def __init__(self):
HistogramPlot.__init__(self)
self.title = 'Model'
def plot_marx_spectrum(id=None, elow=None, ehigh=None, ewidth=None,
norm=None, overplot=False, clearwindow=True):
"""Plots the spectrum as used by MARX.
Parameters
----------
id : int or string or None
If id is None then the default Sherpa id is used. This
dataset must have a grid and source model defined.
elow, ehigh, ewidth : number or None
Only used if all three are given, otherwise the grid
of the data set is used. The elo value gives the start
of the grid (the left edge of the first bin) and ehi
is the end of the grid (the right edge of the last bin),
and ewidth the bin width, all in keV.
norm : number or None
Multiply the flux values by this value, if set.
overplot : bool, optional
If ``True`` then the data is added to the current plot,
otherwise a new plot is created.
clearwindow: bool, optional
If ``True`` then clear out the current plot area of all
existing plots. This is not used if ``overplot`` is set.
"""
vals = _get_chart_spectrum(id=id, elow=elow,
ehigh=ehigh, ewidth=ewidth,
norm=norm)
# In sherpa_contrib.utils.plot_instmap_weights it was found useful
# to convert the Y values to float32 from float64 to avoid
# precision issues confusing the plot (i.e. showing structure that
# isn't meaningful to the user). Is this going to be needed
# here too?
#
hplot = HistogramPlot()
hplot.xlo = vals['xlo']
hplot.xhi = vals['xhi']
hplot.y = vals['y'] / (vals['xhi'] - vals['xlo'])
hplot.xlabel = 'Energy (keV)'
hplot.title = 'MARX Spectrum: {}'.format(vals['model'])
# LaTeX support depends on the backend
if backend.name == 'pylab':
hplot.ylabel = 'Flux (photon cm$^{-2}$ s$^{-1}$ keV$^{-1}$)'
elif backend.name == 'chips':
hplot.ylabel = 'Flux (photon cm^{-2} s^{-1} keV^{-1})'
else:
hplot.ylabel = 'Flux (photon cm^-2 s^-1)'
hplot.plot(overplot=overplot, clearwindow=clearwindow)
def __init__(self):
self.units = None
self.mask = None
HistogramPlot.__init__(self)
self.title = 'Source'
def plot(self, overplot=False, clearwindow=True, **kwargs):
"""Plot the weights values.
Parameters
----------
overplot : bool, optional
If ``True`` then the data is added to the current plot,
otherwise a new plot is created.
clearwindow: bool, optional
If ``True`` then clear out the current plot area of
all existing plots. This is not used if ``overplot`` is
set.
**kwargs
Assumed to be plot preferences that override the
HistogramPlot.histo_plot preferences.
Notes
-----
The data plot preferences are used to control the
appearance of the plot: at present the following fields
are used::
``xlog``
``ylog``
``color``
Examples
--------
>>> hplot.plot()
>>> hplot.plot(ylog=True, color='orange', linestyle='dotted')
"""
# Create a Sherpa histogram plot. Unlike other plot
# classes there is no prepare method, which means
# that we do not have to create a temporary data object,
# but do have to set the attributes manually.
#
# This assumes that it is okay to use a histogram-style
# plot, since pre CIAO 4.2 there were numeric problems with
# bin edges that made these plots look ugly, so a "curve"
# was used.
#
# To avoid issues with numeric accuracy (I have seen a
# case where the variation in the weight values was ~2e-16
# and matplotlib showed this structure), convert the weights
# to 32-bit before plotting. This is a lot simpler than
# dealing with some sort of run-length-encoding scheme to
# group bins that are "close enough" numerically.
#
hplot = HistogramPlot()
hplot.xlo = self.xlo
hplot.xhi = self.xhi
hplot.y = self.weight.astype(np.float32)
hplot.xlabel = 'Energy (keV)'
hplot.ylabel = 'Weights'
hplot.title = 'Weights for {}: '.format(self.id) + \
' {}'.format(self.modelexpr)
# There is no validation of the preference values.
#
# I have removed linestyle from this list since the pylab
# backend default is 'None', which ends up meaning nothing
# appears to get drawn. Which is less-than helpful.
# linecolor also doen't appear to be used, but color is.
#
# names = ['xlog', 'ylog', 'linestyle', 'linecolor']
names = ['xlog', 'ylog', 'color']
prefs = ui.get_data_plot_prefs()
for name in names:
value = prefs.get(name, None)
if value is not None:
hplot.histo_prefs[name] = value
hplot.plot(overplot=overplot, clearwindow=clearwindow, **kwargs)
