def _decorate_histogram(vstats):
import pylab
from matplotlib.transforms import blended_transform_factory as blend
# Shade things inside 1-sigma
pylab.axvspan(vstats.p68[0],vstats.p68[1],
color='gold',alpha=0.5,zorder=-1)
# build transform with x=data, y=axes(0,1)
ax = pylab.gca()
transform = blend(ax.transData, ax.transAxes)
l95,h95 = vstats.p95
l68,h68 = vstats.p68
def marker(s,v):
if v < l95: s,v,ha = '<'+s,l95,'left'
elif v > h95: s,v,ha = '>'+s,h95,'right'
else: ha='center'
pylab.text(v, 0.95, s, va='top', ha=ha,
transform=transform, zorder=3, color='g')
#pylab.axvline(v)
marker('|',vstats.median)
marker('E',vstats.mean)
marker('*',vstats.best)
pylab.text(0.01, 0.95, vstats.label, zorder=2,
backgroundcolor=(1,1,0,0.2),
verticalalignment='top',
horizontalalignment='left',
transform=pylab.gca().transAxes)
pylab.setp([pylab.gca().get_yticklabels()],visible=False)
ticks = (l95, l68, vstats.median, h68, h95)
labels = [format_value(v,h95-l95) for v in ticks]
if len(labels[2]) > 5:
# Drop 68% values if too many digits
ticks,labels= ticks[0::2],labels[0::2]
pylab.xticks(ticks, labels)
def _decorate_histogram(vstats):
import pylab
from matplotlib.transforms import blended_transform_factory as blend
# Shade things inside 1-sigma
pylab.axvspan(vstats.p68[0],
vstats.p68[1],
color='gold',
alpha=0.5,
zorder=-1)
# build transform with x=data, y=axes(0,1)
ax = pylab.gca()
transform = blend(ax.transData, ax.transAxes)
l95, h95 = vstats.p95
l68, h68 = vstats.p68
def marker(symbol, position):
if position < l95:
symbol, position, ha = '<' + symbol, l95, 'left'
elif position > h95:
symbol, position, ha = '>' + symbol, h95, 'right'
else:
symbol, position, ha = symbol, position, 'center'
pylab.text(position,
0.95,
symbol,
va='top',
ha=ha,
transform=transform,
zorder=3,
color='g')
#pylab.axvline(v)
marker('|', vstats.median)
marker('E', vstats.mean)
marker('*', vstats.best)
pylab.text(0.01,
0.95,
vstats.label,
zorder=2,
backgroundcolor=(1, 1, 0, 0.2),
verticalalignment='top',
horizontalalignment='left',
transform=pylab.gca().transAxes)
pylab.setp([pylab.gca().get_yticklabels()], visible=False)
ticks = (l95, l68, vstats.median, h68, h95)
labels = [format_value(v, h95 - l95) for v in ticks]
if len(labels[2]) > 5:
# Drop 68% values if too many digits
ticks, labels = ticks[0::2], labels[0::2]
pylab.xticks(ticks, labels)
def _decorate_histogram(vstats):
import pylab
from matplotlib.transforms import blended_transform_factory as blend
l95, h95 = vstats.p95_range
l68, h68 = vstats.p68_range
# Shade things inside 1-sigma
pylab.axvspan(l68, h68, color='gold', alpha=0.5, zorder=-1)
# build transform with x=data, y=axes(0,1)
ax = pylab.gca()
transform = blend(ax.transData, ax.transAxes)
def marker(symbol, position):
if position < l95:
symbol, position, ha = '<' + symbol, l95, 'left'
elif position > h95:
symbol, position, ha = '>' + symbol, h95, 'right'
else:
symbol, position, ha = symbol, position, 'center'
pylab.text(position,
0.95,
symbol,
va='top',
ha=ha,
transform=transform,
zorder=3,
color='g')
#pylab.axvline(v)
marker('|', vstats.median)
marker('E', vstats.mean)
marker('*', vstats.best)
pylab.text(0.01,
0.95,
vstats.label,
zorder=2,
backgroundcolor=(1, 1, 0, 0.2),
verticalalignment='top',
horizontalalignment='left',
transform=pylab.gca().transAxes)
ax.set_yticklabels([])
def annotate(entry0, entry1, dim, minfvalue=1e-8, nbtests=1):
"""Display some annotations associated to the graphs generated."""
isEarlyStop = False
ha = 'left'
va = 'center'
lastfvalue = min(entry0.evals[-1][0], entry1.evals[-1][0])
if minfvalue < lastfvalue:
isEarlyStop = True
#ha = 'center'
#va = 'top'
if not minfvalue or minfvalue < lastfvalue:
minfvalue = lastfvalue
line = []
data0 = entry0.detEvals([minfvalue])[0]
evals0 = data0.copy()
succ = (np.isnan(evals0) == False)
evals0[np.isnan(evals0)] = entry0.maxevals[np.isnan(evals0)]
line.append(toolsstats.sp(evals0, issuccessful=succ))
data1 = entry1.detEvals([minfvalue])[0]
evals1 = data1.copy()
succ = (np.isnan(evals1) == False)
evals1[np.isnan(evals1)] = entry1.maxevals[np.isnan(evals1)]
line.append(toolsstats.sp(evals1, issuccessful=succ))
# What's the situation?
txt = '%dD' % dim
if (line[0][2] > 0 and line[1][2] > 0 and line[1][2] < 10):
tmp = str(int(line[1][2]))
tmp2 = str(int(line[0][2]))
txt = tmp + '/' + tmp2
dims = dimension_index
ax = plt.gca()
assert line[0][2] > 0 or line[1][2] > 0
signdata = line[1][0] - line[0][0]
if line[0][2] > 0 and line[1][2] > 0:
trans = ax.transData
annotcoord = [minfvalue, line[1][0] / line[0][0]]
elif line[0][2] == 0:
trans = blend(ax.transData, ax.transAxes)
annotcoord = [
minfvalue, -line[1][1] / 2 + 0.5 + offset * (5 - dims[dim])
]
#if va == 'top':
# va = 'bottom'
elif line[1][2] == 0:
trans = blend(ax.transData, ax.transAxes)
annotcoord = [
minfvalue, line[0][1] / 2 + 0.5 - offset * (5 - dims[dim])
]
plt.text(annotcoord[0],
annotcoord[1],
txt,
horizontalalignment=ha,
verticalalignment=va,
transform=trans)
#ranksum test
line0 = np.power(data0, -1.)
line0[np.isnan(line0)] = -entry0.finalfunvals[np.isnan(line0)]
line1 = np.power(data1, -1.)
line1[np.isnan(line1)] = -entry1.finalfunvals[np.isnan(line1)]
# one-tailed statistics: scipy.stats.mannwhitneyu, two-tailed statistics: scipy.stats.ranksumtest
z, p = ranksumtest(line0, line1)
# Set the correct line in data0 and data1
nbstars = 0
# sign of z-value and data must agree
if ((nbtests * p) < 0.05 and (z * signdata) > 0):
nbstars = np.min([5, -np.ceil(np.log10(nbtests * p + 1e-99))])
if nbstars > 0:
xstars = annotcoord[0] * np.power(incrstars, np.arange(
1., 1. + nbstars))
# the additional slicing [0:int(nbstars)] is due to
# np.arange(1., 1. - 0.1 * nbstars, -0.1) not having the right number
# of elements due to numerical error
ystars = [annotcoord[1]] * nbstars
try:
h = plt.plot(xstars,
ystars,
marker='*',
ls='',
color='w',
markersize=5 * linewidth,
markeredgecolor='k',
markerfacecolor='None',
zorder=20,
markeredgewidth=0.4 * linewidth,
transform=trans,
clip_on=False)
except KeyError:
#Version problem
h = plt.plot(xstars,
ystars,
marker='+',
ls='',
color='w',
markersize=2.5 * linewidth,
markeredgecolor='k',
zorder=20,
markeredgewidth=0.2 * linewidth,
transform=trans,
clip_on=False)
def plot(xdata, ydata):
"""Plot the ERT log loss figures.
Two cases: box-whisker plot is used for representing the data of all
functions, otherwise all data is represented using crosses.
"""
res = []
tmp = list(10**np.mean(i[np.isfinite(i)]) for i in ydata)
res.extend(plt.plot(xdata, tmp, ls='-', color='k', lw=3, #marker='+',
markersize=20, markeredgewidth=3))
if max(len(i) for i in ydata) < 20: # TODO: subgroups of function, hopefully.
for i, y in enumerate(ydata):
# plot all single data points
if (np.isfinite(y)==False).any():
assert not (np.isinf(y) * y > 0.).any()
assert not np.isnan(y).any()
ax = plt.gca()
trans = blend(ax.transData, ax.transAxes)
res.extend(plt.plot((xdata[i], ), (0., ),
marker='+', color=flierscolor,
ls='', markersize=20, markeredgewidth=3,
transform=trans, clip_on=False))
res.append(plt.text(xdata[i], 0.02, '%d' % len(y[np.isinf(y)]),
transform=trans, horizontalalignment='left',
verticalalignment='bottom'))
y = y[np.isfinite(y)]
if len(y) == 0:
continue
res.extend(plt.plot([xdata[i]]*len(y), 10**np.array(y),
marker='+', color=flierscolor,
ls='', markersize=20, markeredgewidth=3))
# plot dashed vertical line between min and max
plt.plot([xdata[i]]*2, 10**np.array([min(y), max(y)]),
color='k', # marker='+',
ls='--', linewidth=2) #, markersize=20, markeredgewidth=3)
# plot min and max with different symbol
#plt.plot([xdata[i]], 10**min(np.array(y)),
# marker='+', color='k',
# ls='', markersize=20, markeredgewidth=3)
#plt.plot([xdata[i]], 10**max(np.array(y)),
# marker='+', color='k',
# ls='', markersize=20, markeredgewidth=3)
else:
for i, y in enumerate(ydata):
# plot all single data points
if (np.isfinite(y)==False).any():
assert not (np.isinf(y) * y > 0.).any()
assert not np.isnan(y).any()
ax = plt.gca()
trans = blend(ax.transData, ax.transAxes)
res.extend(plt.plot((xdata[i], ), (0, ),
marker='.', color='k',
ls='', markersize=20, markeredgewidth=3,
transform=trans, clip_on=False))
res.append(plt.text(xdata[i], 0.02, '%d' % len(y[np.isinf(y)]),
transform=trans, horizontalalignment='left',
verticalalignment='bottom'))
y = y[np.isfinite(y)]
dictboxwhisker = boxplot(list(10**np.array(i) for i in ydata),
sym='', notch=0, widths=None,
positions=xdata)
#'medians', 'fliers', 'whiskers', 'boxes', 'caps'
plt.setp(dictboxwhisker['medians'], lw=3)
plt.setp(dictboxwhisker['boxes'], lw=3)
plt.setp(dictboxwhisker['caps'], lw=3)
plt.setp(dictboxwhisker['whiskers'], lw=3)
for i in dictboxwhisker.values():
res.extend(i)
res.extend(plt.plot(xdata, list(10**min(i) for i in ydata), marker='.',
markersize=20, color='k', ls=''))
res.extend(plt.plot(xdata, list(10**max(i) for i in ydata), marker='.',
markersize=20, color='k', ls=''))
return res
def main(dsList0, dsList1, outputdir, verbose=True):
"""Generate a scatter plot figure.
