Step-by-Step guide to changing the look of Sherpa plots

Getting Started

The data used in this thread is available in the sherpa.tar.gz file, as described in the "Getting Started" thread.

For the various plots discussed below we use the same dataset and model as used in the Estimating Errors and Confidence Levels thread:

sherpa> data source_grouped_pi.fits
The inferred file type is PHA.  If this is not what you want, please
specify the type explicitly in the data command.
WARNING: statistical errors specified in the PHA file.
         These are currently IGNORED.  To use them, type:
         READ ERRORS "<filename>[cols CHANNEL,STAT_ERR]" fitsbin
RMF is being input from:
  /data/ciao/rmf.fits
ARF is being input from:
  /data/ciao/arf.fits
sherpa> ignore energy :0.5,8:
sherpa> source = xswabs[abs] * powlaw1d[p1]
abs.nH parameter value [0.1]
p1.gamma parameter value [1]
p1.ref parameter value [4]
p1.ampl parameter value [0.000149261]
sherpa> fit
 LVMQT: V2.0
 LVMQT: initial statistic value = 4583.05
 LVMQT: final statistic value = 83.2873 at iteration 8
           abs.nH  2.4061  10^22/cm^2  
            p1.gamma  1.51851     
            p1.ampl  0.000241434     

Background Information

Plots are created in Sherpa by the LPLOT command.

Many aspects of the appearance of the plots can be changed by setting values in the Sherpa configuration variables (state objects). This is discussed in the Configuration of Sherpa with Sherpa State Objects section of ahelp sherpa, and the variables relevant to plotting are: sherpa.plot, sherpa.dataplot, sherpa.fitplot, sherpa.resplot, and sherpa.multiplot.

By changing the values of the fields in these variables you change how new plots of the given type will be drawn. There are also several utility routines - such as set_log, set_lin, set_erron, and set_erroff to change some of the common options.

Changing "DATA" plots

The default settings for "DATA" plots produces this plot  for the dataset used in this example. The look of the plot is controlled by the fields of the sherpa.dataplot plotting variable. These can be viewed using the print() command:

sherpa> print(sherpa.dataplot)
x_errorbars      =  0
y_errorbars      =  1
errs_style       =  bar
errs_type        =  both
x_log            =  0
y_log            =  0
curvestyle       =  noline
curvecolor       =  default
symbolstyle      =  square
symbolcolor      =  default
symbolsize       =  2
xlabel_size      =  1.5
ylabel_size      =  1.5
zlabel_size      =  1.5
title_size       =  1.5
tickvals_size    =  1.5
prefunc          =  NULL
postfunc         =  NULL

We decide we want the y axis to be drawn in a logarithmic scale and the symbols to be drawn in magenta whilst keeping the remaining options unchanged.

sherpa> sherpa.dataplot.y_log = 1
sherpa> sherpa.dataplot.symbolcolor = "magenta"
sherpa> print(sherpa.dataplot)
x_errorbars      =  0
y_errorbars      =  1
errs_style       =  bar
errs_type        =  both
x_log            =  0
y_log            =  1
curvestyle       =  noline
curvecolor       =  default
symbolstyle      =  square
symbolcolor      =  magenta
symbolsize       =  2
xlabel_size      =  1.5
ylabel_size      =  1.5
zlabel_size      =  1.5
title_size       =  1.5
tickvals_size    =  1.5
prefunc          =  NULL
postfunc         =  NULL
sherpa> lplot data

The new version of the plot looks like this . Note that the color is specified as a string, unlike the chips.symbolcolor field of the ChIPS state object which takes an integer.

We now decide that we want to make further changes to the plot:

sherpa> sherpa.dataplot.x_log = 1
sherpa> sherpa.dataplot.curvestyle = "step"
sherpa> sherpa.dataplot.curvecolor = "cyan"
sherpa> sherpa.dataplot.symbolstyle = "none"
sherpa> sherpa.dataplot.title_size = 2
sherpa> lplot 2 data arf

The resulting plot looks like this . The top plot has been changed to match the settings asked for, whilst the bottom plot - of the ARF - does not use any of these settings, because the ARF plot is controlled by the sherpa.plot plot variable.

Although the plot title was changed to a size of 2, from its default value of 1.5, the new plot does not have a title. This is because we draw two plots here and Sherpa only sets the title when an individual plot - i.e. "lplot data" - is created.

Changing the "ARF" plot to match

In the previous section we changed Sherpa so that all "DATA" plots have a garish color scheme. However, other plots - such as of the ARF - do not share this scheme. We can set the fields in the sherpa.plot plot variable to change this:

sherpa> sherpa.plot.x_log = 1
sherpa> sherpa.plot.y_log = 1
sherpa> sherpa.plot.curvecolor = "cyan"
sherpa> sherpa.plot.title_size = 2
sherpa> lplot 2 data arf

which produces this plot .

