fig.dpi) # pgf
plt.close()
##################################
# 2D
##################################
figure_width = 0.66*text_width
figure_height = figure_width # (figure_width / golden_ratio) #figure_width
#figure_height = 0.75 * figure_width
figure_size = [figure_width, figure_height]
config.load_config_small()
# Custom colormap UIBK Orange
cdict = {'red': ((0.0, 1.0, 1.0),
(1.0, 1.0, 1.0)),
'green': ((0.0, 1.0, 1.0),
(1.0, 0.5, 0.5)),
'blue': ((0.0, 1.0, 1.0),
(1.0, 0.0, 0.0))}
plt.register_cmap(name='UIBK_ORANGES', data=cdict)
p = 5
x = np.linspace(-1, 1, 256)
def plot_convolution3(plotname, tsframe, paperwidth):
colormap = plt.get_cmap('YlOrRd_r')
colormap.set_under([0.2, 0.2, 0.2])
figure_width = paperwidth
figure_height = paperwidth
figure_size = [figure_width, figure_height]
config.load_config_small()
# find peak and scale neighbourhood
poi = np.asarray(np.unravel_index(np.argmax(tsframe), tsframe.shape))
px = poi[1]
py = poi[0]
r = 1
scalefactor = 4.0
peak_area = tsframe[py - r:py + r + 1, px - r:px + r + 1]
peak_area = scipy.ndimage.zoom(peak_area, scalefactor, order=3)
# Point of interest of peak_area
poi_area = np.asarray(
np.unravel_index(np.argmax(peak_area), peak_area.shape))
px_area = poi_area[1]
py_area = poi_area[0]
offset_x_area = (scalefactor * (2 * r + 1) - 1) / 2.0 - px_area
offset_y_area = (scalefactor * (2 * r + 1) - 1) / 2.0 - py_area
width = peak_area.shape[1]
height = peak_area.shape[0]
xs, ys = np.meshgrid(np.arange(0, width + 1), np.arange(0, height + 1))
fig, axes = plt.subplots(nrows=1,
ncols=1,
sharex=False,
sharey=False,
squeeze=True,
figsize=figure_size,
dpi=100)
ax = axes
# Workaround inverted y-axis
ax.invert_yaxis()
# pcolormesh aligns cells on their edges, while imshow aligns them on their centers.
ax.pcolormesh(xs - 0.5,
ys - 0.5,
peak_area,
cmap=colormap,
vmin=0.1,
vmax=np.max(peak_area),
shading="faceted",
linestyle="dashed",
linewidth=0.5,
edgecolor=[0.0, 0.0, 0.0])
# Absolute number
#for i,j in ((x,y) for x in np.arange(0, len(tsframe))
# for y in np.arange(0, len(tsframe[0]))):
# if tsframe[i][j] >= 1:
# ax.annotate(str(int(tsframe[i][j])), xy=(j,i), fontsize=3.5, ha='center', va='center')
r = 3.5
#extent1 = patches.Rectangle((-0.48, -0.48), 11.96, 11.96, lw=1.5, ec=[0.0, 0.0, 0.0, 1.0], fc=[1.0, 1.0, 1.0, 0.0], zorder=98 )
#extent2 = patches.Rectangle((-0.48, -0.48), 11.96, 11.96, lw=0.5, ec=[0.0, 0.0, 1.0, 1.0], fc=[0.0, 0.0, 0.0, 0.0], zorder=99 )
#ax.add_patch(extent1)
#ax.add_patch(extent2)
kernel1 = patches.Rectangle((px_area - r, py_area - r),
2 * r,
2 * r,
lw=1.5,
ec=[0.0, 0.0, 0.0, 1.0],
fc=[1.0, 1.0, 1.0, 0.0])
kernel2 = patches.Rectangle((px_area - r, py_area - r),
2 * r,
2 * r,
lw=0.5,
ec=[0.0, 1.0, 0.0, 1.0],
fc=[0.0, 0.0, 0.0, 0.0])
ax.add_patch(kernel1)
ax.add_patch(kernel2)
ax.set_aspect('equal')
ax.set_xlim([-0.5, width - 0.5])
ax.set_ylim([height - 0.5, -0.5])
#ax.plot(px_area, py_area, "x", ms=2, mew=0.3, color=[1.0, 1.0, 1.0, 1.0], mfc='None')
