def testBootstrapCI(self):
''' test a line plot with confidence intervals from bootstrapping '''
fig,ax = getFigAx(1, name=sys._getframe().f_code.co_name[4:], **figargs) # use test method name as title
assert fig.__class__.__name__ == 'MyFigure'
assert fig.axes_class.__name__ == 'MyAxes'
assert not isinstance(ax,(list,tuple)) # should return a "naked" axes
# settings
varlist = [self.var1, self.var2]
nbins = 15
# add regular histogram for comparison
bins, ptchs = ax.histogram(varlist, bins=nbins, legend=2, alpha=0.5, rwidth=0.8, histtype='bar')
# histtype = 'bar' | 'barstacked' | 'step' | 'stepfilled'
assert len(ptchs) == 2
assert len(bins) == nbins
vmin = np.min([var.min() for var in varlist])
vmax = np.max([var.max() for var in varlist])
#print bins[0], vmin; print bins[-1], vmax
assert bins[0] == vmin and bins[-1] == vmax
# add bootstrap plot with errorbars
support = np.linspace(vmin, vmax, 100)
fitvars = [var.fitDist(dist=self.dist,lflatten=True, lbootstrap=True, nbs=1000) for var in varlist]
ax.bootPlot(fitvars[0], support=support, errorscale=0.5, linewidth=2, lsmooth=False,
percentiles=None, lvar=True, lvarBand=False, lmean=True, reset_color=True)
ax.bootPlot(fitvars[1], support=support, errorscale=0.5, linewidth=2, lsmooth=False,
percentiles=(0.25,0.75), lvar=False, lvarBand=False, lmedian=True, reset_color=False)
# add the actual distribution
ax.linePlot(self.distVar, support=support, linewidth=1, marker='^')
# add statistical info
pstr = "p-values for '{:s}':\n".format(self.dist)
for var in varlist:
pstr += ' {:<9s} {:3.2f}\n'.format(var.name, var.kstest(dist=self.dist, asVar=False))
pstr += ' 2-samples {:3.2f}\n'.format(kstest(varlist[0], varlist[1], lflatten=True))
ax.addLabel(label=pstr, loc=1, lstroke=False, lalphabet=True, size=None, prop=None)
def testBasicHistogram(self):
''' a simple bar plot of two normally distributed samples '''
fig,ax = getFigAx(1, name=sys._getframe().f_code.co_name[4:], **figargs) # use test method name as title
assert fig.__class__.__name__ == 'MyFigure'
assert fig.axes_class.__name__ == 'MyAxes'
assert not isinstance(ax,(list,tuple)) # should return a "naked" axes
# settings
varlist = [self.var1, self.var2]
nbins = 15
# create regular histogram
bins, ptchs = ax.histogram(varlist, bins=nbins, legend=2, alpha=0.5, rwidth=0.8, histtype='bar')
# histtype = 'bar' | 'barstacked' | 'step' | 'stepfilled'
assert len(ptchs) == 2
assert len(bins) == nbins
vmin = np.min([var.min() for var in varlist])
vmax = np.max([var.max() for var in varlist])
#print bins[0], vmin; print bins[-1], vmax
assert bins[0] == vmin and bins[-1] == vmax
# add a KDE plot
support = np.linspace(vmin, vmax, 100)
kdevars = [var.kde(lflatten=True, lbootstrap=True, nbs=10) for var in varlist]
# N.B.: the bootstrapping is just to test bootstrap tolerance in regular plotting methods
ax.linePlot(kdevars, support=support, linewidth=2)
# add label
pval = kstest(varlist[0], varlist[1], lflatten=True)
pstr = 'p-value = {:3.2f}'.format(pval)
ax.addLabel(label=pstr, loc=1, lstroke=False, lalphabet=True, size=None, prop=None)
def testBootstrapCI(self):
''' test a line plot with confidence intervals from bootstrapping '''
fig,ax = getFigAx(1, name=sys._getframe().f_code.co_name[4:], **figargs) # use test method name as title
assert fig.__class__.__name__ == 'MyFigure'
assert fig.axes_class.__name__ == 'MyAxes'
assert not isinstance(ax,(list,tuple)) # should return a "naked" axes