Since the sherpa.plot plot object controls the look of many plots, not just those of the ARF, the following changes will also be seen in these other plots. For instance:

sherpa> lp model

The resulting plot  matches the style of the ARF plot. Since there is only one plot here we can see that the title has indeed been increased.

Changing the "FIT" plots

Although there are several different plot variables:

• sherpa.plot for general plots, such as the ARF and model components
• sherpa.dataplot for plotting the source and background data
• sherpa.resplot for plotting "residual" style plots
• and sherpa.fitplot for plotting fits

they are all essentially identical - except for fits. This is because the sherpa.fitplot object contains extra variables to control how the fitted model is displayed.

Here we change the fit-style plots; although most of fields are the same as seen with previous plots, there are extra ones (beginning with "fit_"):

sherpa> print(sherpa.fitplot)
x_errorbars      =  0
y_errorbars      =  1
errs_style       =  bar
errs_type        =  both
x_log            =  0
y_log            =  0
curvestyle       =  noline
curvecolor       =  default
symbolstyle      =  square
symbolcolor      =  default
symbolsize       =  2
fit_curvestyle   =  step
fit_curvecolor   =  red
fit_symbolstyle  =  none
fit_symbolcolor  =  default
fit_symbolsize   =  2
xlabel_size      =  1.5
ylabel_size      =  1.5
zlabel_size      =  1.5
title_size       =  1.5
tickvals_size    =  1.5
prefunc          =  NULL
postfunc         =  NULL
sherpa> sherpa.fitplot.x_log = 1
sherpa> sherpa.fitplot.y_log = 1
sherpa> sherpa.fitplot.symbolstyle = "bigpoint"
sherpa> sherpa.fitplot.symbolcolor = "blue"
sherpa> sherpa.fitplot.symbolsize = 1
sherpa> sherpa.fitplot.fit_curvecolor = "green"
sherpa> sherpa.fitplot.tickvals_size = 2
sherpa> lp fit

The resulting plot looks like this .

Making global changes

Whilst it is useful to be able to change the look of different plot types, there are a number of changes that you may wish to make to all plots. If you do this often then setting the same fields in the different plot variables is just too much typing!

Sherpa defines several functions to help out - and, as discussed below - you can write your own.

sherpa> set_log

sherpa> sherpa.plot.x_log = 1
sherpa> sherpa.plot.y_log = 1
sherpa> sherpa.dataplot.x_log = 1
sherpa> sherpa.dataplot.y_log = 1
sherpa> sherpa.fitplot.x_log = 1
sherpa> sherpa.fitplot.y_log = 1
sherpa> sherpa.resplot.x_log = 1
sherpa> sherpa.resplot.y_log = 1

sherpa> set_yerroff

sherpa> sherpa.plot.y_errorbars = 0
sherpa> sherpa.dataplot.y_errorbars = 0
sherpa> sherpa.fitplot.y_errorbars = 0
sherpa> sherpa.resplot.y_errorbars = 0

define set_myplot() {
  set_ylog;
  sherpa.resplot.y_log = 0;
}

sherpa> evalfile("myplot.sl")
1
sherpa> set_myplot

Saving your changes

The save_state() command can be used to save your settings so that they can be used in other Sherpa sessions.

If called with no arguments, save_state() will write out the contents of the Sherpa configuration variables to the file $HOME/.sherpa-state-rc. This file will be over-written without warning (so you should not make any changes to it). When Sherpa starts it will automatically load in the settings from this file, so your plots will retain the look you have set for them.

If called with an argument, save_state() will write out the settings to the file name (its argument) instead of $HOME/.sherpa-state-rc. This file can then be read into a Sherpa session using the USE command; this is useful if you want to set up different plot styles for use in different situations.

The beginning of the file will look something like:

% Sherpa state for ciaouser, Wed Jul 30 17:45:13 2006

sherpa.plot.x_errorbars = 0
sherpa.plot.y_errorbars = 0
sherpa.plot.errs_style = "bar"
sherpa.plot.errs_type = "both"
sherpa.plot.x_log = 1
sherpa.plot.y_log = 1
sherpa.plot.curvestyle = "step"
sherpa.plot.curvecolor = "cyan"
sherpa.plot.symbolstyle = "none"
sherpa.plot.symbolcolor = "default"
sherpa.plot.symbolsize = 2
sherpa.plot.xlabel_size = 1.5
sherpa.plot.ylabel_size = 1.5
sherpa.plot.zlabel_size = 1.5
sherpa.plot.title_size = 2
sherpa.plot.tickvals_size = 1.5
sherpa.plot.prefunc = NULL
sherpa.plot.postfunc = NULL

...

(the values will depend on what changes you have made to the plotting variables).