TODO: """
markers = genericsettings.dim_related_markers
colors = genericsettings.dim_related_colors
#plt.rc("axes", labelsize=24, titlesize=24)
#plt.rc("xtick", labelsize=20)
#plt.rc("ytick", labelsize=20)
#plt.rc("font", size=20)
#plt.rc("legend", fontsize=20)
dictFunc0 = dsList0.dictByFunc()
dictFunc1 = dsList1.dictByFunc()
funcs = set(dictFunc0.keys()) & set(dictFunc1.keys())
targets = testbedsettings.current_testbed.ppscatter_target_values
if isinstance(targets, pproc.RunlengthBasedTargetValues):
linewidth = linewidth_rld_based
else:
linewidth = linewidth_default
funInfos = ppfigparam.read_fun_infos()
for f in funcs:
dictDim0 = dictFunc0[f].dictByDim()
dictDim1 = dictFunc1[f].dictByDim()
dims = set(dictDim0.keys()) & set(dictDim1.keys())
#set_trace()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
if linewidth: # plot all reliable aRT values as a line
all_targets = np.array(sorted(set(entry0.target).union(entry1.target), reverse=True))
assert entry0.detSuccessRates([all_targets[0]]) == 1.0
assert entry1.detSuccessRates([all_targets[0]]) == 1.0
all_targets = all_targets[np.where(all_targets <= targets((f, d))[0])[0]] #
xdata_all = np.array(entry0.detERT(all_targets))
ydata_all = np.array(entry1.detERT(all_targets))
# idx of reliable targets: last index where success rate >= 1/2 and aRT <= maxevals
idx = []
for ari in (np.where(entry0.detSuccessRates(all_targets) >= 0.5)[0],
np.where(entry1.detSuccessRates(all_targets) >= 0.5)[0],
np.where(xdata_all <= max(entry0.maxevals))[0],
np.where(ydata_all <= max(entry1.maxevals))[0]
):
if len(ari):
idx.append(ari[-1])
if len(idx) == 4:
max_idx = min(idx)
## at least up to the most difficult given target
## idx = max((idx, np.where(all_targets >= targets((f, d))[-1])[0][-1]))
xdata_all = xdata_all[:max_idx + 1]
ydata_all = ydata_all[:max_idx + 1]
idx = (numpy.isfinite(xdata_all)) * (numpy.isfinite(ydata_all))
assert idx.all()
if idx.any():
plt.plot(xdata_all[idx], ydata_all[idx], colors[i], ls='solid', lw=linewidth,
# TODO: ls has changed, check whether this works out
clip_on=False)
xdata = numpy.array(entry0.detERT(targets((f, d))))
ydata = numpy.array(entry1.detERT(targets((f, d))))
# plot "valid" data, those within maxevals
idx = np.logical_and(xdata < entry0.mMaxEvals(),
ydata < entry1.mMaxEvals())
# was:
# (numpy.isinf(xdata) == False) *
# (numpy.isinf(ydata) == False) *
# (xdata < entry0.mMaxEvals()) *
# (ydata < entry1.mMaxEvals()))
if idx.any():
try:
plt.plot(xdata[idx], ydata[idx], ls='',
markersize=markersize,
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
clip_on=False)
except KeyError:
plt.plot(xdata[idx], ydata[idx], ls='', markersize=markersize,
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
clip_on=False)
#try:
# plt.scatter(xdata[idx], ydata[idx], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3)
#except ValueError:
# set_trace()
# plot beyond maxevals but finite data
idx = ((numpy.isinf(xdata) == False) *
(numpy.isinf(ydata) == False) *
np.logical_or(xdata >= entry0.mMaxEvals(),
ydata >= entry1.mMaxEvals()))
if idx.any():
try:
plt.plot(xdata[idx], ydata[idx], ls='',
markersize=markersize + markersize_addon_beyond_maxevals,
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=1,
clip_on=False)
except KeyError:
plt.plot(xdata[idx], ydata[idx], ls='', markersize=markersize,
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=2,
clip_on=False)
#ax = plt.gca()
ax = plt.axes()
# plot data on the right edge
idx = numpy.isinf(xdata) * (numpy.isinf(ydata) == False)
if idx.any():
# This (seems to) transform inf to the figure limits!?
trans = blend(ax.transAxes, ax.transData)
#plt.scatter([1.]*numpy.sum(idx), ydata[idx], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1.]*numpy.sum(idx), ydata[idx],
markersize=markersize + markersize_addon_beyond_maxevals, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=1,
transform=trans, clip_on=False)
except KeyError:
plt.plot([1.]*numpy.sum(idx), ydata[idx],
markersize=markersize, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=2,
transform=trans, clip_on=False)
#set_trace()
# plot data on the left edge
idx = (numpy.isinf(xdata)==False) * numpy.isinf(ydata)
if idx.any():
# This (seems to) transform inf to the figure limits!?
trans = blend(ax.transData, ax.transAxes)
# plt.scatter(xdata[idx], [1.-offset]*numpy.sum(idx), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot(xdata[idx], [1.-offset]*numpy.sum(idx),
markersize=markersize + markersize_addon_beyond_maxevals, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=1,
transform=trans, clip_on=False)
except KeyError:
plt.plot(xdata[idx], [1.-offset]*numpy.sum(idx),
markersize=markersize, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=2,
transform=trans, clip_on=False)
# plot data in the top corner
idx = numpy.isinf(xdata) * numpy.isinf(ydata)
if idx.any():
# plt.scatter(xdata[idx], [1.-offset]*numpy.sum(idx), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1.-offset]*numpy.sum(idx), [1.-offset]*numpy.sum(idx),
markersize=markersize + markersize_addon_beyond_maxevals, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=1,
transform=ax.transAxes, clip_on=False)
except KeyError:
plt.plot([1.-offset]*numpy.sum(idx), [1.-offset]*numpy.sum(idx),
markersize=markersize, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=2,
transform=ax.transAxes, clip_on=False)
targetlabels = targets.labels()
if isinstance(targets, pproc.RunlengthBasedTargetValues):
text = (str(len(targetlabels)) + ' target RLs/dim: ' +
targetlabels[0] + '..' +
targetlabels[len(targetlabels)-1] + '\n')
text += ' from ' + testbedsettings.current_testbed.best_algorithm_filename
else:
text = (str(len(targetlabels)) + ' targets in ' +
targetlabels[0] + '..' +
targetlabels[len(targetlabels)-1])
# add number of instances
text += '\n'
num_of_instances_alg0 = []
num_of_instances_alg1 = []
for d in dims:
num_of_instances_alg0.append((dictDim0[d][0]).nbRuns())
num_of_instances_alg1.append((dictDim1[d][0]).nbRuns())
# issue a warning if the numbers of instances are inconsistent:
if (len(set(num_of_instances_alg0)) > 1):
warnings.warn('Inconsistent numbers of instances over dimensions found for ALG0:\n\
found instances %s' % str(num_of_instances_alg0))
if (len(set(num_of_instances_alg1)) > 1):
warnings.warn('Inconsistent numbers of instances over dimensions found for ALG1:\n\
found instances %s' % str(num_of_instances_alg1))
if (len(set(num_of_instances_alg0)) == 1 and len(set(num_of_instances_alg0)) == 1):
text += '%s and %s instances' % (num_of_instances_alg0[0], num_of_instances_alg1[0])
else:
for n in num_of_instances_alg0:
text += '%d, ' % n
text = text.rstrip(', ')
text += ' and '
for n in num_of_instances_alg1:
text += '%d, ' % n
text = text.rstrip(', ')
text += ' instances'
plt.text(0.01, 0.98, text, horizontalalignment="left",
verticalalignment="top", transform=plt.gca().transAxes, size='small')
beautify()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
minbnd, maxbnd = plt.xlim()
plt.plot((entry0.mMaxEvals(), entry0.mMaxEvals()),
# (minbnd, entry1.mMaxEvals()), ls='-', color=colors[i],
(max([minbnd, entry1.mMaxEvals()/max_evals_line_length]), entry1.mMaxEvals()), ls='-', color=colors[i],
zorder=-1)
plt.plot(# (minbnd, entry0.mMaxEvals()),
(max([minbnd, entry0.mMaxEvals()/max_evals_line_length]), entry0.mMaxEvals()),
(entry1.mMaxEvals(), entry1.mMaxEvals()), ls='-',
color=colors[i], zorder=-1)
plt.xlim(minbnd, maxbnd)
plt.ylim(minbnd, maxbnd)
#Set the boundaries again: they changed due to new plots.
#plt.axvline(entry0.mMaxEvals(), ls='--', color=colors[i])
#plt.axhline(entry1.mMaxEvals(), ls='--', color=colors[i])
fontSize = genericsettings.getFontSize(funInfos.values())
if f in funInfos.keys():
plt.ylabel(funInfos[f], fontsize=fontSize)
filename = os.path.join(outputdir, 'ppscatter_f%03d' % f)
saveFigure(filename, verbose=verbose)
plt.close()
def main(dsList0, dsList1, outputdir, verbose=True):
"""Generate a scatter plot figure."""
#plt.rc("axes", labelsize=24, titlesize=24)
#plt.rc("xtick", labelsize=20)
#plt.rc("ytick", labelsize=20)
#plt.rc("font", size=20)
#plt.rc("legend", fontsize=20)
dictFunc0 = dsList0.dictByFunc()
dictFunc1 = dsList1.dictByFunc()
funcs = set(dictFunc0.keys()) & set(dictFunc1.keys())
for f in funcs:
dictDim0 = dictFunc0[f].dictByDim()
dictDim1 = dictFunc1[f].dictByDim()
dims = set(dictDim0.keys()) & set(dictDim1.keys())
#set_trace()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
xdata = numpy.array(entry0.detERT(targets))
ydata = numpy.array(entry1.detERT(targets))
tmp = (numpy.isinf(xdata)==False) * (numpy.isinf(ydata)==False)
if tmp.any():
try:
plt.plot(xdata[tmp], ydata[tmp], ls='', markersize=markersize,
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3)
except KeyError:
plt.plot(xdata[tmp], ydata[tmp], ls='', markersize=markersize,
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3)
#try:
# plt.scatter(xdata[tmp], ydata[tmp], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3)
#except ValueError:
# set_trace()
#ax = plt.gca()
ax = plt.axes()
tmp = numpy.isinf(xdata) * (numpy.isinf(ydata)==False)
if tmp.any():
trans = blend(ax.transAxes, ax.transData)
#plt.scatter([1.]*numpy.sum(tmp), ydata[tmp], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1.]*numpy.sum(tmp), ydata[tmp], markersize=markersize, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
except KeyError:
plt.plot([1.]*numpy.sum(tmp), ydata[tmp], markersize=markersize, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
#set_trace()
tmp = (numpy.isinf(xdata)==False) * numpy.isinf(ydata)
if tmp.any():
trans = blend(ax.transData, ax.transAxes)
# plt.scatter(xdata[tmp], [1.-offset]*numpy.sum(tmp), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot(xdata[tmp], [1.-offset]*numpy.sum(tmp), markersize=markersize, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
except KeyError:
plt.plot(xdata[tmp], [1.-offset]*numpy.sum(tmp), markersize=markersize, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
tmp = numpy.isinf(xdata) * numpy.isinf(ydata)
if tmp.any():
# plt.scatter(xdata[tmp], [1.-offset]*numpy.sum(tmp), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1.-offset]*numpy.sum(tmp), [1.-offset]*numpy.sum(tmp), markersize=markersize, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=ax.transAxes, clip_on=False)
except KeyError:
plt.plot([1.-offset]*numpy.sum(tmp), [1.-offset]*numpy.sum(tmp), markersize=markersize, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=ax.transAxes, clip_on=False)
#set_trace()
beautify()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
minbnd, maxbnd = plt.xlim()
plt.plot((entry0.mMaxEvals(), entry0.mMaxEvals()),
# (minbnd, entry1.mMaxEvals()), ls='-', color=colors[i],
(max([minbnd, entry1.mMaxEvals()/10.]), entry1.mMaxEvals()), ls='-', color=colors[i],
zorder=-1)
plt.plot(# (minbnd, entry0.mMaxEvals()),
(max([minbnd, entry0.mMaxEvals()/10.]), entry0.mMaxEvals()),
(entry1.mMaxEvals(), entry1.mMaxEvals()), ls='-',
color=colors[i], zorder=-1)
plt.xlim(minbnd, maxbnd)
plt.ylim(minbnd, maxbnd)
#Set the boundaries again: they changed due to new plots.
#plt.axvline(entry0.mMaxEvals(), ls='--', color=colors[i])
#plt.axhline(entry1.mMaxEvals(), ls='--', color=colors[i])
if isBenchmarkinfosFound:
try:
plt.ylabel(funInfos[f])
except IndexError:
pass
filename = os.path.join(outputdir, 'ppscatter_f%03d' % f)
saveFigure(filename, verbose=verbose)
plt.close()
def main(dsList0, dsList1, outputdir, settings):
"""Generate a scatter plot figure.