ax.plot(px_area,
py_area,
"o",
ms=1.5,
mew=0.3,
mec=[0.0, 0.0, 0.0, 1.0],
mfc=[0.0, 1.0, 0.0, 1.0])
ax.plot(px_area + offset_x_area,
py_area + offset_y_area,
"o",
ms=1.5,
mew=0.3,
mec=[0.0, 0.0, 0.0, 1.0],
mfc=[0.0, 0.0, 1.0, 1.0])
ax.xaxis.tick_top()
ax.tick_params(axis='both',
which='both',
left='off',
right='off',
bottom='off',
top='off',
labeltop='on',
pad=1,
labelsize=4)
ax.xaxis.set_major_locator(plt.NullLocator())
ax.yaxis.set_major_locator(plt.NullLocator())
fig.tight_layout()
fig.savefig(plotname + ".pdf",
pad_inches=0,
bbox_inches='tight',
dpi=fig.dpi) # pdf
fig.savefig(plotname + ".pgf",
pad_inches=0,
bbox_inches='tight',
dpi=fig.dpi) # pgf
def plot_convolution2(plotname, tsframe, tsframe_max, paperwidth):
colormap = plt.get_cmap('YlOrRd_r')
colormap.set_under([0.2, 0.2, 0.2])
figure_width = paperwidth
figure_height = paperwidth
figure_size = [figure_width, figure_height]
config.load_config_small()
width = tsframe.shape[1]
height = tsframe.shape[0]
xs, ys = np.meshgrid(np.arange(0, width + 1), np.arange(0, height + 1))
fig, axes = plt.subplots(nrows=1,
ncols=1,
sharex=False,
sharey=False,
squeeze=True,
figsize=figure_size,
dpi=100)
ax = axes
# Workaround inverted y-axis
ax.invert_yaxis()
# pcolormesh aligns cells on their edges, while imshow aligns them on their centers.
ax.pcolormesh(xs - 0.5,
ys - 0.5,
tsframe,
cmap=colormap,
vmin=0.1,
vmax=tsframe_max,
shading="faceted",
linestyle="dashed",
linewidth=0.5,
edgecolor=[0.0, 0.0, 0.0])
# Absolute number
#for i,j in ((x,y) for x in np.arange(0, len(tsframe))
# for y in np.arange(0, len(tsframe[0]))):
# if tsframe[i][j] >= 1:
# ax.annotate(str(int(tsframe[i][j])), xy=(j,i), fontsize=3.5, ha='center', va='center')
#find peak
poi = np.asarray(np.unravel_index(np.argmax(tsframe), tsframe.shape))
px = poi[1]
py = poi[0]
ax.plot(px,
py,
"o",
ms=1.5,
mew=0.3,
mec=[0.0, 0.0, 0.0, 1.0],
mfc=[0.0, 0.0, 1.0, 1.0])
ax.plot(px - 1,
py,
"o",
ms=1.5,
mew=0.3,
mec=[0.0, 0.0, 0.0, 1.0],
mfc=[1.0, 1.0, 1.0, 1.0])
ax.plot(px + 1,
py,
"o",
ms=1.5,
mew=0.3,
mec=[0.0, 0.0, 0.0, 1.0],
mfc=[1.0, 1.0, 1.0, 1.0])
ax.plot(px,
py - 1,
"o",
ms=1.52,
mew=0.3,
mec=[0.0, 0.0, 0.0, 1.0],
mfc=[1.0, 1.0, 1.0, 1.0])
ax.plot(px,
py + 1,
"o",
ms=1.5,
mew=0.3,
mec=[0.0, 0.0, 0.0, 1.0],
mfc=[1.0, 1.0, 1.0, 1.0])
cx = (tsframe.shape[1] - 1) / 2
cy = (tsframe.shape[0] - 1) / 2
ax.plot(cx, cy, "x", ms=2, mew=0.3, color=[1.0, 1.0, 1.0, 1.0], mfc='None')
r = 1.5
kernel1 = patches.Rectangle((px - r, py - r),
2 * r,
2 * r,
lw=1.5,
ec=[0.0, 0.0, 0.0, 1.0],
fc=[1.0, 1.0, 1.0, 0.0])
kernel2 = patches.Rectangle((px - r, py - r),
2 * r,
2 * r,
lw=0.5,
ec=[0.0, 0.0, 1.0, 1.0],
fc=[0.0, 0.0, 0.0, 0.0])
ax.add_patch(kernel1)
ax.add_patch(kernel2)
ax.set_aspect('equal')
ax.set_xlim([-0.5, width - 0.5])
ax.set_ylim([height - 0.5, -0.5])
ax.xaxis.tick_top()
ax.tick_params(axis='both',
which='both',
left='off',
right='off',
bottom='off',
top='off',
labeltop='on',
