GitLab

To ease the process of restructuring the isisscripts I've added
a simple test that currently only checks if a file '.check' is
present in any of the subfolders. The idea is, that this file lists
all files that are accepted for the restructured setup. If any file
in a subfolder is not present, the test will fail.

Currently only the presence of '.check' is tested, not is contents.

Per default the test is skipped, to run it the BUILDCHECK variable
must be define. The simplest way to do this is to run

  'BUILDCHECK= make'

I've also changed the documentation test to only execute when BUILDCHECK
is given and fail properly. Building the isisscripts and getting a huge
list of undocumented functions did not seem to motivate anyone to fix that :)

This is a major change breaking backwards comp: Some functions are renamed to have
a more consistend naming scheme.

The color conversion functions, like hsl2rgb, have this form and should be
read as from space to space (here HSL to RGB). RGB colors are always a
triplet of 8 bit values. All other spaces are a triplet of Double_Type
in the range (0-1). Another useful converter is rgb2hex, here hex represents
either a 24 bit value representing a RGB tuple, or a string representing this
value in hex notation (e.g., 0xff0000 equals "ff0000", and is full red).

There were a number of functions that produced a RGB color (as hex value).
This functions are now all renamed to color_<something>, e.g., color_wavelength
calculates the hex value that corresponds to a given wavelength. In this sence
'color' should always be understood as RGB hex value (notice, 'hex' should be
understood either as hex value (that is, 'color') or as the string (see above)).

The biggest changes concerns the addition of a set of functions which are all
interfaced via 'get_color_palette'. This function allows to choose a palette name
(see 'get_color_palette_names') and how many colors the palette should contain.
See the help of the various palettes for more information.
Via this interface it is also very easy to generate colormaps to be used with
plotting functions.

Example: cmap = get_color_palette("hsluv", 256); % parameters to the generator
% function are passed via qualifiers. Not all palettes are generated, rather
% a static set of specific colors.

Further, I moved the xfig_mix_color to 'plot' as it is an extension of xfig (something
to be fixed in the future). I also deleted the X11 rgb.txt reader functions as this file
is no longer provided on most systems. If we want to provide a list of color names, maybe
we should pack a file with the isisscripts?

Function returned wrong order of results (and index was mixed up)

With this qualifier the normalization value can be retrieve.
This is useful to calculate the flux of components

Merge branch 'lag_energy_spectrum_ingram_refbands' of serpens.sternwarte.uni-erlangen.de:remeis/isisscripts into lag_energy_spectrum_ingram_refbands

* Helpful in combination with get_params if some parameters are tied
  and others are not
* Without keep_null qualifier merge_struct_arrays(get_params())
  filters the NULL entries in the tied field, (i.e., all untied
  parameters are filtered out), which causes problems if one wants to
  filter the struct later.

* Helpful in combination with get_params if some parameters are tied
  and others are not
* Without keep_null qualifier merge_struct_arrays(get_params())
  filters the NULL entries in the tied field, (i.e., all untied
  parameters are filtered out), which causes problems if one wants to
  filter the struct later.

Please make sure there is spacing in the description! Otherwise 'help'
will pick up on the words that are not preceeded by a space.

Often the range was outside of what cerf can handle causing NaNs.
Using the gsl module fixes this.

Hopefully this works, not sure what I was thinking...

The reduced "square" of parameter ranges given by
width is moved such that the center of the smaller
"square" is, if possible, placed at the current
parameter values. If this causes the "square" to
reach outside of the parameter ranges, the maximum
shift still inside is used.

Similar to "sigma" for the gaussian initializer, this
allows to run the chain in a specific range but initialize
it in a smaller range.

Potentially this should be done for all read, open, and
write calls. But for the initial open there are also user
errors that must cause an abort.

This does not really do what we want, but hoefully it stops all jobs

This does not really do what we want, but hoefully it stops all jobs