# settings
varlist = [self.var1, self.var2]
nbins = 15
# add regular histogram for comparison
bins, ptchs = ax.histogram(varlist, bins=nbins, legend=2, alpha=0.5, rwidth=0.8, histtype='bar')
# histtype = 'bar' | 'barstacked' | 'step' | 'stepfilled'
assert len(ptchs) == 2
assert len(bins) == nbins
vmin = np.min([var.min() for var in varlist])
vmax = np.max([var.max() for var in varlist])
#print bins[0], vmin; print bins[-1], vmax
assert bins[0] == vmin and bins[-1] == vmax
# add bootstrap plot with errorbars
support = np.linspace(vmin, vmax, 100)
fitvars = [var.fitDist(dist=self.dist,lflatten=True, lbootstrap=True, nbs=1000) for var in varlist]
ax.bootPlot(fitvars[0], support=support, errorscale=0.5, linewidth=2, lsmooth=False,
percentiles=None, lvar=True, lvarBand=False, lmean=True, reset_color=True)
ax.bootPlot(fitvars[1], support=support, errorscale=0.5, linewidth=2, lsmooth=False,
percentiles=(0.25,0.75), lvar=False, lvarBand=False, lmedian=True, reset_color=False)
# add the actual distribution
ax.linePlot(self.distVar, support=support, linewidth=1, marker='^')
# add statistical info
pstr = "p-values for '{:s}':\n".format(self.dist)
for var in varlist:
pstr += ' {:<9s} {:3.2f}\n'.format(var.name, var.kstest(dist=self.dist, asVar=False))
pstr += ' 2-samples {:3.2f}\n'.format(kstest(varlist[0], varlist[1], lflatten=True))
ax.addLabel(label=pstr, loc=1, lstroke=False, lalphabet=True, size=None, prop=None)
def testBasicHistogram(self):
''' a simple bar plot of two normally distributed samples '''
fig,ax = getFigAx(1, name=sys._getframe().f_code.co_name[4:], **figargs) # use test method name as title
assert fig.__class__.__name__ == 'MyFigure'
assert fig.axes_class.__name__ == 'MyAxes'
assert not isinstance(ax,(list,tuple)) # should return a "naked" axes
# settings
varlist = [self.var1, self.var2]
nbins = 15
# create regular histogram
bins, ptchs = ax.histogram(varlist, bins=nbins, legend=2, alpha=0.5, rwidth=0.8, histtype='bar')
# histtype = 'bar' | 'barstacked' | 'step' | 'stepfilled'
assert len(ptchs) == 2
assert len(bins) == nbins
vmin = np.min([var.min() for var in varlist])
vmax = np.max([var.max() for var in varlist])
#print bins[0], vmin; print bins[-1], vmax
assert bins[0] == vmin and bins[-1] == vmax
# add a KDE plot
support = np.linspace(vmin, vmax, 100)
kdevars = [var.kde(lflatten=True, lbootstrap=True, nbs=10) for var in varlist]
# N.B.: the bootstrapping is just to test bootstrap tolerance in regular plotting methods
ax.linePlot(kdevars, support=support, linewidth=2)
# add label
pval = kstest(varlist[0], varlist[1], lflatten=True)
pstr = 'p-value = {:3.2f}'.format(pval)
ax.addLabel(label=pstr, loc=1, lstroke=False, lalphabet=True, size=None, prop=None)
def testSamplePlot(self):
""" test a line and band plot showing the mean/median and given percentiles """
fig, ax = getFigAx(1, name=sys._getframe().f_code.co_name[4:], **figargs) # use test method name as title
assert fig.__class__.__name__ == "MyFigure"
assert fig.axes_class.__name__ == "MyAxes"
assert not isinstance(ax, (list, tuple)) # should return a "naked" axes
# settings
varlist = [self.var1, self.var2]
nbins = 15
# add regular histogram for comparison
bins, ptchs = ax.histogram(varlist, bins=nbins, legend=2, alpha=0.5, rwidth=0.8, histtype="bar")
# histtype = 'bar' | 'barstacked' | 'step' | 'stepfilled'
assert len(ptchs) == 2
assert len(bins) == nbins
