import copy
import numpy as np
import lenstronomy.Util.class_creator as class_creator
from lenstronomy.Plots.model_band_plot import ModelBandPlot
from lenstronomy.Analysis.image_reconstruction import check_solver_error
__all__ = ["ModelPlot"]
[docs]
class ModelPlot(object):
"""Class that manages the summary plots of a lens model The class uses the same
conventions as being used in the FittingSequence and interfaces with the ImSim
module.
The linear inversion is re-done given the likelihood settings in the init of this
class (make sure this is the same as you perform the FittingSequence) to make sure
the linear amplitude parameters are computed as they are not part of the output of
the FittingSequence results.
"""
[docs]
def __init__(
self,
multi_band_list,
kwargs_model,
kwargs_params,
image_likelihood_mask_list=None,
bands_compute=None,
multi_band_type="multi-linear",
source_marg=False,
linear_prior=None,
arrow_size=0.02,
cmap_string="gist_heat",
fast_caustic=True,
linear_solver=True,
):
"""
:param multi_band_list: list of [[kwargs_data, kwargs_psf, kwargs_numerics], [], ..]
:param multi_band_type: string, option when having multiple imaging data sets modelled simultaneously.
Options are:
- 'multi-linear': linear amplitudes are inferred on single data set
- 'linear-joint': linear amplitudes ae jointly inferred
- 'single-band': single band
:param kwargs_model: model keyword arguments
:param bands_compute: (optional), bool list to indicate which band to be included in the modeling
:param image_likelihood_mask_list: list of image likelihood mask
(same size as image_data with 1 indicating being evaluated and 0 being left out)
:param kwargs_params: keyword arguments of 'kwargs_lens', 'kwargs_source' etc. as coming as kwargs_result from
FittingSequence class
:param source_marg:
:param linear_prior:
:param arrow_size:
:param cmap_string:
:param fast_caustic: boolean; if True, uses fast (but less accurate) caustic calculation method
:param linear_solver: bool, if True (default) fixes the linear amplitude parameters 'amp' (avoid sampling) such
that they get overwritten by the linear solver solution.
"""
if bands_compute is None:
bands_compute = [True] * len(multi_band_list)
if multi_band_type == "single-band":
multi_band_type = "multi-linear" # this makes sure that the linear inversion outputs are coming in a list
self._imageModel = class_creator.create_im_sim(
multi_band_list,
multi_band_type,
kwargs_model,
bands_compute=bands_compute,
linear_solver=linear_solver,
image_likelihood_mask_list=image_likelihood_mask_list,
)
kwargs_params_copy = copy.deepcopy(kwargs_params)
kwargs_params_copy.pop("kwargs_tracer_source", None)
model, error_map, cov_param, param = self._imageModel.image_linear_solve(
inv_bool=True, **kwargs_params_copy
)
if linear_solver is False:
if len(multi_band_list) > 1:
raise ValueError(
"plotting the solution without the linear solver currently only works with one band."
)
im_sim = class_creator.create_im_sim(
multi_band_list,
"single-band",
kwargs_model,
bands_compute=bands_compute,
linear_solver=linear_solver,
image_likelihood_mask_list=image_likelihood_mask_list,
)
# overwrite model with initial input without linear solver applied
model[0] = im_sim.image(**kwargs_params_copy)
# retrieve amplitude parameters directly from kwargs_list
param[0] = im_sim.linear_param_from_kwargs(
kwargs_params["kwargs_source"],
kwargs_params["kwargs_lens_light"],
kwargs_params["kwargs_ps"],
)
else:
# overwrite the keyword list with the linear solved 'amp' values
for key in kwargs_params_copy.keys():
kwargs_params[key] = kwargs_params_copy[key]
check_solver_error(param)
log_l, _ = self._imageModel.likelihood_data_given_model(
source_marg=source_marg, linear_prior=linear_prior, **kwargs_params_copy
)
n_data = self._imageModel.num_data_evaluate
if n_data > 0:
print(
log_l * 2 / n_data,
"reduced X^2 of all evaluated imaging data combined "
"(without degrees of freedom subtracted).",
)
self._band_plot_list = []
self._index_list = []
index = 0
for i in range(len(multi_band_list)):
if bands_compute[i] is True:
if multi_band_type == "joint-linear":
param_i = param
cov_param_i = cov_param
else:
param_i = param[index]
cov_param_i = cov_param[index]
bandplot = ModelBandPlot(
multi_band_list,
kwargs_model,
model[index],
error_map[index],
cov_param_i,
param_i,
copy.deepcopy(kwargs_params),
likelihood_mask_list=image_likelihood_mask_list,
band_index=i,
arrow_size=arrow_size,
cmap_string=cmap_string,
fast_caustic=fast_caustic,
)
self._band_plot_list.append(bandplot)
self._index_list.append(index)
else:
self._index_list.append(-1)
index += 1
def _select_band(self, band_index):
"""
:param band_index: index of imaging band to be plotted