TODO: """
markers = genericsettings.dim_related_markers
colors = genericsettings.dim_related_colors
#plt.rc("axes", labelsize=24, titlesize=24)
#plt.rc("xtick", labelsize=20)
#plt.rc("ytick", labelsize=20)
#plt.rc("font", size=20)
#plt.rc("legend", fontsize=20)
dictFunc0 = dsList0.dictByFunc()
dictFunc1 = dsList1.dictByFunc()
funcs = set(dictFunc0.keys()) & set(dictFunc1.keys())
targets = testbedsettings.current_testbed.ppscatter_target_values
if isinstance(targets, pproc.RunlengthBasedTargetValues):
linewidth = linewidth_rld_based
else:
linewidth = linewidth_default
funInfos = ppfigparam.read_fun_infos()
for f in funcs:
dictDim0 = dictFunc0[f].dictByDim()
dictDim1 = dictFunc1[f].dictByDim()
dims = set(dictDim0.keys()) & set(dictDim1.keys())
#set_trace()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
if linewidth: # plot all reliable aRT values as a line
all_targets = np.array(
sorted(set(entry0.target).union(entry1.target),
reverse=True))
assert entry0.detSuccessRates([all_targets[0]]) == 1.0
assert entry1.detSuccessRates([all_targets[0]]) == 1.0
all_targets = all_targets[np.where(
all_targets <= targets((f, d))[0])[0]] #
xdata_all = np.array(entry0.detERT(all_targets))
ydata_all = np.array(entry1.detERT(all_targets))
# idx of reliable targets: last index where success rate >= 1/2 and aRT <= maxevals
idx = []
for ari in (np.where(
entry0.detSuccessRates(all_targets) >= 0.5)[0],
np.where(
entry1.detSuccessRates(all_targets) >= 0.5)[0],
np.where(xdata_all <= max(entry0.maxevals))[0],
np.where(ydata_all <= max(entry1.maxevals))[0]):
if len(ari):
idx.append(ari[-1])
if len(idx) == 4:
max_idx = min(idx)
## at least up to the most difficult given target
## idx = max((idx, np.where(all_targets >= targets((f, d))[-1])[0][-1]))
xdata_all = xdata_all[:max_idx + 1]
ydata_all = ydata_all[:max_idx + 1]
idx = (numpy.isfinite(xdata_all)) * (
numpy.isfinite(ydata_all))
assert idx.all()
if idx.any():
plt.plot(
xdata_all[idx],
ydata_all[idx],
colors[i],
ls='solid',
lw=linewidth,
# TODO: ls has changed, check whether this works out
clip_on=False)
xdata = numpy.array(entry0.detERT(targets((f, d))))
ydata = numpy.array(entry1.detERT(targets((f, d))))
# plot "valid" data, those within maxevals
idx = np.logical_and(xdata < entry0.mMaxEvals(),
ydata < entry1.mMaxEvals())
# was:
# (numpy.isinf(xdata) == False) *
# (numpy.isinf(ydata) == False) *
# (xdata < entry0.mMaxEvals()) *
# (ydata < entry1.mMaxEvals()))
if idx.any():
try:
plt.plot(xdata[idx],
ydata[idx],
ls='',
markersize=markersize,
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=3,
clip_on=False)
except KeyError:
plt.plot(xdata[idx],
ydata[idx],
ls='',
markersize=markersize,
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=3,
clip_on=False)
#try:
# plt.scatter(xdata[idx], ydata[idx], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3)
#except ValueError:
# set_trace()
# plot beyond maxevals but finite data
idx = ((numpy.isinf(xdata) == False) *
(numpy.isinf(ydata) == False) *
np.logical_or(xdata >= entry0.mMaxEvals(),
ydata >= entry1.mMaxEvals()))
if idx.any():
try:
plt.plot(xdata[idx],
ydata[idx],
ls='',
markersize=markersize +
markersize_addon_beyond_maxevals,
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=1,
clip_on=False)
except KeyError:
plt.plot(xdata[idx],
ydata[idx],
ls='',
markersize=markersize,
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=2,
clip_on=False)
#ax = plt.gca()
ax = plt.axes()
# plot data on the right edge
idx = numpy.isinf(xdata) * (numpy.isinf(ydata) == False)
if idx.any():
# This (seems to) transform inf to the figure limits!?
trans = blend(ax.transAxes, ax.transData)
#plt.scatter([1.]*numpy.sum(idx), ydata[idx], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1.] * numpy.sum(idx),
ydata[idx],
markersize=markersize +
markersize_addon_beyond_maxevals,
ls='',
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=1,
transform=trans,
clip_on=False)
except KeyError:
plt.plot([1.] * numpy.sum(idx),
ydata[idx],
markersize=markersize,
ls='',
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=2,
transform=trans,
clip_on=False)
#set_trace()
# plot data on the left edge
idx = (numpy.isinf(xdata) == False) * numpy.isinf(ydata)
if idx.any():
# This (seems to) transform inf to the figure limits!?
trans = blend(ax.transData, ax.transAxes)
# plt.scatter(xdata[idx], [1.-offset]*numpy.sum(idx), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot(xdata[idx], [1. - offset] * numpy.sum(idx),
markersize=markersize +
markersize_addon_beyond_maxevals,
ls='',
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=1,
transform=trans,
clip_on=False)
except KeyError:
plt.plot(xdata[idx], [1. - offset] * numpy.sum(idx),
markersize=markersize,
ls='',
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=2,
transform=trans,
clip_on=False)
# plot data in the top corner
idx = numpy.isinf(xdata) * numpy.isinf(ydata)
if idx.any():
# plt.scatter(xdata[idx], [1.-offset]*numpy.sum(idx), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1. - offset] * numpy.sum(idx),
[1. - offset] * numpy.sum(idx),
markersize=markersize +
markersize_addon_beyond_maxevals,
ls='',
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=1,
transform=ax.transAxes,
clip_on=False)
except KeyError:
plt.plot([1. - offset] * numpy.sum(idx),
[1. - offset] * numpy.sum(idx),
markersize=markersize,
ls='',
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=2,
transform=ax.transAxes,
clip_on=False)
targetlabels = targets.labels()
if isinstance(targets, pproc.RunlengthBasedTargetValues):
text = (str(len(targetlabels)) + ' target RLs/dim: ' +
targetlabels[0] + '..' +
targetlabels[len(targetlabels) - 1] + '\n')
text += ' from ' + testbedsettings.current_testbed.reference_algorithm_filename
else:
text = (str(len(targetlabels)) + ' targets in ' + targetlabels[0] +
'..' + targetlabels[len(targetlabels) - 1])
# add number of instances
text += '\n'
num_of_instances_alg0 = []
num_of_instances_alg1 = []
for d in dims:
num_of_instances_alg0.append((dictDim0[d][0]).nbRuns())
num_of_instances_alg1.append((dictDim1[d][0]).nbRuns())
# issue a warning if the numbers of instances are inconsistent:
if (len(set(num_of_instances_alg0)) > 1):
warnings.warn(
'Inconsistent numbers of instances over dimensions found for ALG0:\n\
found instances %s' % str(num_of_instances_alg0))
if (len(set(num_of_instances_alg1)) > 1):
warnings.warn(
'Inconsistent numbers of instances over dimensions found for ALG1:\n\
found instances %s' % str(num_of_instances_alg1))
if (len(set(num_of_instances_alg0)) == 1
and len(set(num_of_instances_alg0)) == 1):
text += '%s and %s instances' % (num_of_instances_alg0[0],
num_of_instances_alg1[0])
else:
for n in num_of_instances_alg0:
text += '%d, ' % n
text = text.rstrip(', ')
text += ' and '
for n in num_of_instances_alg1:
text += '%d, ' % n
text = text.rstrip(', ')
text += ' instances'
plt.text(0.01,
0.98,
text,
horizontalalignment="left",
verticalalignment="top",
transform=plt.gca().transAxes,
size='small')
beautify()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
minbnd, maxbnd = plt.xlim()
plt.plot(
(entry0.mMaxEvals(), entry0.mMaxEvals()),
# (minbnd, entry1.mMaxEvals()), ls='-', color=colors[i],
(max([minbnd,
entry1.mMaxEvals() / max_evals_line_length
]), entry1.mMaxEvals()),
ls='-',
color=colors[i],
zorder=-1)
plt.plot( # (minbnd, entry0.mMaxEvals()),
(max([minbnd,
entry0.mMaxEvals() / max_evals_line_length
]), entry0.mMaxEvals()),
(entry1.mMaxEvals(), entry1.mMaxEvals()),
ls='-',
color=colors[i],
zorder=-1)
plt.xlim(minbnd, maxbnd)
plt.ylim(minbnd, maxbnd)
#Set the boundaries again: they changed due to new plots.
#plt.axvline(entry0.mMaxEvals(), ls='--', color=colors[i])
#plt.axhline(entry1.mMaxEvals(), ls='--', color=colors[i])
fontSize = genericsettings.getFontSize(funInfos.values())
if f in funInfos.keys():
plt.ylabel(funInfos[f], fontsize=fontSize)
filename = os.path.join(outputdir, 'ppscatter_f%03d' % f)
save_figure(filename, dsList0[0].algId)
plt.close()
def main(dsList0, dsList1, outputdir, verbose=True):
"""Generate a scatter plot figure.
TODO: """
#plt.rc("axes", labelsize=24, titlesize=24)
#plt.rc("xtick", labelsize=20)
#plt.rc("ytick", labelsize=20)
#plt.rc("font", size=20)
#plt.rc("legend", fontsize=20)
dictFunc0 = dsList0.dictByFunc()
dictFunc1 = dsList1.dictByFunc()
funcs = set(dictFunc0.keys()) & set(dictFunc1.keys())
for f in funcs:
dictDim0 = dictFunc0[f].dictByDim()
dictDim1 = dictFunc1[f].dictByDim()
dims = set(dictDim0.keys()) & set(dictDim1.keys())
#set_trace()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
if linewidth: # plot all reliable ERT values as a line
all_targets = np.array(sorted(set(entry0.target).union(entry1.target), reverse=True))
assert entry0.detSuccessRates([all_targets[0]]) == 1.0
assert entry1.detSuccessRates([all_targets[0]]) == 1.0
all_targets = all_targets[np.where(all_targets <= targets((f, d))[0])[0]] #
xdata_all = np.array(entry0.detERT(all_targets))
ydata_all = np.array(entry1.detERT(all_targets))
# idx of reliable targets: last index where success rate >= 1/2 and ERT <= maxevals
idx = []
for ari in (np.where(entry0.detSuccessRates(all_targets) >= 0.5)[0],
np.where(entry1.detSuccessRates(all_targets) >= 0.5)[0],
np.where(xdata_all <= max(entry0.maxevals))[0],
np.where(ydata_all <= max(entry1.maxevals))[0]
):
if len(ari):
idx.append(ari[-1])
if len(idx) == 4:
max_idx = min(idx)
## at least up to the most difficult given target
## idx = max((idx, np.where(all_targets >= targets((f, d))[-1])[0][-1]))
xdata_all = xdata_all[:max_idx + 1]
ydata_all = ydata_all[:max_idx + 1]
idx = (numpy.isfinite(xdata_all)) * (numpy.isfinite(ydata_all))
assert idx.all()
if idx.any():
plt.plot(xdata_all[idx], ydata_all[idx], colors[i], ls='solid', lw=linewidth,
# TODO: ls has changed, check whether this works out
clip_on=False)
xdata = numpy.array(entry0.detERT(targets((f, d))))
ydata = numpy.array(entry1.detERT(targets((f, d))))
tmp = (numpy.isinf(xdata)==False) * (numpy.isinf(ydata)==False)
if tmp.any():
try:
plt.plot(xdata[tmp], ydata[tmp], ls='',
markersize=markersize,
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
clip_on=False)
except KeyError:
plt.plot(xdata[tmp], ydata[tmp], ls='', markersize=markersize,
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
clip_on=False)
#try:
# plt.scatter(xdata[tmp], ydata[tmp], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3)
#except ValueError:
# set_trace()
#ax = plt.gca()
ax = plt.axes()
tmp = numpy.isinf(xdata) * (numpy.isinf(ydata)==False)
if tmp.any():
trans = blend(ax.transAxes, ax.transData)
#plt.scatter([1.]*numpy.sum(tmp), ydata[tmp], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1.]*numpy.sum(tmp), ydata[tmp], markersize=markersize, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
except KeyError:
plt.plot([1.]*numpy.sum(tmp), ydata[tmp], markersize=markersize, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
#set_trace()
tmp = (numpy.isinf(xdata)==False) * numpy.isinf(ydata)
if tmp.any():
trans = blend(ax.transData, ax.transAxes)
# plt.scatter(xdata[tmp], [1.-offset]*numpy.sum(tmp), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot(xdata[tmp], [1.-offset]*numpy.sum(tmp), markersize=markersize, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
except KeyError:
plt.plot(xdata[tmp], [1.-offset]*numpy.sum(tmp), markersize=markersize, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
tmp = numpy.isinf(xdata) * numpy.isinf(ydata)
if tmp.any():
# plt.scatter(xdata[tmp], [1.-offset]*numpy.sum(tmp), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1.-offset]*numpy.sum(tmp), [1.-offset]*numpy.sum(tmp), markersize=markersize, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=ax.transAxes, clip_on=False)
except KeyError:
plt.plot([1.-offset]*numpy.sum(tmp), [1.-offset]*numpy.sum(tmp), markersize=markersize, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=ax.transAxes, clip_on=False)
#set_trace()
beautify()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
minbnd, maxbnd = plt.xlim()
plt.plot((entry0.mMaxEvals(), entry0.mMaxEvals()),
# (minbnd, entry1.mMaxEvals()), ls='-', color=colors[i],
(max([minbnd, entry1.mMaxEvals()/max_evals_line_length]), entry1.mMaxEvals()), ls='-', color=colors[i],
zorder=-1)
plt.plot(# (minbnd, entry0.mMaxEvals()),
(max([minbnd, entry0.mMaxEvals()/max_evals_line_length]), entry0.mMaxEvals()),
(entry1.mMaxEvals(), entry1.mMaxEvals()), ls='-',
color=colors[i], zorder=-1)
plt.xlim(minbnd, maxbnd)
plt.ylim(minbnd, maxbnd)
#Set the boundaries again: they changed due to new plots.
