GitLab

To avoid problems that might occur when, for instance, using an integer
instead of a float, we now cast the module parameters to their expected
type. This will also have the advantage to raise an error when a
parameter can't be interpreted with its expected type.

Also, use integer for age parameters as we are now using Myr as unit.

Thanks to Médéric Boquien.

It's a dictionnary to hold the spectral lines associated with the SED.
We use a disctionnary to be able to differentiate several sets of lines
(e.g. lines from old and young populations).

TODO: Complete the code to add the line contributions to the integrated flux.

Rename lumin_contributions to luminosities and use a 2 axis numpy array
instead of a list of arrays. This should accelerate some operations.

The star formation history is now stored in its own property.
TODO: maybe automatically add age and instantaneous SFR to the info dictionnary when setting the SFH.

To perform the convolution with the SFH, the rate of ionizing  photons
must not be stored as the log of its value.

Using Gyr leads to some errors when comparing times (because they are no
really equals when they shoud). So we switch to the use of Myr as it is
the precision for our times.

Maybe later, we'll go further and use integers for these values.

Rename it to dustatt_calzleit in preparation of a power law module.

* Multiplicate the Calzetti-Leitherer continuum by a power law.
* Add a UV bump.
* The luminosity under 91.2 nm to 0.
* Extrapolate the continuum over 2200 nm until attenuation reach 0.

By default, an exception is raised when adding an existing key to the
info dictionary. Make it possible to bypass this, for instance to add
age if we want to combine various SFH.

If we redshift a SED using a module, we must not redshift the wavelength
to compute the λ, Fν array tuple. Thus we make it optional and not
default.

Convert the warehouse a its own module and use submodules for each
"back-end".
Also, make the back-end configurable with a parameter of the psum
analysis module.

It's better (for now) to use only get_from_cache and get None when the
object is not cached.

As there is a specific module to redshift the spectrum, do not
automatically apply the redshift when computing a flux.

Simplify some module names:
dh2002_ir > dh2002
sfh_2exp > sfh2exp

Rename m2005_sfh to m2005 as it does not deal with the star formation
history.

The Marasont 2005 module is modified to use the output of a star
formation history module. The SFH is now a complete SFR vs time tuple
(taken from the SED information dictionary) and the age of the galaxy is
defined by this SFH (the maximum time).
The Maraston 2005 data class is also changed to make the convolution in
such SFH.

s/[Pp]arametre/[Pp]arameter/g
s/[Dd]ictionnary/[Dd]ictionary/g

A list is added to the SED object to store the key names from the info
dictionary whose values are proportional to the galaxy mass.
This is used by the statistical module to multiplicate the value by the
galaxy mass (correction factor) in the analysis.

Dictionary associating to each "output" parameter its description.

Suppress the 'unit' part of the tuple associated to each parametre
in the parametre_list dictionary of the SED modules. It's not
useful. It's better to indicate the unit in the description text.

A set must be used to compare the two lists because they may have
different orders.

Also, add the list of missing parametres and of unexpected parametres to
the error message.

The effect of the module is _not_ negative!

The modules can now be named with any prefix using a dot (e.g.
dh2002_ir.1). That means that a 'pristine' module can't have a dot in
it's name.

The add_component SED method was removed. To add components to a sed,
add_info, add_module (NEW) and add_contribution (NEW) must be used.

The modules now add more information in the sed info dictionary, even
the parametres that can be used for the statistical analysis. In
particular, mh2005_sfh add the SFR and the average SFR.

The sfr variable was renamed to sfh.