Release Notes
[v0.4.3] - 2025-12-18
[v0.4.3] - 2025-12-18
Added the LGPL-3.0+ licence
Updated GitHub workflows.
Updated Example notebooks.
Model and ModelCollection implementation
Redid how data is meant to be saved to the database file is handled. The type of the data
is now preserved. I.e. a list will always be a list etc. The types that can be saved are
int, float, string, list, tuple, and ndarray.
Updated how models and references are stored in the model collection. They are now stored in lists instead of dictionaries. This change is only internal and normal user experience remains the same.
If a model contains a refid_<name> hdf5 attribute when created it will automatically create
a normal attribute called ref_<name> that points towards that model object. An error raised
if a reference models with that name does not exist.
Models are not tied to a model collection. One model can belong to several different model collections.
Adding a model to a collection will also add any reference models that are linked through
normal attributes called ref_<name>.
Updated how get_mask works. Any attribute associated with the Model can now be accessed using
the . method, the same as for where.
The where method now return a model collection containing the original model objects rather than copies.
Renamed ModelTemplate to ModelBase.
Added FILE_TYPE, VERSION, and CREATED attributes to the saved hdf5 file.
CCSNe Models
Added the masscoord_mass attribute to the mandatory attributes of the CCSNe models.
Added a zone attribute to CCSNe models with an onion shell structure. This is a ndarray with the name
of the shell for each mass coordinate. Removed the ability to specify individual shells
in get_mask. Instead use .zone == <name> instead.
Added warning if a model contains data points from the Mrem zone. And ability to skip these points.
Removed the onion_lbound attribute from the CCSNe models. The reason is because the index was hardcoded it makes
excluding datapoints tricky. This can be avoided by calculating the bounds using the zone attribute instead
Plotting
Added a new plot type - A traditional histogram. The new plot type can be accessed using
the hist function. Additionally, it is now possible to automatically add a histogram to
of the data show in plot using the hist, yhist and xhist keywords. This replaces
the old hist_host/mhist functions to plot circular histograms.
get_data now return the datapoints mapped to the Model object rather than the model name.
Added add_weights and add_weights_ccsne functions that will calculate and add the weight
for each datapoint in a data returned from get_data.
Split the main plotting functions into two steps.
Added test based on Tutorial 3. Does not verify output only that the commands run without error. Added check that the tutorial notebooks run without error.
Utils
Updated how the default kwargs works. You can now pass the whole kwargs dictionary to another function so that removed/added items are synced between the functions.
Shortcuts can now be "inherited".