Model

Model#

class jetset.base_model.Model(name='no_name', nu_size=200, model_type='base_model', scale='lin-lin', cosmo=None, nu_min=None, nu_max=None)[source]#

Bases: object

Base class for analytical and numerical SED models.

Attributes Summary

version

Return the package version used to create this model.

Methods Summary

add_user_par(name, val[, units, val_min, ...])

Add a user-defined parameter to the model.

build_table([restframe])

Build SED table for the current model state.

clone()

Return a deep clone of the model via in-memory serialization.

dep_func_get_default_args(par_func)

Validate dependency-function arguments against model parameters.

eval([fill_SED, nu, get_model, loglog, label])

Evaluate model fluxes.

get_par_by_name(par_name)

Return parameter by name.

get_par_by_type(par_type)

Return first parameter matching a parameter type.

get_residuals(data[, log_log, filter_UL])

Compute residuals between observed data and model prediction.

lin_func(lin_nu)

Return model values in linear space.

load_model(file_name_or_obj[, from_string])

Load a serialized model.

log_func(log_nu)

Return model values in log10 space.

make_dependent_par(par, depends_on, par_expr)

Make dependent par.

plot_model([plot_obj, clean, sed_data, ...])

Plot model SED and optional spectral components.

save_model(file_name)

Serialize the model to disk.

sed_table([restframe])

Return SED table, evaluating the model if needed.

set_fit_range([down_tol, up_tol])

Set fit ranges for numeric parameters.

set_nu_grid([nu_min, nu_max, nu_size])

Set model frequency-grid settings.

set_par(par_name, val)

Set a parameter value by name.

show_best_fit_pars()

Display best-fit parameter values.

show_model()

Print model summary and parameters.

show_pars([sort_key])

Display model parameters.

Attributes Documentation

version#

Return the package version used to create this model.

Returns:

Version string.

Return type:

str

Methods Documentation

add_user_par(name, val, units='', val_min=None, val_max=None)[source]#

Add a user-defined parameter to the model.

Parameters:

  • name (str) – Parameter name.

  • val (float) – Initial value.

  • units (str, optional) – Parameter units label.

  • val_min (float, optional) – Lower physical bound.

  • val_max (float, optional) – Upper physical bound.

build_table(restframe='obs')[source]#

Build SED table for the current model state.

Parameters:

restframe ({'obs', 'src'}, optional) – Frame used to build frequency/flux columns.

clone()[source]#

Return a deep clone of the model via in-memory serialization.

Returns:

Cloned model instance.

Return type:

Model

dep_func_get_default_args(par_func)[source]#

Validate dependency-function arguments against model parameters.

Parameters:

par_func (callable) – Function used as dependency expression.

Returns:

Ordered list of parameter names accepted by par_func.

Return type:

list of str

Raises:

RuntimeError – If par_func uses argument names not present in the model.

eval(fill_SED=True, nu=None, get_model=False, loglog=False, label=None, **kwargs)[source]#

Evaluate model fluxes.

Parameters:

  • fill_SED (bool, optional) – If True, update the model SED object after evaluation.

  • nu (array-like or float, optional) – Frequency grid in Hz (or log10(Hz) if loglog is True). If omitted, the internal nu_min/nu_max/nu_size grid is used.

  • get_model (bool, optional) – If True, return evaluated values.

  • loglog (bool, optional) – If True, treat input/output frequency and model values in log10 space.

  • label (str, optional) – Reserved for subclasses/plotting integrations.

  • **kwargs – Extra keyword arguments accepted for subclass compatibility.

Returns:

Evaluated model array when get_model is True, otherwise None.

Return type:

ndarray or None

get_par_by_name(par_name)[source]#

Return parameter by name.

Parameters:

par_name (str) – Parameter name.

Returns:

Matching parameter, if found.

Return type:

ModelParameter or None

get_par_by_type(par_type)[source]#

Return first parameter matching a parameter type.

Parameters:

par_type (str) – Parameter type label.

Returns:

Matching parameter, if found.

Return type:

ModelParameter or None

get_residuals(data, log_log=False, filter_UL=True)[source]#

Compute residuals between observed data and model prediction.