#plt.axvline(entry0.mMaxEvals(), ls='--', color=colors[i])
#plt.axhline(entry1.mMaxEvals(), ls='--', color=colors[i])
try:
plt.ylabel(funInfos[f])
except IndexError:
pass
filename = os.path.join(outputdir, 'ppscatter_f%03d' % f)
saveFigure(filename, verbose=verbose)
plt.close()
def main(dsList0, dsList1, outputdir, verbose=True):
"""Generate a scatter plot figure.
TODO: """
#plt.rc("axes", labelsize=24, titlesize=24)
#plt.rc("xtick", labelsize=20)
#plt.rc("ytick", labelsize=20)
#plt.rc("font", size=20)
#plt.rc("legend", fontsize=20)
dictFunc0 = dsList0.dictByFunc()
dictFunc1 = dsList1.dictByFunc()
funcs = set(dictFunc0.keys()) & set(dictFunc1.keys())
if isinstance(targets, pproc.RunlengthBasedTargetValues):
linewidth = linewidth_rld_based
else:
linewidth = linewidth_default
for f in funcs:
dictDim0 = dictFunc0[f].dictByDim()
dictDim1 = dictFunc1[f].dictByDim()
dims = set(dictDim0.keys()) & set(dictDim1.keys())
#set_trace()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
if linewidth: # plot all reliable ERT values as a line
all_targets = np.array(
sorted(set(entry0.target).union(entry1.target),
reverse=True))
assert entry0.detSuccessRates([all_targets[0]]) == 1.0
assert entry1.detSuccessRates([all_targets[0]]) == 1.0
all_targets = all_targets[np.where(
all_targets <= targets((f, d))[0])[0]] #
xdata_all = np.array(entry0.detERT(all_targets))
ydata_all = np.array(entry1.detERT(all_targets))
# idx of reliable targets: last index where success rate >= 1/2 and ERT <= maxevals
idx = []
for ari in (np.where(
entry0.detSuccessRates(all_targets) >= 0.5)[0],
np.where(
entry1.detSuccessRates(all_targets) >= 0.5)[0],
np.where(xdata_all <= max(entry0.maxevals))[0],
np.where(ydata_all <= max(entry1.maxevals))[0]):
if len(ari):
idx.append(ari[-1])
if len(idx) == 4:
max_idx = min(idx)
## at least up to the most difficult given target
## idx = max((idx, np.where(all_targets >= targets((f, d))[-1])[0][-1]))
xdata_all = xdata_all[:max_idx + 1]
ydata_all = ydata_all[:max_idx + 1]
idx = (numpy.isfinite(xdata_all)) * (
numpy.isfinite(ydata_all))
assert idx.all()
if idx.any():
plt.plot(
xdata_all[idx],
ydata_all[idx],
colors[i],
ls='solid',
lw=linewidth,
# TODO: ls has changed, check whether this works out
clip_on=False)
xdata = numpy.array(entry0.detERT(targets((f, d))))
ydata = numpy.array(entry1.detERT(targets((f, d))))
# plot "valid" data, those within maxevals
idx = np.logical_and(xdata < entry0.mMaxEvals(),
ydata < entry1.mMaxEvals())
# was:
# (numpy.isinf(xdata) == False) *
# (numpy.isinf(ydata) == False) *
# (xdata < entry0.mMaxEvals()) *
# (ydata < entry1.mMaxEvals()))
if idx.any():
try:
plt.plot(xdata[idx],
ydata[idx],
ls='',
markersize=markersize,
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=3,
clip_on=False)
except KeyError:
plt.plot(xdata[idx],
ydata[idx],
ls='',
markersize=markersize,
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=3,
clip_on=False)
#try:
# plt.scatter(xdata[idx], ydata[idx], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3)
#except ValueError:
# set_trace()
# plot beyond maxevals but finite data
idx = ((numpy.isinf(xdata) == False) *
(numpy.isinf(ydata) == False) *
np.logical_or(xdata >= entry0.mMaxEvals(),
ydata >= entry1.mMaxEvals()))
if idx.any():
try:
plt.plot(xdata[idx],
ydata[idx],
ls='',
markersize=markersize +
markersize_addon_beyond_maxevals,
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=1,
clip_on=False)
except KeyError:
plt.plot(xdata[idx],
ydata[idx],
ls='',
markersize=markersize,
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=2,
clip_on=False)
#ax = plt.gca()
ax = plt.axes()
# plot data on the right edge
idx = numpy.isinf(xdata) * (numpy.isinf(ydata) == False)
if idx.any():
# This (seems to) transform inf to the figure limits!?
trans = blend(ax.transAxes, ax.transData)
#plt.scatter([1.]*numpy.sum(idx), ydata[idx], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1.] * numpy.sum(idx),
ydata[idx],
markersize=markersize +
markersize_addon_beyond_maxevals,
ls='',
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=1,
transform=trans,
clip_on=False)
except KeyError:
plt.plot([1.] * numpy.sum(idx),
ydata[idx],
markersize=markersize,
ls='',
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=2,
transform=trans,
clip_on=False)
#set_trace()
# plot data on the left edge
idx = (numpy.isinf(xdata) == False) * numpy.isinf(ydata)
if idx.any():
# This (seems to) transform inf to the figure limits!?
trans = blend(ax.transData, ax.transAxes)
# plt.scatter(xdata[idx], [1.-offset]*numpy.sum(idx), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot(xdata[idx], [1. - offset] * numpy.sum(idx),
markersize=markersize +
markersize_addon_beyond_maxevals,
ls='',
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=1,
transform=trans,
clip_on=False)
except KeyError:
plt.plot(xdata[idx], [1. - offset] * numpy.sum(idx),
markersize=markersize,
ls='',
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=2,
transform=trans,
clip_on=False)
# plot data in the top corner
idx = numpy.isinf(xdata) * numpy.isinf(ydata)
if idx.any():
# plt.scatter(xdata[idx], [1.-offset]*numpy.sum(idx), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1. - offset] * numpy.sum(idx),
[1. - offset] * numpy.sum(idx),
markersize=markersize +
markersize_addon_beyond_maxevals,
ls='',
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=1,
transform=ax.transAxes,
clip_on=False)
except KeyError:
plt.plot([1. - offset] * numpy.sum(idx),
[1. - offset] * numpy.sum(idx),
markersize=markersize,
ls='',
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=2,
transform=ax.transAxes,
clip_on=False)
#set_trace()
beautify()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
minbnd, maxbnd = plt.xlim()
plt.plot(
(entry0.mMaxEvals(), entry0.mMaxEvals()),
# (minbnd, entry1.mMaxEvals()), ls='-', color=colors[i],
(max([minbnd,
entry1.mMaxEvals() / max_evals_line_length
]), entry1.mMaxEvals()),
ls='-',
color=colors[i],
zorder=-1)
plt.plot( # (minbnd, entry0.mMaxEvals()),
(max([minbnd,
entry0.mMaxEvals() / max_evals_line_length
]), entry0.mMaxEvals()),
(entry1.mMaxEvals(), entry1.mMaxEvals()),
ls='-',
color=colors[i],
zorder=-1)
plt.xlim(minbnd, maxbnd)
plt.ylim(minbnd, maxbnd)
#Set the boundaries again: they changed due to new plots.
#plt.axvline(entry0.mMaxEvals(), ls='--', color=colors[i])
#plt.axhline(entry1.mMaxEvals(), ls='--', color=colors[i])
try:
plt.ylabel(funInfos[f])
except IndexError:
pass
filename = os.path.join(outputdir, 'ppscatter_f%03d' % f)
saveFigure(filename, verbose=verbose)
if f == 1:
algName1 = toolsdivers.str_to_latex(
toolsdivers.strip_pathname1(entry1.algId))
algName0 = toolsdivers.str_to_latex(
toolsdivers.strip_pathname1(entry0.algId))
save_single_functions_html(os.path.join(
outputdir, genericsettings.two_algorithm_file_name),
"%s vs %s" % (algName1, algName0),
algorithmCount=AlgorithmCount.TWO)
plt.close()
def main(dsList0, dsList1, outputdir, verbose=True):
plt.rc("axes", labelsize=20, titlesize=24)
plt.rc("xtick", labelsize=20)
plt.rc("ytick", labelsize=20)
plt.rc("font", size=20)
plt.rc("legend", fontsize=20)
dictFunc0 = dsList0.dictByFunc()
dictFunc1 = dsList1.dictByFunc()
funcs = set(dictFunc0.keys()) & set(dictFunc1.keys())
nbmarkers = 46
inc = 45./(nbmarkers-1)
targets = numpy.power(10, numpy.arange(-40, 5 + inc, inc)/5.)
for f in funcs:
dictDim0 = dictFunc0[f].dictByDim()
dictDim1 = dictFunc1[f].dictByDim()
dims = set(dictDim0.keys()) & set(dictDim1.keys())
#set_trace()
for i, d in enumerate((2, 3, 5, 10, 20, 40)):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
xdata = numpy.array(entry0.detERT(targets))
ydata = numpy.array(entry1.detERT(targets))
#targets, xdata, ydata = generateData(entry0, entry1)
#plt.plot(xdata, ydata, ls='', color=colors[i], marker=markers[i],
# markerfacecolor='None', markeredgecolor=colors[i],
# markersize=10, markeredgewidth=3)
tmp = (numpy.isinf(xdata)==False) * (numpy.isinf(ydata)==False)
if tmp.any():
try:
plt.plot(xdata[tmp], ydata[tmp], ls='', markersize=10,
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3)
except KeyError:
plt.plot(xdata[tmp], ydata[tmp], ls='', markersize=10,
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3)
#try:
# plt.scatter(xdata[tmp], ydata[tmp], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3)
#except ValueError:
# set_trace()
#ax = plt.gca()
ax = plt.axes()
tmp = numpy.isinf(xdata) * (numpy.isinf(ydata)==False)
if tmp.any():
trans = blend(ax.transAxes, ax.transData)
#plt.scatter([1.]*numpy.sum(tmp), ydata[tmp], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1.]*numpy.sum(tmp), ydata[tmp], markersize=10, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
except KeyError:
plt.plot([1.]*numpy.sum(tmp), ydata[tmp], markersize=10, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
#set_trace()
tmp = (numpy.isinf(xdata)==False) * numpy.isinf(ydata)
if tmp.any():
trans = blend(ax.transData, ax.transAxes)
# plt.scatter(xdata[tmp], [1.-offset]*numpy.sum(tmp), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot(xdata[tmp], [1.-offset]*numpy.sum(tmp), markersize=10, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
except KeyError:
plt.plot(xdata[tmp], [1.-offset]*numpy.sum(tmp), markersize=10, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
tmp = numpy.isinf(xdata) * numpy.isinf(ydata)
if tmp.any():
# plt.scatter(xdata[tmp], [1.-offset]*numpy.sum(tmp), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1.-offset]*numpy.sum(tmp), [1.-offset]*numpy.sum(tmp), markersize=10, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=ax.transAxes, clip_on=False)
except KeyError:
plt.plot([1.-offset]*numpy.sum(tmp), [1.-offset]*numpy.sum(tmp), markersize=10, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=ax.transAxes, clip_on=False)
#set_trace()
beautify()
#if isBenchmarkinfosFound:
# plt.ylabel(funInfos[f])
filename = os.path.join(outputdir, 'scatter_f%d' % f)
saveFigure(filename, figFormat=figFormat, verbose=verbose)
plt.close()
plt.rcdefaults()
def main(dsList0, dsList1, outputdir, verbose=True):
"""Generate a scatter plot figure.
