9. API

9.1. Model

class hazel.model.Model(config=None, working_mode='synthesis', verbose=0, debug=False, rank=0, randomization=None, root='')

Bases: object

Methods

add_chromosphere(atmosphere)

Programmatically add a chromosphere

add_parametric(atmosphere)

Programmatically add a parametric atmosphere

add_photosphere(atmosphere)

Programmatically add a photosphere

add_spectral(spectral)

Programmatically add a spectral region

add_straylight(atmosphere)

Programmatically add a straylight atmosphere

add_topology(atmosphere_order)

Add a new topology

backtracking_parabolic(dchi2, ddchi2[, ...])

Do the backtracking to get an optimal value of lambda in the LM algorithm

compute_chi2([only_chi2, weights])

Compute chi2 for all spectral regions

compute_uncertainty()

Compute the uncertainty in the parameters at the minimum with the current Hessian

find_active_parameters(cycle)

Find all active parameters in all active atmospheres in the current cycle

flatten_parameters_to_reference(cycle)

Flatten all current parameters to the reference atmosphere

init_sir()

Initialize SIR for this synthesis.

init_sir_external()

Initialize SIR for this synthesis

invert([randomize, randomization_ind])

Invert all atmospheres

invert_external(algorithm[, use_jacobian])

Invert all atmospheres

modified_svd_inverse(H[, tol])

Compute the inverse of the Hessian matrix using a modified SVD, by thresholding each subpsace separately

normalize_ff(nodes)

Normalize all filling factors so that they add to one to avoid later problems.

randomize()

Randomize all free parameters to lie uniformly in the interval [-2,2] in the transformed domain

remove_atmosphere(atm)

Remove unused atmospheres

remove_unused_atmosphere()

Remove unused atmospheres

set_new_model(nodes)

Set the nodes of the current model to the values passed on the arguments

set_nlte(option)

Set calculation of Ca II 8542 A to NLTE

set_topologies(topologies)

Set the topologies to be used in the inversion

setup()

Setup the model for synthesis/inversion.

synthesize([perturbation])

Synthesize all atmospheres

synthesize_and_compute_rf([compute_rf, ...])

Compute response functions for all free parameters according to all active_parameters

synthesize_spectral_region(spectral_region)

Synthesize all atmospheres for a single spectral region and normalize to the continuum of the quiet Sun at disk center

use_configuration(config_dict)

Use a configuration file

backtracking_brent

close_output

exit_hazel

open_output

read_observation

write_output

add_chromosphere(atmosphere)

Programmatically add a chromosphere

Parameters:

atmosphere (dict) – Dictionary containing the following data ‘Name’, ‘Spectral region’, ‘Height’, ‘Line’, ‘Wavelength’, ‘Reference atmospheric model’, ‘Ranges’, ‘Nodes’

Return type:

None

add_parametric(atmosphere)

Programmatically add a parametric atmosphere

Parameters:

atmosphere (dict) – Dictionary containing the following data ‘Name’, ‘Spectral region’, ‘Wavelength’, ‘Reference atmospheric model’, ‘Type’, ‘Ranges’, ‘Nodes’

Return type:

None

add_photosphere(atmosphere)

Programmatically add a photosphere

Parameters:

atmosphere (dict) – Dictionary containing the following data ‘Name’, ‘Spectral region’, ‘Height’, ‘Line’, ‘Wavelength’, ‘Reference atmospheric model’, ‘Ranges’, ‘Nodes’

Return type:

None

add_spectral(spectral)

Programmatically add a spectral region

Parameters:

spectral (dict) – Dictionary containing the following data ‘Name’, ‘Wavelength’, ‘Topology’, ‘Weights Stokes’, ‘Wavelength file’, ‘Wavelength weight file’, ‘Observations file’, ‘Mask file’

Return type:

None

add_straylight(atmosphere)

Programmatically add a straylight atmosphere

Parameters:

atmosphere (dict) – Dictionary containing the following data ‘Name’, ‘Spectral region’, ‘Reference atmospheric model’, ‘Ranges’, ‘Nodes’

Return type:

None

add_topology(atmosphere_order)

Add a new topology

Parameters:

topology (str) – Topology

Return type:

None

backtracking_brent(dchi2, ddchi2, maxiter=10, bounds=[-3.0, 3.0], tol=0.01)
backtracking_parabolic(dchi2, ddchi2, direction='down', maxiter=5, lambda_init=0.001, current_chi2=10000000000.0)

Do the backtracking to get an optimal value of lambda in the LM algorithm

Parameters:
  • dchi2 (float) – Gradient of the chi2

  • ddchi2 (float) – Second order derivatives with which the Hessian is computed

  • direction (str, optional) – Direction on which do the backtracking (‘down’/’up’ for decreasing/increasing lambda)

  • maxiter (int) – Maximum number of iterations

  • lambda_init (float) – Initial value of lambda

  • current_chi2 (float) – Current best chi2 to compare with those of the backtracking

Returns:

  • lambda_opt (float) – Optimal value of lambda found. Bracketed value if bracketing has been possible or just the best value otherwise

  • bracketed (bool) – True if the best value has been bracketed

  • best_chi2 (float) – Best value of chi2 found

close_output()
compute_chi2(only_chi2=False, weights=None)

Compute chi2 for all spectral regions

Parameters:
  • obs (float) – Vector of observations

  • only_chi2 (bool) – Control whether the gradient and Hessian is returned

Return type:

None

compute_uncertainty()

Compute the uncertainty in the parameters at the minimum with the current Hessian

Parameters:

None

Return type:

None

exit_hazel()
find_active_parameters(cycle)

Find all active parameters in all active atmospheres in the current cycle

Parameters:

cycle (int) – Cycle to consider

Return type:

None

flatten_parameters_to_reference(cycle)

Flatten all current parameters to the reference atmosphere

Parameters:

cycle (int) – Current cycle

Return type:

None

init_sir()

Initialize SIR for this synthesis. This version does not make use of any external file, which might be not safe when running in MPI mode.

Parameters:

None

Return type:

None

init_sir_external()

Initialize SIR for this synthesis

Parameters:

None

Return type:

None

invert(randomize=False, randomization_ind=None)

Invert all atmospheres

Parameters:

None

Return type:

None

invert_external(algorithm, use_jacobian=False, **kwargs)

Invert all atmospheres

Parameters:

None

Return type:

None

modified_svd_inverse(H, tol=1e-08)

Compute the inverse of the Hessian matrix using a modified SVD, by thresholding each subpsace separately

Parameters:
  • H (float) – Hessian matrix

  • tol (float) – Tolerance for the singular value of each subspace

Return type:

None

normalize_ff(nodes)

Normalize all filling factors so that they add to one to avoid later problems. We use a softmax function to make sure they all add to one and can be unconstrained

ff_i = exp(x_i) / sum(exp(x_i))

Parameters:

None

Return type:

None

open_output()
randomize()

Randomize all free parameters to lie uniformly in the interval [-2,2] in the transformed domain

read_observation()
remove_atmosphere(atm)

Remove unused atmospheres

Parameters:

None

Return type:

None

remove_unused_atmosphere()

Remove unused atmospheres

Parameters:

None

Return type:

None

set_new_model(nodes)

Set the nodes of the current model to the values passed on the arguments

Parameters:

nodes (float) – Array with the new set of nodes

Return type:

None

set_nlte(option)

Set calculation of Ca II 8542 A to NLTE

Parameters:

option (bool) – Set to True to use NLTE, False to use LTE

set_topologies(topologies)

Set the topologies to be used in the inversion

Parameters:

topologies (list) – List of strings with the names of the topologies

Return type:

None

setup()

Setup the model for synthesis/inversion. This setup includes adding the topologies, removing unused atmospheres, reading the number of cycles for the inversion and some sanity checks

Parameters:

None

Return type:

None

synthesize(perturbation=False)

Synthesize all atmospheres

Parameters:

perturbation (bool) – Set to True if you are synthesizing with a perturbation. In this case, the synthesis is saved in spectrum.stokes_perturbed instead of spectrum.stokes

Return type:

None

synthesize_and_compute_rf(compute_rf=False, include_jacobian=False)

Compute response functions for all free parameters according to all active_parameters

Parameters:

compute_rf (bool (optional, default False)) – If True, then compute the response functions. If not, just compute the synthesis.

Return type:

None

synthesize_spectral_region(spectral_region, perturbation=False)

Synthesize all atmospheres for a single spectral region and normalize to the continuum of the quiet Sun at disk center

Parameters:
  • spectral_region (str) – Spectral region to synthesize

  • perturbation (bool) – Set to True if you are synthesizing with a perturbation. In this case, the synthesis is saved in spectrum.stokes_perturbed instead of spectrum.stokes

Return type:

None

use_configuration(config_dict)

Use a configuration file

Parameters:

config_dict (dict) – Dictionary containing all the options from the configuration file previously read

Return type:

None

write_output(randomization=0)

9.2. Photosphere

class hazel.photosphere.SIR_atmosphere(working_mode, name='', root='', verbose=0)

Bases: General_atmosphere

Methods

add_active_line(lines, spectrum, wvl_range, ...)

Add an active lines in this atmosphere

allocate_info_cycles(n_cycles)

Set the appropriate variables to store per-cycle models

get_parameters()

Get the curent parameters as a model

init_reference([check_borders])

Initialize the reference atmosphere to the values of the parameters, doing the inverse transformation if in inversion mode

interpolate_nodes(log_tau, reference, nodes, ...)

Generate a model atmosphere by interpolating the defined nodes.

interpolate_nodes_rf(log_tau, reference, ...)