pad=1,
labelsize=4)
#ax.xaxis.set_major_locator(plt.NullLocator())
#ax.yaxis.set_major_locator(plt.NullLocator())
fig.tight_layout()
fig.savefig(plotname + ".pdf",
pad_inches=0,
bbox_inches='tight',
dpi=fig.dpi) # pdf
fig.savefig(plotname + ".pgf",
pad_inches=0,
bbox_inches='tight',
dpi=fig.dpi) # pgf
def plot_convolution(plotname, tsframe, tsframe_max, paperwidth):
colormap = plt.get_cmap('YlOrRd_r')
colormap.set_under([0.2, 0.2, 0.2])
figure_width = paperwidth
figure_height = paperwidth
figure_size = [figure_width, figure_height]
config.load_config_small()
width = tsframe.shape[1]
height = tsframe.shape[0]
xs, ys = np.meshgrid(np.arange(0, width + 1), np.arange(0, height + 1))
fig, axes = plt.subplots(nrows=1,
ncols=1,
sharex=False,
sharey=False,
squeeze=True,
figsize=figure_size,
dpi=100)
ax = axes
# Workaround inverted y-axis
ax.invert_yaxis()
# pcolormesh aligns cells on their edges, while imshow aligns them on their centers.
ax.pcolormesh(xs - 0.5,
ys - 0.5,
tsframe,
cmap=colormap,
vmin=0.1,
vmax=tsframe_max,
shading="faceted",
linestyle="dashed",
linewidth=0.5,
edgecolor=[0.0, 0.0, 0.0])
# Absolute number
#for i,j in ((x,y) for x in np.arange(0, len(tsframe))
# for y in np.arange(0, len(tsframe[0]))):
# if tsframe[i][j] >= 1:
# ax.annotate(str(int(tsframe[i][j])), xy=(j,i), fontsize=3.5, ha='center', va='center')
ax.set_aspect('equal')
ax.set_xlim([-0.5, width - 0.5])
ax.set_ylim([height - 0.5, -0.5])
ax.xaxis.tick_top()
ax.tick_params(axis='both',
which='both',
left='off',
right='off',
bottom='off',
top='off',
labeltop='on',
pad=1,
labelsize=4)
#ax.xaxis.set_major_locator(plt.NullLocator())
#ax.yaxis.set_major_locator(plt.NullLocator())
fig.tight_layout()
fig.savefig(plotname + ".pdf",
pad_inches=0,
bbox_inches='tight',
dpi=fig.dpi) # pdf
fig.savefig(plotname + ".pgf",
pad_inches=0,
bbox_inches='tight',
dpi=fig.dpi) # pgf
def plot_convolution3(plotname, tsframe, paperwidth):
colormap=plt.get_cmap('YlOrRd_r')
colormap.set_under([0.2, 0.2, 0.2])
figure_width = paperwidth
figure_height = paperwidth
figure_size = [figure_width, figure_height]
config.load_config_small()
# find peak and scale neighbourhood
poi = np.asarray(np.unravel_index(np.argmax(tsframe), tsframe.shape))
px = poi[1]
py = poi[0]
r = 1
scalefactor = 4.0
peak_area = tsframe[py-r:py+r+1, px-r:px+r+1]
peak_area = scipy.ndimage.zoom(peak_area, scalefactor, order=3)
# Point of interest of peak_area
poi_area = np.asarray(np.unravel_index(np.argmax(peak_area), peak_area.shape))
px_area = poi_area[1]
py_area = poi_area[0]
offset_x_area = (scalefactor*(2*r+1)-1)/2.0 - px_area
offset_y_area = (scalefactor*(2*r+1)-1)/2.0 - py_area
width = peak_area.shape[1]; height = peak_area.shape[0]
xs,ys = np.meshgrid(np.arange(0, width+1), np.arange(0, height+1))
fig, axes = plt.subplots(nrows=1, ncols=1, sharex=False, sharey=False, squeeze=True, figsize=figure_size, dpi=100)
ax = axes
# Workaround inverted y-axis
ax.invert_yaxis()