TODO: """
#plt.rc("axes", labelsize=24, titlesize=24)
#plt.rc("xtick", labelsize=20)
#plt.rc("ytick", labelsize=20)
#plt.rc("font", size=20)
#plt.rc("legend", fontsize=20)
dictFunc0 = dsList0.dictByFunc()
dictFunc1 = dsList1.dictByFunc()
funcs = set(dictFunc0.keys()) & set(dictFunc1.keys())
for f in funcs:
dictDim0 = dictFunc0[f].dictByDim()
dictDim1 = dictFunc1[f].dictByDim()
dims = set(dictDim0.keys()) & set(dictDim1.keys())
#set_trace()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
if linewidth: # plot all reliable ERT values as a line
all_targets = np.array(sorted(set(entry0.target).union(entry1.target), reverse=True))
assert entry0.detSuccessRates([all_targets[0]]) == 1.0
assert entry1.detSuccessRates([all_targets[0]]) == 1.0
all_targets = all_targets[np.where(all_targets <= targets((f, d))[0])[0]] #
xdata_all = np.array(entry0.detERT(all_targets))
ydata_all = np.array(entry1.detERT(all_targets))
# idx of reliable targets: last index where success rate >= 1/2 and ERT <= maxevals
idx = []
for ari in (np.where(entry0.detSuccessRates(all_targets) >= 0.5)[0],
np.where(entry1.detSuccessRates(all_targets) >= 0.5)[0],
np.where(xdata_all <= max(entry0.maxevals))[0],
np.where(ydata_all <= max(entry1.maxevals))[0]
):
if len(ari):
idx.append(ari[-1])
if len(idx) == 4:
max_idx = min(idx)
## at least up to the most difficult given target
## idx = max((idx, np.where(all_targets >= targets((f, d))[-1])[0][-1]))
xdata_all = xdata_all[:max_idx + 1]
ydata_all = ydata_all[:max_idx + 1]
idx = (numpy.isfinite(xdata_all)) * (numpy.isfinite(ydata_all))
assert idx.all()
if idx.any():
plt.plot(xdata_all[idx], ydata_all[idx], colors[i], ls='solid', lw=linewidth,
# TODO: ls has changed, check whether this works out
clip_on=False)
xdata = numpy.array(entry0.detERT(targets((f, d))))
ydata = numpy.array(entry1.detERT(targets((f, d))))
tmp = (numpy.isinf(xdata)==False) * (numpy.isinf(ydata)==False)
if tmp.any():
try:
plt.plot(xdata[tmp], ydata[tmp], ls='',
markersize=markersize,
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
clip_on=False)
except KeyError:
plt.plot(xdata[tmp], ydata[tmp], ls='', markersize=markersize,
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
clip_on=False)
#try:
# plt.scatter(xdata[tmp], ydata[tmp], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3)
#except ValueError:
# set_trace()
#ax = plt.gca()
ax = plt.axes()
tmp = numpy.isinf(xdata) * (numpy.isinf(ydata)==False)
if tmp.any():
trans = blend(ax.transAxes, ax.transData)
#plt.scatter([1.]*numpy.sum(tmp), ydata[tmp], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1.]*numpy.sum(tmp), ydata[tmp], markersize=markersize, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
except KeyError:
plt.plot([1.]*numpy.sum(tmp), ydata[tmp], markersize=markersize, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
#set_trace()
tmp = (numpy.isinf(xdata)==False) * numpy.isinf(ydata)
if tmp.any():
trans = blend(ax.transData, ax.transAxes)
# plt.scatter(xdata[tmp], [1.-offset]*numpy.sum(tmp), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot(xdata[tmp], [1.-offset]*numpy.sum(tmp), markersize=markersize, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
except KeyError:
plt.plot(xdata[tmp], [1.-offset]*numpy.sum(tmp), markersize=markersize, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=trans, clip_on=False)
tmp = numpy.isinf(xdata) * numpy.isinf(ydata)
if tmp.any():
# plt.scatter(xdata[tmp], [1.-offset]*numpy.sum(tmp), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1.-offset]*numpy.sum(tmp), [1.-offset]*numpy.sum(tmp), markersize=markersize, ls='',
marker=markers[i], markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=ax.transAxes, clip_on=False)
except KeyError:
plt.plot([1.-offset]*numpy.sum(tmp), [1.-offset]*numpy.sum(tmp), markersize=markersize, ls='',
marker='x', markerfacecolor='None',
markeredgecolor=colors[i], markeredgewidth=3,
transform=ax.transAxes, clip_on=False)
#set_trace()
beautify()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
minbnd, maxbnd = plt.xlim()
plt.plot((entry0.mMaxEvals(), entry0.mMaxEvals()),
# (minbnd, entry1.mMaxEvals()), ls='-', color=colors[i],
(max([minbnd, entry1.mMaxEvals()/max_evals_line_length]), entry1.mMaxEvals()), ls='-', color=colors[i],
zorder=-1)
plt.plot(# (minbnd, entry0.mMaxEvals()),
(max([minbnd, entry0.mMaxEvals()/max_evals_line_length]), entry0.mMaxEvals()),
(entry1.mMaxEvals(), entry1.mMaxEvals()), ls='-',
color=colors[i], zorder=-1)
plt.xlim(minbnd, maxbnd)
plt.ylim(minbnd, maxbnd)
#Set the boundaries again: they changed due to new plots.
#plt.axvline(entry0.mMaxEvals(), ls='--', color=colors[i])
#plt.axhline(entry1.mMaxEvals(), ls='--', color=colors[i])
try:
plt.ylabel(funInfos[f])
except IndexError:
pass
filename = os.path.join(outputdir, 'ppscatter_f%03d' % f)
saveFigure(filename, verbose=verbose)
plt.close()
def main(dsList0, dsList1, outputdir, verbose=True):
"""Generate a scatter plot figure.
TODO: """
# plt.rc("axes", labelsize=24, titlesize=24)
# plt.rc("xtick", labelsize=20)
# plt.rc("ytick", labelsize=20)
# plt.rc("font", size=20)
# plt.rc("legend", fontsize=20)
dictFunc0 = dsList0.dictByFunc()
dictFunc1 = dsList1.dictByFunc()
funcs = set(dictFunc0.keys()) & set(dictFunc1.keys())
if isinstance(targets, pproc.RunlengthBasedTargetValues):
linewidth = linewidth_rld_based
else:
linewidth = linewidth_default
for f in funcs:
dictDim0 = dictFunc0[f].dictByDim()
dictDim1 = dictFunc1[f].dictByDim()
dims = set(dictDim0.keys()) & set(dictDim1.keys())
# set_trace()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
if linewidth: # plot all reliable ERT values as a line
all_targets = np.array(sorted(set(entry0.target).union(entry1.target), reverse=True))
assert entry0.detSuccessRates([all_targets[0]]) == 1.0
assert entry1.detSuccessRates([all_targets[0]]) == 1.0
all_targets = all_targets[np.where(all_targets <= targets((f, d))[0])[0]] #
xdata_all = np.array(entry0.detERT(all_targets))
ydata_all = np.array(entry1.detERT(all_targets))
# idx of reliable targets: last index where success rate >= 1/2 and ERT <= maxevals
idx = []
for ari in (
np.where(entry0.detSuccessRates(all_targets) >= 0.5)[0],
np.where(entry1.detSuccessRates(all_targets) >= 0.5)[0],
np.where(xdata_all <= max(entry0.maxevals))[0],
np.where(ydata_all <= max(entry1.maxevals))[0],
):
if len(ari):
idx.append(ari[-1])
if len(idx) == 4:
max_idx = min(idx)
## at least up to the most difficult given target
## idx = max((idx, np.where(all_targets >= targets((f, d))[-1])[0][-1]))
xdata_all = xdata_all[: max_idx + 1]
ydata_all = ydata_all[: max_idx + 1]
idx = (numpy.isfinite(xdata_all)) * (numpy.isfinite(ydata_all))
assert idx.all()
if idx.any():
plt.plot(
xdata_all[idx],
ydata_all[idx],
colors[i],
ls="solid",
lw=linewidth,
# TODO: ls has changed, check whether this works out
clip_on=False,
)
xdata = numpy.array(entry0.detERT(targets((f, d))))
ydata = numpy.array(entry1.detERT(targets((f, d))))
# plot "valid" data, those within maxevals
idx = np.logical_and(xdata < entry0.mMaxEvals(), ydata < entry1.mMaxEvals())
# was:
# (numpy.isinf(xdata) == False) *
# (numpy.isinf(ydata) == False) *
# (xdata < entry0.mMaxEvals()) *
# (ydata < entry1.mMaxEvals()))
if idx.any():
try:
plt.plot(
xdata[idx],
ydata[idx],
ls="",
markersize=markersize,
marker=markers[i],
markerfacecolor="None",
markeredgecolor=colors[i],
markeredgewidth=3,
clip_on=False,
)
except KeyError:
plt.plot(
xdata[idx],
ydata[idx],
ls="",
markersize=markersize,
marker="x",
markerfacecolor="None",
markeredgecolor=colors[i],
markeredgewidth=3,
clip_on=False,
)
# try:
# plt.scatter(xdata[idx], ydata[idx], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3)
# except ValueError:
# set_trace()
# plot beyond maxevals but finite data
idx = (
(numpy.isinf(xdata) == False)
* (numpy.isinf(ydata) == False)
* np.logical_or(xdata >= entry0.mMaxEvals(), ydata >= entry1.mMaxEvals())
)
if idx.any():
try:
plt.plot(
xdata[idx],
ydata[idx],
ls="",
markersize=markersize + markersize_addon_beyond_maxevals,
marker=markers[i],
markerfacecolor="None",
markeredgecolor=colors[i],
markeredgewidth=1,
clip_on=False,
)
except KeyError:
plt.plot(
xdata[idx],
ydata[idx],
ls="",
markersize=markersize,
marker="x",
markerfacecolor="None",
markeredgecolor=colors[i],
markeredgewidth=2,
clip_on=False,
)
# ax = plt.gca()
ax = plt.axes()
# plot data on the right edge
idx = numpy.isinf(xdata) * (numpy.isinf(ydata) == False)
if idx.any():