vmin = np.min([var.min() for var in varlist])
vmax = np.max([var.max() for var in varlist])
# print bins[0], vmin; print bins[-1], vmax
assert bins[0] == vmin and bins[-1] == vmax
# add bootstrap plot with errorbars
support = np.linspace(vmin, vmax, 100)
fitvars = [var.fitDist(dist=self.dist, lflatten=True, lbootstrap=True, nbs=1000) for var in varlist]
# add the actual distribution
ax.linePlot(self.distVar, support=support, linewidth=1, marker="^")
# add simple sample (using bootstrap axis as sample
ax.samplePlot(
fitvars[0],
support=support,
linewidth=2,
lsmooth=False,
sample_axis="bootstrap",
percentiles=(0.05, 0.95),
lmedian=True,
reset_color=True,
)
# replicate axis and add random noise to mean
rndfit = fitvars[1].insertAxis(axis="sample", iaxis=0, length=100)
rndfit.data_array += np.random.randn(*rndfit.shape) / 100.0
ax.samplePlot(
rndfit,
support=support,
linewidth=2,
lsmooth=False,
bootstrap_axis="bootstrap",
sample_axis=("no_axis", "sample"),
percentiles=(0.25, 0.75),
lmean=True,
reset_color=False,
)
# add statistical info
pstr = "p-values for '{:s}':\n".format(self.dist)
for var in varlist:
pstr += " {:<9s} {:3.2f}\n".format(var.name, var.kstest(dist=self.dist, asVar=False))
pstr += " 2-samples {:3.2f}\n".format(kstest(varlist[0], varlist[1], lflatten=True))
ax.addLabel(label=pstr, loc=1, lstroke=False, lalphabet=True, size=None, prop=None)
def testSamplePlot(self):
''' test a line and band plot showing the mean/median and given percentiles '''
fig,ax = getFigAx(1, name=sys._getframe().f_code.co_name[4:], **figargs) # use test method name as title
assert fig.__class__.__name__ == 'MyFigure'
assert fig.axes_class.__name__ == 'MyAxes'
assert not isinstance(ax,(list,tuple)) # should return a "naked" axes
# settings
varlist = [self.var1, self.var2]
nbins = 15
# add regular histogram for comparison
bins, ptchs = ax.histogram(varlist, bins=nbins, legend=2, alpha=0.5, rwidth=0.8, histtype='bar')
# histtype = 'bar' | 'barstacked' | 'step' | 'stepfilled'
assert len(ptchs) == 2
assert len(bins) == nbins
vmin = np.min([var.min() for var in varlist])
vmax = np.max([var.max() for var in varlist])
#print bins[0], vmin; print bins[-1], vmax
assert bins[0] == vmin and bins[-1] == vmax
# add bootstrap plot with errorbars
support = np.linspace(vmin, vmax, 100)
fitvars = [var.fitDist(dist=self.dist,lflatten=True, lbootstrap=True, nbs=1000) for var in varlist]
# add the actual distribution
ax.linePlot(self.distVar, support=support, linewidth=1, marker='^')
# add simple sample (using bootstrap axis as sample
ax.samplePlot(fitvars[0], support=support, linewidth=2, lsmooth=False,
sample_axis='bootstrap', percentiles=(0.05,0.95), lmedian=True, reset_color=True)
# replicate axis and add random noise to mean
rndfit = fitvars[1].insertAxis(axis='sample',iaxis=0,length=100)
rndfit.data_array += np.random.randn(*rndfit.shape)/100.
ax.samplePlot(rndfit, support=support, linewidth=2, lsmooth=False,
bootstrap_axis='bootstrap', sample_axis=('no_axis','sample'),
percentiles=(0.25,0.75), lmean=True, reset_color=False)
# add statistical info
pstr = "p-values for '{:s}':\n".format(self.dist)
for var in varlist:
pstr += ' {:<9s} {:3.2f}\n'.format(var.name, var.kstest(dist=self.dist, asVar=False))
pstr += ' 2-samples {:3.2f}\n'.format(kstest(varlist[0], varlist[1], lflatten=True))
ax.addLabel(label=pstr, loc=1, lstroke=False, lalphabet=True, size=None, prop=None)