# pcolormesh aligns cells on their edges, while imshow aligns them on their centers.
ax.pcolormesh(xs-0.5, ys-0.5, peak_area, cmap=colormap, vmin=0.1, vmax=np.max(peak_area),
shading="faceted", linestyle="dashed", linewidth=0.5, edgecolor=[0.0, 0.0, 0.0])
# Absolute number
#for i,j in ((x,y) for x in np.arange(0, len(tsframe))
# for y in np.arange(0, len(tsframe[0]))):
# if tsframe[i][j] >= 1:
# ax.annotate(str(int(tsframe[i][j])), xy=(j,i), fontsize=3.5, ha='center', va='center')
r = 3.5
#extent1 = patches.Rectangle((-0.48, -0.48), 11.96, 11.96, lw=1.5, ec=[0.0, 0.0, 0.0, 1.0], fc=[1.0, 1.0, 1.0, 0.0], zorder=98 )
#extent2 = patches.Rectangle((-0.48, -0.48), 11.96, 11.96, lw=0.5, ec=[0.0, 0.0, 1.0, 1.0], fc=[0.0, 0.0, 0.0, 0.0], zorder=99 )
#ax.add_patch(extent1)
#ax.add_patch(extent2)
kernel1 = patches.Rectangle((px_area-r, py_area-r), 2*r, 2*r, lw=1.5, ec=[0.0, 0.0, 0.0, 1.0], fc=[1.0, 1.0, 1.0, 0.0] )
kernel2 = patches.Rectangle((px_area-r, py_area-r), 2*r, 2*r, lw=0.5, ec=[0.0, 1.0, 0.0, 1.0], fc=[0.0, 0.0, 0.0, 0.0] )
ax.add_patch(kernel1)
ax.add_patch(kernel2)
ax.set_aspect('equal')
ax.set_xlim([-0.5, width-0.5])
ax.set_ylim([height-0.5, -0.5])
#ax.plot(px_area, py_area, "x", ms=2, mew=0.3, color=[1.0, 1.0, 1.0, 1.0], mfc='None')
ax.plot(px_area, py_area, "o", ms=1.5, mew=0.3, mec=[0.0, 0.0, 0.0, 1.0], mfc=[0.0, 1.0, 0.0, 1.0])
ax.plot(px_area+offset_x_area, py_area+offset_y_area, "o", ms=1.5, mew=0.3, mec=[0.0, 0.0, 0.0, 1.0], mfc=[0.0, 0.0, 1.0, 1.0])
ax.xaxis.tick_top()
ax.tick_params(axis='both', which='both', left='off', right='off', bottom='off', top='off', labeltop='on',
pad=1, labelsize=4)
ax.xaxis.set_major_locator(plt.NullLocator())
ax.yaxis.set_major_locator(plt.NullLocator())
fig.tight_layout()
fig.savefig(plotname+".pdf", pad_inches=0, bbox_inches='tight', dpi=fig.dpi) # pdf
fig.savefig(plotname+".pgf", pad_inches=0, bbox_inches='tight', dpi=fig.dpi) # pgf
def plot_convolution2(plotname, tsframe, tsframe_max, paperwidth):
colormap=plt.get_cmap('YlOrRd_r')
colormap.set_under([0.2, 0.2, 0.2])
figure_width = paperwidth
figure_height = paperwidth
figure_size = [figure_width, figure_height]
config.load_config_small()
width = tsframe.shape[1]; height = tsframe.shape[0]
xs,ys = np.meshgrid(np.arange(0, width+1), np.arange(0, height+1))
fig, axes = plt.subplots(nrows=1, ncols=1, sharex=False, sharey=False, squeeze=True, figsize=figure_size, dpi=100)
ax = axes
# Workaround inverted y-axis
ax.invert_yaxis()