# This (seems to) transform inf to the figure limits!?
trans = blend(ax.transAxes, ax.transData)
# plt.scatter([1.]*numpy.sum(idx), ydata[idx], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot(
[1.0] * numpy.sum(idx),
ydata[idx],
markersize=markersize + markersize_addon_beyond_maxevals,
ls="",
marker=markers[i],
markerfacecolor="None",
markeredgecolor=colors[i],
markeredgewidth=1,
transform=trans,
clip_on=False,
)
except KeyError:
plt.plot(
[1.0] * numpy.sum(idx),
ydata[idx],
markersize=markersize,
ls="",
marker="x",
markerfacecolor="None",
markeredgecolor=colors[i],
markeredgewidth=2,
transform=trans,
clip_on=False,
)
# set_trace()
# plot data on the left edge
idx = (numpy.isinf(xdata) == False) * numpy.isinf(ydata)
if idx.any():
# This (seems to) transform inf to the figure limits!?
trans = blend(ax.transData, ax.transAxes)
# plt.scatter(xdata[idx], [1.-offset]*numpy.sum(idx), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot(
xdata[idx],
[1.0 - offset] * numpy.sum(idx),
markersize=markersize + markersize_addon_beyond_maxevals,
ls="",
marker=markers[i],
markerfacecolor="None",
markeredgecolor=colors[i],
markeredgewidth=1,
transform=trans,
clip_on=False,
)
except KeyError:
plt.plot(
xdata[idx],
[1.0 - offset] * numpy.sum(idx),
markersize=markersize,
ls="",
marker="x",
markerfacecolor="None",
markeredgecolor=colors[i],
markeredgewidth=2,
transform=trans,
clip_on=False,
)
# plot data in the top corner
idx = numpy.isinf(xdata) * numpy.isinf(ydata)
if idx.any():
# plt.scatter(xdata[idx], [1.-offset]*numpy.sum(idx), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot(
[1.0 - offset] * numpy.sum(idx),
[1.0 - offset] * numpy.sum(idx),
markersize=markersize + markersize_addon_beyond_maxevals,
ls="",
marker=markers[i],
markerfacecolor="None",
markeredgecolor=colors[i],
markeredgewidth=1,
transform=ax.transAxes,
clip_on=False,
)
except KeyError:
plt.plot(
[1.0 - offset] * numpy.sum(idx),
[1.0 - offset] * numpy.sum(idx),
markersize=markersize,
ls="",
marker="x",
markerfacecolor="None",
markeredgecolor=colors[i],
markeredgewidth=2,
transform=ax.transAxes,
clip_on=False,
)
# set_trace()
beautify()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
minbnd, maxbnd = plt.xlim()
plt.plot(
(entry0.mMaxEvals(), entry0.mMaxEvals()),
# (minbnd, entry1.mMaxEvals()), ls='-', color=colors[i],
(max([minbnd, entry1.mMaxEvals() / max_evals_line_length]), entry1.mMaxEvals()),
ls="-",
color=colors[i],
zorder=-1,
)
plt.plot( # (minbnd, entry0.mMaxEvals()),
(max([minbnd, entry0.mMaxEvals() / max_evals_line_length]), entry0.mMaxEvals()),
(entry1.mMaxEvals(), entry1.mMaxEvals()),
ls="-",
color=colors[i],
zorder=-1,
)
plt.xlim(minbnd, maxbnd)
plt.ylim(minbnd, maxbnd)
# Set the boundaries again: they changed due to new plots.
# plt.axvline(entry0.mMaxEvals(), ls='--', color=colors[i])
# plt.axhline(entry1.mMaxEvals(), ls='--', color=colors[i])
try:
plt.ylabel(funInfos[f])
except IndexError:
pass
filename = os.path.join(outputdir, "ppscatter_f%03d" % f)
saveFigure(filename, verbose=verbose)
if f == 1:
algName1 = toolsdivers.str_to_latex(toolsdivers.strip_pathname1(entry1.algId))
algName0 = toolsdivers.str_to_latex(toolsdivers.strip_pathname1(entry0.algId))
save_single_functions_html(
os.path.join(outputdir, genericsettings.two_algorithm_file_name),
"%s vs %s" % (algName1, algName0),
algorithmCount=AlgorithmCount.TWO,
)
plt.close()
def generate_line_plot(
# data parameters
data_dir=data_dir,
data_file=data_file,
read_csv_kw={},
strike_zone_data_file=None,
x_column=0,
y_column=-1,
x_minus_yvalues=None,
yerr_columns=None,
year=year,
first_day=None,
last_day=334,
hline_ypos=None,
# figure parameters
width=width,
height_to_width=height_to_width,
width_pixels_save=width_pixels_save,
axes_box=axes_box,
# text parameters
title_txt='',
ylab_txt='',
ylab_rotation=0,
hline_labels=None,
hline_label_units='',
presidential_2020=False,
hline_lab_xpos=0.85,
watermark=watermark,
watermark_pos=watermark_pos,
# size parameters
size_scale=size_scale,
font_size=font_size,
circle_size=3,
# thickness parameters
data_lw=1,
grid_lw=0.25,
spine_lw=0.25,
hline_lw=1,
tick_lw=0.25,
# color parameters
col_D=col_D,
col_R=col_R,
zone_colors=zone_colors,
shading=True,
color_reverse=False,
shading_kw=dict(zorder=0, alpha=0.15, lw=0),
yerr_kw=dict(color='grey', alpha=0.3, lw=0),
hline_color='dimgrey',
hline_lab_colors=None,
data_line_color='k',
yticklab_format=False,
grid_alpha=0.5,
strike_colors=['dimgrey', 'lightgrey'],
# axes parameters
ylim=None,
# tick parameters
tick_length=3,
tick_pad=1,
yticks_interval=1,
# distance/pad parameters
title_pad=3,
ylab_pad=0.1,
xtick_pad=0.015,
hline_lab_pad=0.06,
# output parameters
out_path=out_path,
out_format=out_format,
# other customs
plot_customization=False,
thumbnail=False,
):
# load data
data_path = os.path.join(data_dir, data_file)
data = pd.read_csv(data_path, **read_csv_kw)
#data.drop_duplicates(x_column, inplace=True)
# read in column labels
fname, ext = os.path.splitext(data_file)
column_file = fname + '_columns' + ext
column_path = os.path.join(data_dir, column_file)
columns = pd.read_csv(column_path).columns
data.columns = columns
data.sort_values(by=x_column, inplace=True)
# convert dates to datetimes
days = data[x_column].values
day_of_year = days
days = np.array([doy2dt(int(d)) for d in days])
# specify data to plot
vals = data[y_column].values
if x_minus_yvalues is not None:
vals = x_minus_yvalues - vals
# use data to fill title text
title_fillers = dict(
last_value=vals[-1],
party='',
)
# change hline_labels for thumnail versions, set fontsizes
# for hline labels
hlabel_font_size = font_size
if thumbnail and hline_label_units == "given":
font_size = font_size * 1.25
hlabel_font_size = font_size * 1.5
elif thumbnail and not plot_customization and hline_labels is not None:
font_size = font_size * 1.25
hlabel_font_size = font_size * 1.5
rounded_last_value = round(float(title_fillers['last_value']), 1)
if presidential_2020:
if rounded_last_value > 0:
hline_labels[1] = "Biden +" + str(
rounded_last_value) + hline_label_units
else:
hline_labels[1] = "Trump +" + str(
rounded_last_value) + hline_label_units
elif rounded_last_value > 0:
if hline_label_units == "seats":
hline_labels[1] = "D+" + str(round(
rounded_last_value / 10, 1)) + " " + hline_label_units
else:
hline_labels[1] = "D+" + str(
rounded_last_value) + hline_label_units
elif rounded_last_value < 0:
if hline_label_units == "seats":
hline_labels[1] = "R+" + str(round(
rounded_last_value / 10, 1)) + " " + hline_label_units
else:
hline_labels[1] = "Current R+" + str(
rounded_last_value) + hline_label_units
else:
hline_labels[1] = "Currently Tie"
hline_labels[0] = ""
# prepare figurex
fig = pl.figure(figsize=(width, width * height_to_width))
ax = fig.add_axes(axes_box)
# plot main data line
ax.plot(days,
vals,
color=data_line_color,
lw=data_lw * size_scale,
zorder=1000)
# plot circle on most recent data point
col = zone_colors[1] if vals[-1] > 0 else zone_colors[0]
ax.plot(days[-1],
vals[-1],
marker='o',
markersize=circle_size * size_scale,
markerfacecolor=col,
markeredgewidth=0,
zorder=1001)
# plot error bars
err = None
if yerr_columns is not None:
err = data[yerr_columns].values
if len(yerr_columns) == 1:
# then column with symmetrical value to be added
# otherwise abs values were assumed
err = err.squeeze()
err = np.array([vals - err, vals + err]).T
if x_minus_yvalues is not None:
err = x_minus_yvalues - err
err = err[:, ::-1] # keep [lower, upper]
ax.fill_between(
days,
err[:, 0],
err[:, 1],
**yerr_kw,
)
# x axis limits
first_day = first_day or np.min(day_of_year)
last_day = last_day or np.max(day_of_year)
first_date = doy2dt(first_day)
last_date = doy2dt(last_day)
ax.set_xlim(first_date, last_date)
# y axis limits
if ylim is None:
mindat = err[:, 0] if err is not None else vals
maxdat = err[:, 1] if err is not None else vals
minn = np.floor(np.min(mindat))
maxx = np.ceil(np.max(maxdat))
pad = np.round((maxx - minn) / 5)
y0, y1 = minn - pad, maxx + pad
if pad > 10:
def r210(val, fx=np.ceil):
return int(fx(val / 10.0)) * 10
y0 = r210(y0, np.ceil)
y1 = r210(y1, np.floor)
ylim = (y0, y1)
ax.set_ylim(ylim)
ylim = ax.get_ylim()
# x ticks at month boundaries
date_range = pd.date_range(first_date, last_date)
tick_dates = date_range[date_range.day == 1]
tick_dates_dt = np.array([td.to_datetime64() for td in tick_dates])
ax.set_xticks(tick_dates_dt)
ax.set_xticklabels([])
# x tick labels at centers of months
def get_mid(mo):
rang = calendar.monthrange(year, mo)
return int(np.round(rang[-1] / 2))
if not thumbnail:
mid_months = np.array([get_mid(d.month) for d in date_range])
tick_label_dates = date_range[date_range.day == mid_months]
for td in tick_label_dates:
loc = td.to_datetime64()
ax.text(loc,
-xtick_pad * size_scale,
mo_names[td.month],
color='k',
ha='center',
va='center',
fontsize=font_size,
transform=blend(ax.transData, ax.transAxes))
# y ticks at integer values
ax.set_yticks(
np.arange(ylim[0], ylim[1] + yticks_interval, yticks_interval))
if yticklab_format:
# ytick labels with formatting specifications
ytls = []
for t in ax.get_yticks():
plus = '+' if t != 0 else ''
ytls.append(f'{plus}{abs(t):0.0f}%')
ytls[0] = f'R {ytls[0]}'
ytls[-1] = f'D {ytls[-1]}'
ax.set_yticklabels(ytls)
# color ytick labels
for lab, pos in zip(ax.get_yticklabels(), ax.get_yticks()):
if pos > 0:
col = zone_colors[1]
elif pos < 0:
col = zone_colors[0]
else:
col = 'k'
lab.set_color(col)
# tick parameters
ax.tick_params(labelsize=font_size,
length=tick_length * size_scale,
pad=tick_pad * size_scale,
width=tick_lw * size_scale)
ax.tick_params(which='minor', length=0)
# grid and spines
ax.grid('on', linewidth=grid_lw * size_scale, alpha=grid_alpha)
ax.spines['right'].set_visible(False)
ax.spines['top'].set_visible(False)
pl.setp(ax.spines.values(), linewidth=spine_lw * size_scale)
# colors for hline labels and shading
col0, col1 = zone_colors