Parameters:

  • data (ObsData or table-like) – Input data table containing nu_data, nuFnu_data, dnuFnu_data and UL columns.

  • log_log (bool, optional) – If True, return frequency axis in log10(Hz).

  • filter_UL (bool, optional) – If True, exclude upper-limit points.

Returns:

(nu_axis, residuals).

Return type:

tuple of ndarray

lin_func(lin_nu)[source]#

Return model values in linear space.

Parameters:

lin_nu (ndarray) – Frequency array in Hz.

Returns:

Model values in linear nuFnu units.

Return type:

ndarray

classmethod load_model(file_name_or_obj, from_string=False)[source]#

Load a serialized model.

Parameters:

  • file_name_or_obj (str or bytes or file-like) – Serialized model source accepted by the pickle loader.

  • from_string (bool, optional) – If True, interpret file_name_or_obj as in-memory content.

Returns:

Reconstructed and evaluated model instance.

Return type:

Model

Raises:

RuntimeError – If deserialization fails or object type is not valid.

log_func(log_nu)[source]#

Return model values in log10 space.

Parameters:

log_nu (ndarray) – Frequency array in log10(Hz).

Returns:

Model values in log10 space.

Return type:

ndarray

make_dependent_par(par, depends_on, par_expr, verbose=True, set_par_expr_source_code=True, master_pars=None)[source]#

Make dependent par.

Parameters:

  • par (object) – Parameter object or parameter name.

  • depends_on (object) – Names of master parameters used by the dependency.

  • par_expr (object) – Expression defining a dependent parameter.

  • verbose (bool, optional) – If True, print additional information.

  • set_par_expr_source_code (bool, optional) – If True, store source code for the dependency expression.

  • master_pars (object, optional) – Master-parameter objects used by the dependency.

plot_model(plot_obj=None, clean=False, sed_data=None, frame='obs', skip_components=False, label=None, line_style='-', density=False)[source]#

Plot model SED and optional spectral components.

Parameters:

  • plot_obj (PlotSED, optional) – Existing plotting object. If omitted, a new one is created.

  • clean (bool, optional) – If True, clear existing model lines from plot_obj.

  • sed_data (ObsData, optional) – Optional observed data used by the plotting helper.

  • frame ({'obs', 'src'}, optional) – Output frame for plotting.

  • skip_components (bool, optional) – If True, do not plot individual spectral components.

  • label (str, optional) – Label for the model curve.

  • line_style (str, optional) – Matplotlib line style.

  • density (bool, optional) – If True, use density representation in plotting helper.

Returns:

Plot object with model curves added.

Return type:

PlotSED

save_model(file_name)[source]#

Serialize the model to disk.

Parameters:

file_name (str) – Output pickle file path.

sed_table(restframe='obs')[source]#

Return SED table, evaluating the model if needed.

Parameters:

restframe ({'obs', 'src'}, optional) – Frame used to build frequency/flux columns.

Returns:

SED table for the current model state.

Return type:

astropy.table.Table or None

set_fit_range(down_tol=0.1, up_tol=100)[source]#

Set fit ranges for numeric parameters.

Parameters:

  • down_tol (float, optional) – Multiplicative factor for lower fit bound.

  • up_tol (float, optional) – Relative factor for upper fit bound.

set_nu_grid(nu_min=None, nu_max=None, nu_size=None)[source]#

Set model frequency-grid settings.

Parameters:

  • nu_min (float, optional) – Minimum frequency in Hz.

  • nu_max (float, optional) – Maximum frequency in Hz.

  • nu_size (int, optional) – Number of samples in the grid.

set_par(par_name, val)[source]#

Set a parameter value by name.

Parameters:

  • par_name (str) – Parameter name.

  • val (float or int) – New parameter value.

show_best_fit_pars()[source]#

Display best-fit parameter values.

show_model()[source]#

Print model summary and parameters.

show_pars(sort_key='par type')[source]#

Display model parameters.

Parameters:

sort_key (str, optional) – Column used to sort displayed parameters.

Returns:

Output returned by ModelParameterArray.show_pars.

Return type:

object