# pcolormesh aligns cells on their edges, while imshow aligns them on their centers.
ax.pcolormesh(xs-0.5, ys-0.5, tsframe, cmap=colormap, vmin=0.1, vmax=tsframe_max,
shading="faceted", linestyle="dashed", linewidth=0.5, edgecolor=[0.0, 0.0, 0.0])
# Absolute number
#for i,j in ((x,y) for x in np.arange(0, len(tsframe))
# for y in np.arange(0, len(tsframe[0]))):
# if tsframe[i][j] >= 1:
# ax.annotate(str(int(tsframe[i][j])), xy=(j,i), fontsize=3.5, ha='center', va='center')
#find peak
poi = np.asarray(np.unravel_index(np.argmax(tsframe), tsframe.shape))
px = poi[1]
py = poi[0]
ax.plot(px, py, "o", ms=1.5, mew=0.3, mec=[0.0, 0.0, 0.0, 1.0], mfc=[0.0, 0.0, 1.0, 1.0])
ax.plot(px-1, py, "o", ms=1.5, mew=0.3, mec=[0.0, 0.0, 0.0, 1.0], mfc=[1.0, 1.0, 1.0, 1.0])
ax.plot(px+1, py, "o", ms=1.5, mew=0.3, mec=[0.0, 0.0, 0.0, 1.0], mfc=[1.0, 1.0, 1.0, 1.0])
ax.plot(px, py-1, "o", ms=1.52, mew=0.3, mec=[0.0, 0.0, 0.0, 1.0], mfc=[1.0, 1.0, 1.0, 1.0])
ax.plot(px, py+1, "o", ms=1.5, mew=0.3, mec=[0.0, 0.0, 0.0, 1.0], mfc=[1.0, 1.0, 1.0, 1.0])
cx = (tsframe.shape[1]-1)/2
cy = (tsframe.shape[0]-1)/2
ax.plot(cx, cy, "x", ms=2, mew=0.3, color=[1.0, 1.0, 1.0, 1.0], mfc='None')
r = 1.5
kernel1 = patches.Rectangle((px-r, py-r), 2*r, 2*r, lw=1.5, ec=[0.0, 0.0, 0.0, 1.0], fc=[1.0, 1.0, 1.0, 0.0] )
kernel2 = patches.Rectangle((px-r, py-r), 2*r, 2*r, lw=0.5, ec=[0.0, 0.0, 1.0, 1.0], fc=[0.0, 0.0, 0.0, 0.0] )
ax.add_patch(kernel1)
ax.add_patch(kernel2)
ax.set_aspect('equal')
ax.set_xlim([-0.5, width-0.5])
ax.set_ylim([height-0.5, -0.5])
ax.xaxis.tick_top()
ax.tick_params(axis='both', which='both', left='off', right='off', bottom='off', top='off', labeltop='on',
pad=1, labelsize=4)
#ax.xaxis.set_major_locator(plt.NullLocator())
#ax.yaxis.set_major_locator(plt.NullLocator())
fig.tight_layout()
fig.savefig(plotname+".pdf", pad_inches=0, bbox_inches='tight', dpi=fig.dpi) # pdf
fig.savefig(plotname+".pgf", pad_inches=0, bbox_inches='tight', dpi=fig.dpi) # pgf
# Create legend from custom artist/label lists
#plt.figlegend([artist for artist in artists_category], [label for label in unique_labels],
# fontsize='x-small', labelspacing=0.5, handletextpad=0.05,
# loc='center right', bbox_to_anchor=(0, 0.5, 1, 1), bbox_transform=plt.gcf().transFigure )
legend = plt.legend([artist for artist in artists_category], [label for label in unique_labels], ncol=3, loc='center',
fontsize='small', labelspacing=0.5, handletextpad=0.05, columnspacing=0.5, borderpad=0.5,
fancybox=True, shadow=False, framealpha=1.0)
plotname = "scatterplotmatrix_legend"
fig.savefig(plotname+".pdf", pad_inches=0, dpi=fig.dpi) # pdf
fig.savefig(plotname+".pgf", pad_inches=0, dpi=fig.dpi) # pgf
plt.close()
'''
'''
#######################################################
# Bean plot (standardized features)
#######################################################
import numpy as np
import statsmodels.api as sm
import matplotlib.pyplot as plt
config.load_config_small()
figure_size = [1*text_width, 1*text_height]
n_columns = 54
n_rows = 12 # 36
rows = np.array([[0, 12], [12, 24], [24, 36]])
#labels = np.arange(0, 54, 1)