if color_reverse:
col0, col1 = col1, col0
# horizontal line
if hline_ypos is not None:
ax.axhline(hline_ypos, color=hline_color, lw=hline_lw * size_scale)
#***************
if hline_labels is not None:
pad_data_units = hline_lab_pad * np.diff(ax.get_ylim())
txt_kw = dict(fontsize=hlabel_font_size,
ha='center',
va='center',
transform=blend(ax.transAxes, ax.transData))
if hline_lab_colors is None:
hline_lab_colors = [col0, col1]
ax.text(
hline_lab_xpos,
hline_ypos - pad_data_units,
hline_labels[0].format(**title_fillers),
color=hline_lab_colors[0],
**txt_kw,
)
ax.text(
hline_lab_xpos,
hline_ypos + pad_data_units,
hline_labels[1].format(**title_fillers),
color=hline_lab_colors[1],
**txt_kw,
)
# background shading of plot regions
if shading:
ax.axhspan(ax.get_ylim()[0], hline_ypos, color=col0, **shading_kw)
ax.axhspan(hline_ypos, ax.get_ylim()[1], color=col1, **shading_kw)
# target bar
if strike_zone_data_file is not None:
strike_path = os.path.join(data_dir, strike_zone_data_file)
strike_data = pd.read_csv(strike_path)
values = np.array(strike_data.columns).astype(float)
if x_minus_yvalues is not None:
values = x_minus_yvalues - values
s1_lo, s1_hi, s2_lo, s2_hi = values
pairs = [s1_lo, s1_hi], [s2_lo, s2_hi]
zorders = [150, 149]
for (lo, hi), col, zo in zip(pairs, strike_colors, zorders):
ax.plot([election_day] * 2, [lo, hi],
lw=2 * size_scale,
color=col,
zorder=zo,
alpha=0.7,
transform=blend(ax.transData, ax.transData))
# text labels
if thumbnail:
ax.set_title(title_txt.format(**title_fillers),
y=0.9,
fontsize=font_size,
weight='bold')
else:
ax.set_title(title_txt.format(**title_fillers),
pad=title_pad * size_scale,
fontsize=font_size,
weight='bold')
ax.text(-ylab_pad * size_scale,
0.5,
ylab_txt,
fontsize=font_size,
rotation=ylab_rotation,
ha='center',
va='center',
transform=ax.transAxes)
# watermark
if not thumbnail:
ax.text(watermark_pos[0],
watermark_pos[1],
watermark,
fontsize=font_size - 4,
color='black',
alpha=0.3,
style='italic',
ha='center',
va='center',
zorder=200,
transform=ax.transAxes)
# CUSTOM type special plot
if plot_customization:
new_ax_offset = 3
ax2 = ax.twinx()
ax.spines['right'].set_visible(True)
ax.yaxis.tick_right()
ax2.yaxis.tick_left()
spec_ypos = 6
spec_span = 3
spec_col = '#f58025'
ax.axhline(spec_ypos, color=spec_col, lw=0.5 * size_scale)
ax.axhspan(
spec_ypos - spec_span,
spec_ypos + spec_span,
color=spec_col,
alpha=0.15,
zorder=0,
lw=0,
)
pad_data_units = 0.05 * np.diff(ax.get_ylim())
ax.text(0.79,
spec_ypos - pad_data_units,
'Special elections',
color=spec_col,
fontsize=font_size - 4,
alpha=0.8,
ha='center',
va='center',
transform=blend(ax.transAxes, ax.transData))
ax2.set_ylim(ax.get_ylim())
ax2.set_yticks(ax.get_yticks())
ax2.tick_params(labelsize=font_size,
length=tick_length * size_scale,
pad=tick_pad * size_scale,
width=tick_lw * size_scale)
ax2.tick_params(which='minor', length=0)
# ytick labels with formatting specifications
ytls, ytls2 = [], []
def make_str(t):
plus = '+' if t != 0 else ''
let = 'D' if t > 0 else 'R'
if t == 0: let = ''
return f'{let}{plus}{abs(t):0.0f}%'
for t in ax.get_yticks():
ytls.append(make_str(t))
ytls2.append(make_str(t - new_ax_offset))
if not thumbnail:
ax.set_yticklabels(ytls)
ax2.set_yticklabels(ytls2)
ax.text(1 + ylab_pad * size_scale,
0.5,
'2020 generic ballot D-R',
fontsize=font_size,
rotation=-ylab_rotation,
ha='center',
va='center',
transform=ax.transAxes)
lv = title_fillers['last_value']
lv = lv - new_ax_offset
if lv > 0:
title_fillers['party'] = 'D+'
elif lv < 0:
title_fillers['party'] = 'R+'
title_fillers['last_value'] = lv
if not thumbnail:
ax.set_title(title_txt.format(**title_fillers),
pad=title_pad * size_scale,
fontsize=font_size,
weight='bold')
else:
ax.set_title(title_txt.format(**title_fillers),
y=0.9,
fontsize=font_size,
weight='bold')
if thumbnail:
ax2.get_xaxis().set_visible(False)
ax2.get_yaxis().set_visible(False)
if thumbnail:
ax.get_xaxis().set_visible(False)
ax.get_yaxis().set_visible(False)
# save out figure
dpi = width_pixels_save / width
if not out_path.endswith(out_format):
out_path = f'{out_path}.{out_format}'
if thumbnail:
pl.savefig(out_path, dpi=dpi, bbox_inches='tight')
else:
pl.savefig(out_path, dpi=dpi)
pl.close(fig)
def generate_histogram(
# data parameters
data_dir=data_dir,
data_file=data_file,
vline_xpos=None,
xvals=None,
data_factor=100.0,
# figure parameters
width=width,
height_to_width=height_to_width,
width_pixels_save=width_pixels_save,
axes_box=axes_box,
# text parameters
title_txt='',
ylab_txt='',
ylab_rotation=0,
xlab_txt='',
vline_labels=None,
vline_lab_ypos=0.92,
last_date_label_pos=(0.01, 0.01),
watermark=watermark,
watermark_pos=watermark_pos,
# size parameters
size_scale=size_scale,
font_size=font_size,
# thickness parameters
bar_width=0.8,
grid_lw=0.25,
spine_lw=0.25,
vline_lw=1,
tick_lw=0.25,
# color parameters
bar_color='black',
col_D='#1660CE',
col_R='#C62535',
zone_colors=zone_colors,
vline_color='dimgrey',
grid_alpha=0.5,
shading=True,
color_reverse=False,
shading_kw=dict(zorder=0, alpha=0.15, lw=0),
# axes parameters
ylim=(0, 1.15),
xlim=(0, 100),
# tick parameters
tick_length=3,
tick_pad=1,
xticks_interval=1,
yticks_interval=1,
# distance/pad parameters
title_pad=3,
ylab_pad=6,
xlab_pad=3,
xtick_pad=0.015,
vline_lab_pad=0.09,
vline_font_size=font_size,
# output parameters
out_path=out_path,
out_format=out_format,
# generate a thumnail image
thumbnail=False,
):
# load data
data_path = os.path.join(data_dir, data_file)
data = pd.read_csv(data_path)
data = data.values.squeeze()
data = data * data_factor
if xvals is None:
xvals = np.arange(len(data))
if thumbnail:
font_size = font_size * 1.25
vline_font_size = font_size * 1.25
# use data to fill title text
def compute_rv_median(p, xs):
p = p / np.sum(p)
cum = np.cumsum(p)
#x1 = xs[np.argwhere(cum <= 0.5).squeeze()[-1]]
#x2 = xs[np.argwhere(cum >= 0.5).squeeze()[0]]
#return np.mean([x1, x2])
return xs[np.argwhere(cum >= 0.5).squeeze()[0]]
title_fillers = dict(
mean_value=np.sum(data * xvals) / np.sum(data),
median_value=compute_rv_median(data, xvals),
)
# prepare figure
fig = pl.figure(figsize=(width, width * height_to_width))
ax = fig.add_axes(axes_box)
# plot main data line
ax.bar(xvals, data, color=bar_color, width=bar_width, zorder=100)
# x axis limits
if xlim is None:
thresh = 0.0001 # cuts of tails below this y value
cs = np.cumsum(data)
try:
x0 = np.argwhere(cs < thresh).squeeze()[-1]
except IndexError:
x0 = 0
cs_r = np.cumsum(data[::-1])
try:
x1 = len(data) - np.argwhere(cs_r < thresh).squeeze()[-1]
except IndexError:
x1 = len(data) - 1
xlim = (xvals[x0], xvals[x1])
ax.set_xlim(xlim)
xlim = ax.get_xlim()
# y axis limits
ylim = [ylim[0], np.max(data) * ylim[1]]
ax.set_ylim(ylim)
ylim = ax.get_ylim()
# x ticks
ax.set_xticks(np.arange(xlim[0], xlim[1], xticks_interval))
# x tick labels
# y ticks
ax.set_yticks(
np.arange(ylim[0], ylim[1] + yticks_interval, yticks_interval))
# tick parameters
ax.tick_params(labelsize=font_size,
length=tick_length * size_scale,
pad=tick_pad * size_scale,
width=tick_lw * size_scale)
ax.tick_params(which='minor', length=0)
# grid and spines
ax.grid('on', linewidth=grid_lw * size_scale, alpha=grid_alpha)
ax.spines['right'].set_visible(False)
ax.spines['top'].set_visible(False)
pl.setp(ax.spines.values(), linewidth=spine_lw * size_scale)
# colors for vline labels and shading
col0, col1 = zone_colors
if color_reverse:
col0, col1 = col1, col0
# vertical line
if vline_xpos is not None:
ax.axvline(
vline_xpos, #+ (bar_width / 2 + (1-bar_width) / 2),
color=vline_color,
lw=vline_lw * size_scale,
zorder=100)
if vline_labels is not None:
pad_data_units = vline_lab_pad * np.diff(ax.get_xlim())
txt_kw = dict(fontsize=vline_font_size,
ha='center',
va='center',
transform=blend(ax.transData, ax.transAxes))
ax.text(
vline_xpos - pad_data_units,
vline_lab_ypos,
vline_labels[0],
color=col0,
**txt_kw,
)
ax.text(
vline_xpos + pad_data_units,
vline_lab_ypos,
vline_labels[1],
color=col1,
**txt_kw,
)
# background shading of plot regions
if shading:
ax.axvspan(ax.get_xlim()[0], vline_xpos, color=col0, **shading_kw)
ax.axvspan(vline_xpos, ax.get_xlim()[1], color=col1, **shading_kw)
# text labels
ax.set_title(title_txt.format(**title_fillers),
pad=title_pad * size_scale,
fontsize=font_size,
weight='bold')
ax.text(-ylab_pad * size_scale,
0.5,
ylab_txt,
fontsize=font_size,
rotation=ylab_rotation,
ha='center',
va='center',
transform=ax.transAxes)
ax.set_xlabel(
xlab_txt,
fontsize=font_size,
labelpad=xlab_pad * size_scale,
)
# last date label
if thumbnail: last_date_label_pos = None
if last_date_label_pos is not None:
h_filename = data_file.split('_')[0] + '_estimate_history'
c_filename = h_filename + '_columns'
hdata_path = os.path.join(data_dir, h_filename + '.csv')
cdata_path = os.path.join(data_dir, c_filename + '.csv')
try:
hist = pd.read_csv(hdata_path)
columns = pd.read_csv(cdata_path).columns
hist.columns = columns
last_date = doy2dt(np.max(hist.date.values))
datestr = last_date.strftime('%Y/%m/%d')
txt = f'Data date: {datestr}'
except:
txt = ''
ax.text(last_date_label_pos[0],
last_date_label_pos[1],
txt,
fontsize=font_size - 10,
color='black',
alpha=0.3,
style='italic',
ha='left',
va='bottom',
zorder=200,
transform=fig.transFigure)
# watermark
if not thumbnail:
ax.text(watermark_pos[0],
watermark_pos[1],
watermark,
fontsize=font_size - 4,
color='black',
alpha=0.3,
style='italic',
ha='center',
va='center',
zorder=200,
transform=ax.transAxes)
# adjustments if a thumbnail image
if thumbnail:
ax.get_xaxis().set_visible(False)
ax.get_yaxis().set_visible(False)
# save out figure
dpi = width_pixels_save / width
if not out_path.endswith(out_format):
out_path = f'{out_path}.{out_format}'
if thumbnail:
pl.savefig(out_path, dpi=dpi, bbox_inches='tight')
else:
pl.savefig(out_path, dpi=dpi)
pl.close(fig)
def plot(xdata, ydata):
"""Plot the ERT log loss figures.