:return: bandplot() instance of selected band, raises when band is not computed
"""
i = self._index_list[band_index]
if i == -1:
raise ValueError("band %s is not computed or out of range." % band_index)
i = int(i)
return self._band_plot_list[i]
[docs]
def reconstruction_all_bands(self, **kwargs):
"""
:param kwargs: arguments of plotting
:return: 3 x n_data plot with data, model, reduced residual plots of all the images/bands that are being modeled
"""
n_bands = len(self._band_plot_list)
import matplotlib.pyplot as plt
f, axes = plt.subplots(n_bands, 3, figsize=(12, 4 * n_bands))
if n_bands == 1: # make sure axis can be called as 2d array
_axes = np.empty((1, 3), dtype=object)
_axes[:] = axes
axes = _axes
i = 0
for band_index in self._index_list:
if band_index >= 0:
axes[i, 0].set_title("image " + str(band_index))
self.data_plot(ax=axes[i, 0], band_index=band_index, **kwargs)
self.model_plot(
ax=axes[i, 1], image_names=True, band_index=band_index, **kwargs
)
self.normalized_residual_plot(
ax=axes[i, 2], v_min=-6, v_max=6, band_index=band_index, **kwargs
)
i += 1
return f, axes
[docs]
def data_plot(self, band_index=0, **kwargs):
"""Illustrates data.
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.data_plot(**kwargs)
[docs]
def model_plot(self, band_index=0, **kwargs):
"""Illustrates model.
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.model_plot(**kwargs)
[docs]
def convergence_plot(self, band_index=0, **kwargs):
"""Illustrates lensing convergence in data frame.
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.convergence_plot(**kwargs)
[docs]
def substructure_plot(self, band_index=0, **kwargs):
"""Illustrates substructure in the lens system.
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.substructure_plot(**kwargs)
[docs]
def normalized_residual_plot(self, band_index=0, **kwargs):
"""Illustrates normalized residuals between data and model fit.
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.normalized_residual_plot(**kwargs)
[docs]
def absolute_residual_plot(self, band_index=0, **kwargs):
"""Illustrates absolute residuals between data and model fit.
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.absolute_residual_plot(**kwargs)
[docs]
def source_plot(self, band_index=0, **kwargs):
"""Illustrates reconstructed source (de-lensed de-convolved)
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.source_plot(**kwargs)
[docs]
def error_map_source_plot(self, band_index=0, **kwargs):
"""Illustrates surface brightness variance in the reconstruction in the source
plane.
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.error_map_source_plot(**kwargs)
[docs]
def magnification_plot(self, band_index=0, **kwargs):
"""Illustrates lensing magnification in the field of view of the data frame.
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.magnification_plot(**kwargs)
[docs]
def deflection_plot(self, band_index=0, **kwargs):
"""Illustrates lensing deflections on the field of view of the data frame.
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.deflection_plot(**kwargs)
[docs]
def decomposition_plot(self, band_index=0, **kwargs):
"""Illustrates decomposition of model components.
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.decomposition_plot(**kwargs)
[docs]
def subtract_from_data_plot(self, band_index=0, **kwargs):
"""Subtracts individual model components from the data.
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.subtract_from_data_plot(**kwargs)
[docs]
def plot_main(self, band_index=0, **kwargs):
"""Plot a set of 'main' modelling diagnostics.
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.plot_main(**kwargs)
[docs]
def plot_separate(self, band_index=0):
"""Plot a set of 'main' modelling diagnostics.
:param band_index: index of band
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.plot_separate()
[docs]
def plot_subtract_from_data_all(self, band_index=0):
"""Plot a set of 'main' modelling diagnostics.
:param band_index: index of band
:return: plot instance
"""
plot_band = self._select_band(band_index)
return plot_band.plot_subtract_from_data_all()
[docs]
def plot_extinction_map(self, band_index=0, **kwargs):
"""
:param band_index: index of band
:param kwargs: arguments of plotting
:return: plot instance of differential extinction map
"""
plot_band = self._select_band(band_index)
return plot_band.plot_extinction_map(**kwargs)
[docs]
def source(self, band_index=0, **kwargs):
"""
:param band_index: index of band
:param kwargs: keyword arguments accessible in model_band_plot.source()
:return: 2d array of source surface brightness
"""
plot_band = self._select_band(band_index)
return plot_band.source(**kwargs)