Two cases: box-whisker plot is used for representing the data of all
functions, otherwise all data is represented using crosses.
"""
res = []
tmp = list(10**numpy.mean(i[numpy.isfinite(i)]) for i in ydata)
res.extend(
plt.plot(
xdata,
tmp,
ls='-',
color='k',
lw=3, #marker='+',
markersize=20,
markeredgewidth=3))
if max(len(i)
for i in ydata) < 20: # TODO: subgroups of function, hopefully.
for i, y in enumerate(ydata):
# plot all single data points
if (numpy.isfinite(y) == False).any():
assert not (numpy.isinf(y) * y > 0.).any()
assert not numpy.isnan(y).any()
ax = plt.gca()
trans = blend(ax.transData, ax.transAxes)
res.extend(
plt.plot((xdata[i], ), (0., ),
marker='+',
color=flierscolor,
ls='',
markersize=20,
markeredgewidth=3,
transform=trans,
clip_on=False))
res.append(
plt.text(xdata[i],
0.02,
'%d' % len(y[numpy.isinf(y)]),
transform=trans,
horizontalalignment='left',
verticalalignment='bottom'))
y = y[numpy.isfinite(y)]
if len(y) == 0:
continue
res.extend(
plt.plot([xdata[i]] * len(y),
10**numpy.array(y),
marker='+',
color=flierscolor,
ls='',
markersize=20,
markeredgewidth=3))
# plot dashed vertical line between min and max
plt.plot(
[xdata[i]] * 2,
10**numpy.array([min(y), max(y)]),
color='k', # marker='+',
ls='--',
linewidth=2) #, markersize=20, markeredgewidth=3)
# plot min and max with different symbol
#plt.plot([xdata[i]], 10**min(numpy.array(y)),
# marker='+', color='k',
# ls='', markersize=20, markeredgewidth=3)
#plt.plot([xdata[i]], 10**max(numpy.array(y)),
# marker='+', color='k',
# ls='', markersize=20, markeredgewidth=3)
else:
for i, y in enumerate(ydata):
# plot all single data points
if (numpy.isfinite(y) == False).any():
assert not (numpy.isinf(y) * y > 0.).any()
assert not numpy.isnan(y).any()
ax = plt.gca()
trans = blend(ax.transData, ax.transAxes)
res.extend(
plt.plot((xdata[i], ), (0, ),
marker='.',
color='k',
ls='',
markersize=20,
markeredgewidth=3,
transform=trans,
clip_on=False))
res.append(
plt.text(xdata[i],
0.02,
'%d' % len(y[numpy.isinf(y)]),
transform=trans,
horizontalalignment='left',
verticalalignment='bottom'))
y = y[numpy.isfinite(y)]
dictboxwhisker = boxplot(list(10**numpy.array(i) for i in ydata),
sym='',
notch=0,
widths=None,
positions=xdata)
#'medians', 'fliers', 'whiskers', 'boxes', 'caps'
plt.setp(dictboxwhisker['medians'], lw=3)
plt.setp(dictboxwhisker['boxes'], lw=3)
plt.setp(dictboxwhisker['caps'], lw=3)
plt.setp(dictboxwhisker['whiskers'], lw=3)
for i in dictboxwhisker.values():
res.extend(i)
res.extend(
plt.plot(xdata,
list(10**min(i) for i in ydata),
marker='.',
markersize=20,
color='k',
ls=''))
res.extend(
plt.plot(xdata,
list(10**max(i) for i in ydata),
marker='.',
markersize=20,
color='k',
ls=''))
return res
def annotate(entry0, entry1, dim, minfvalue=1e-8, nbtests=1):
"""Display some annotations associated to the graphs generated."""
isEarlyStop = False
ha = 'left'
va = 'center'
lastfvalue = min(entry0.evals[-1][0], entry1.evals[-1][0])
if minfvalue < lastfvalue:
isEarlyStop = True
#ha = 'center'
#va = 'top'
if not minfvalue or minfvalue < lastfvalue:
minfvalue = lastfvalue
line = []
data0 = entry0.detEvals([minfvalue])[0]
evals0 = data0.copy()
succ = (np.isnan(evals0) == False)
evals0[np.isnan(evals0)] = entry0.maxevals[np.isnan(evals0)]
line.append(toolsstats.sp(evals0, issuccessful=succ))
data1 = entry1.detEvals([minfvalue])[0]
evals1 = data1.copy()
succ = (np.isnan(evals1) == False)
evals1[np.isnan(evals1)] = entry1.maxevals[np.isnan(evals1)]
line.append(toolsstats.sp(evals1, issuccessful=succ))
# What's the situation?
txt = '%dD' % dim
if (line[0][2] > 0 and line[1][2] > 0 and line[1][2] < 10):
tmp = str(int(line[1][2]))
tmp2 = str(int(line[0][2]))
txt = tmp + '/' + tmp2
dims = dimension_index
ax = plt.gca()
assert line[0][2] > 0 or line[1][2] > 0
signdata = line[1][0] - line[0][0]
if line[0][2] > 0 and line[1][2] > 0:
trans = ax.transData
annotcoord = [minfvalue, line[1][0]/line[0][0]]
elif line[0][2] == 0:
trans = blend(ax.transData, ax.transAxes)
annotcoord = [minfvalue, -line[1][1]/2 + 0.5 + offset*(5-dims[dim])]
#if va == 'top':
# va = 'bottom'
elif line[1][2] == 0:
trans = blend(ax.transData, ax.transAxes)
annotcoord = [minfvalue, line[0][1]/2 + 0.5 - offset*(5-dims[dim])]
plt.text(annotcoord[0], annotcoord[1], txt, horizontalalignment=ha,
verticalalignment=va, transform=trans)
#ranksum test
line0 = np.power(data0, -1.)
line0[np.isnan(line0)] = -entry0.finalfunvals[np.isnan(line0)]
line1 = np.power(data1, -1.)
line1[np.isnan(line1)] = -entry1.finalfunvals[np.isnan(line1)]
# one-tailed statistics: scipy.stats.mannwhitneyu, two-tailed statistics: scipy.stats.ranksumtest
z, p = ranksumtest(line0, line1)
# Set the correct line in data0 and data1
nbstars = 0
# sign of z-value and data must agree
if ((nbtests * p) < 0.05 and (z * signdata) > 0):
nbstars = np.min([5, -np.ceil(np.log10(nbtests * p + 1e-99))])
if nbstars > 0:
xstars = annotcoord[0] * np.power(incrstars, np.arange(1., 1. + nbstars))
# the additional slicing [0:int(nbstars)] is due to
# np.arange(1., 1. - 0.1 * nbstars, -0.1) not having the right number
# of elements due to numerical error
ystars = [annotcoord[1]] * nbstars
try:
h = plt.plot(xstars, ystars, marker='*', ls='', color='w',
markersize=5*linewidth, markeredgecolor='k',
markerfacecolor='None',
zorder=20, markeredgewidth = 0.4 * linewidth,
transform=trans, clip_on=False)
except KeyError:
#Version problem
h = plt.plot(xstars, ystars, marker='+', ls='', color='w',
markersize=2.5*linewidth, markeredgecolor='k',
zorder=20, markeredgewidth = 0.2 * linewidth,
transform=trans, clip_on=False)
def main(dsList0, dsList1, outputdir, verbose=True):
"""Generate a scatter plot figure."""
#plt.rc("axes", labelsize=24, titlesize=24)
#plt.rc("xtick", labelsize=20)
#plt.rc("ytick", labelsize=20)
#plt.rc("font", size=20)
#plt.rc("legend", fontsize=20)
dictFunc0 = dsList0.dictByFunc()
dictFunc1 = dsList1.dictByFunc()
funcs = set(dictFunc0.keys()) & set(dictFunc1.keys())
for f in funcs:
dictDim0 = dictFunc0[f].dictByDim()
dictDim1 = dictFunc1[f].dictByDim()
dims = set(dictDim0.keys()) & set(dictDim1.keys())
#set_trace()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
xdata = numpy.array(entry0.detERT(targets))
ydata = numpy.array(entry1.detERT(targets))
tmp = (numpy.isinf(xdata) == False) * (numpy.isinf(ydata) == False)
if tmp.any():
try:
plt.plot(xdata[tmp],
ydata[tmp],
ls='',
markersize=markersize,
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=3)
except KeyError:
plt.plot(xdata[tmp],
ydata[tmp],
ls='',
markersize=markersize,
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=3)
#try:
# plt.scatter(xdata[tmp], ydata[tmp], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3)
#except ValueError:
# set_trace()
#ax = plt.gca()
ax = plt.axes()
tmp = numpy.isinf(xdata) * (numpy.isinf(ydata) == False)
if tmp.any():
trans = blend(ax.transAxes, ax.transData)
#plt.scatter([1.]*numpy.sum(tmp), ydata[tmp], s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1.] * numpy.sum(tmp),
ydata[tmp],
markersize=markersize,
ls='',
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=3,
transform=trans,
clip_on=False)
except KeyError:
plt.plot([1.] * numpy.sum(tmp),
ydata[tmp],
markersize=markersize,
ls='',
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=3,
transform=trans,
clip_on=False)
#set_trace()
tmp = (numpy.isinf(xdata) == False) * numpy.isinf(ydata)
if tmp.any():
trans = blend(ax.transData, ax.transAxes)
# plt.scatter(xdata[tmp], [1.-offset]*numpy.sum(tmp), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot(xdata[tmp], [1. - offset] * numpy.sum(tmp),
markersize=markersize,
ls='',
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=3,
transform=trans,
clip_on=False)
except KeyError:
plt.plot(xdata[tmp], [1. - offset] * numpy.sum(tmp),
markersize=markersize,
ls='',
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=3,
transform=trans,
clip_on=False)
tmp = numpy.isinf(xdata) * numpy.isinf(ydata)
if tmp.any():
# plt.scatter(xdata[tmp], [1.-offset]*numpy.sum(tmp), s=10, marker=markers[i],
# facecolor='None', edgecolor=colors[i], linewidth=3,
# transform=trans)
try:
plt.plot([1. - offset] * numpy.sum(tmp),
[1. - offset] * numpy.sum(tmp),
markersize=markersize,
ls='',
marker=markers[i],
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=3,
transform=ax.transAxes,
clip_on=False)
except KeyError:
plt.plot([1. - offset] * numpy.sum(tmp),
[1. - offset] * numpy.sum(tmp),
markersize=markersize,
ls='',
marker='x',
markerfacecolor='None',
markeredgecolor=colors[i],
markeredgewidth=3,
transform=ax.transAxes,
clip_on=False)
#set_trace()
beautify()
for i, d in enumerate(dimensions):
try:
entry0 = dictDim0[d][0] # should be only one element
entry1 = dictDim1[d][0] # should be only one element
except (IndexError, KeyError):
continue
minbnd, maxbnd = plt.xlim()
plt.plot(
(entry0.mMaxEvals(), entry0.mMaxEvals()),
# (minbnd, entry1.mMaxEvals()), ls='-', color=colors[i],
(max([minbnd, entry1.mMaxEvals() / 10.]), entry1.mMaxEvals()),
ls='-',
color=colors[i],
zorder=-1)
plt.plot( # (minbnd, entry0.mMaxEvals()),
(max([minbnd, entry0.mMaxEvals() / 10.]), entry0.mMaxEvals()),
(entry1.mMaxEvals(), entry1.mMaxEvals()),
ls='-',
color=colors[i],
zorder=-1)
plt.xlim(minbnd, maxbnd)
plt.ylim(minbnd, maxbnd)
#Set the boundaries again: they changed due to new plots.
#plt.axvline(entry0.mMaxEvals(), ls='--', color=colors[i])
#plt.axhline(entry1.mMaxEvals(), ls='--', color=colors[i])
if isBenchmarkinfosFound:
try:
plt.ylabel(funInfos[f])
except IndexError:
pass
filename = os.path.join(outputdir, 'ppscatter_f%03d' % f)
saveFigure(filename, verbose=verbose)
plt.close()