Generate a model atmosphere by interpolating the defined nodes.

list_lines()

List the lines available in SIR for synthesis

load_reference_model(model_file, verbose)

Load a reference model or a model for every pixel for synthesis/inversion

model_to_nodes()

Transform from model to nodes

nodes_to_model()

Transform from nodes to model

reset_reference()

Reset reference model to the original reference loaded from file

set_parameters(model, ff, vmac)

Set the parameters of the current model to those passed as argument

set_reference([cycle])

Set reference model to that of the current parameters

synthesize(stokes_in[, returnRF, nlte])

Carry out the synthesis and returns the Stokes parameters and the response functions to all physical variables at all depths

to_physical()

Transform the atmospheric parameters from transformed domain to physical domain given the ranges.

to_transformed()

Transform the atmospheric parameters from transformed domain to physical domain given the ranges.

load_nlte_model

print_parameters

add_active_line(lines, spectrum, wvl_range, verbose)

Add an active lines in this atmosphere

Parameters:
  • lines (str) – Line to activate

  • spectrum (Spectrum) – Spectrum object

  • wvl_range (float) – Vector containing wavelength range over which to synthesize this line

Return type:

None

get_parameters()

Get the curent parameters as a model

Parameters:

None

Returns:

model

Return type:

a 6xN photspheric model

interpolate_nodes(log_tau, reference, nodes, nodes_location)

Generate a model atmosphere by interpolating the defined nodes. The interpolation order depends on the number of nodes.

Parameters:
  • log_tau (float) – Vector of log optical depth at 500 nm

  • reference (float) – Vector with the reference atmosphere to which the nodes are added

  • nodes (float) – List with the position of the nodes

Returns:

Vector with the interpolated atmosphere

Return type:

real

interpolate_nodes_rf(log_tau, reference, nodes, lower, upper)

Generate a model atmosphere by interpolating the defined nodes. The interpolation order depends on the number of nodes.

Parameters:
  • log_tau (float) – Vector of log optical depth at 500 nm

  • reference (float) – Vector with the reference atmosphere to which the nodes are added

  • nodes (float) – List with the position of the nodes

Returns:

Vector with the interpolated atmosphere

Return type:

real

list_lines()

List the lines available in SIR for synthesis

load_nlte_model(verbose)
load_reference_model(model_file, verbose)

Load a reference model or a model for every pixel for synthesis/inversion

Parameters:
  • model_file (str) – String with the name of the file. Extensions can currently be “1d” or “h5”

  • verbose (bool) – Verbosity

Return type:

None

model_to_nodes()

Transform from model to nodes

Parameters:

None

Return type:

None

nodes_to_model()

Transform from nodes to model

Parameters:

None

Return type:

None

print_parameters(first=False, error=False)
set_parameters(model, ff, vmac)

Set the parameters of the current model to those passed as argument

Parameters:
  • model_in (float) – Array with the model

  • ff (float) – Value of the filling factor

  • vmac (float) – Value of the macroturbulent velocity

Return type:

None

synthesize(stokes_in, returnRF=False, nlte=False)

Carry out the synthesis and returns the Stokes parameters and the response functions to all physical variables at all depths

Parameters:
  • stokes_in (float) – An array of size [4 x nLambda] with the input Stokes parameter. It is irrelevant in this case because we assume that all SIR atmospheres have the Planck function as boundary.

  • returnRF (bool, optional) – Return response functions

Returns:

  • stokes (float) –

    Stokes parameters, with the first index containing the wavelength displacement and the remaining

    containing I, Q, U and V. Size (5,nLambda)

  • rf (float (optional)) –

    Response functions to T, Pe, vmic, B, v, theta, phi, all of size (4,nLambda,nDepth), plus the RF to macroturbulence of size (4,nLambda)

    It is not returned if returnRF=False

9.3. Chromosphere

class hazel.chromosphere.Hazel_atmosphere(working_mode, name='')

Bases: General_atmosphere

Methods

add_active_line(line, spectrum, wvl_range)

Add an active lines in this atmosphere

allocate_info_cycles(n_cycles)

Set the appropriate variables to store per-cycle models

init_reference([check_borders])

Initialize the reference atmosphere to the values of the parameters, doing the inverse transformation if in inversion mode

load_reference_model(model_file, verbose)

Load a reference model or a model for every pixel for synthesis/inversion

nodes_to_model()

Transform from nodes to model

reset_reference()

Reset reference model to the original reference loaded from file

set_parameters(pars, ff)

Set the parameters of this model chromosphere

set_reference([cycle])

Set reference model to that of the current parameters

synthesize([stokes, returnRF, nlte])

Carry out the synthesis and returns the Stokes parameters

to_physical()

Transform the atmospheric parameters from transformed domain to physical domain given the ranges.

to_transformed()

Transform the atmospheric parameters from transformed domain to physical domain given the ranges.

print_parameters

select_coordinate_system

add_active_line(line, spectrum, wvl_range)

Add an active lines in this atmosphere

Parameters:
  • lines (str) – Line to activate: [‘10830’,’5876’]

  • spectrum (Spectrum) – Spectrum object

  • wvl_range (float) – Vector containing wavelength range over which to synthesize this line

Return type:

None

load_reference_model(model_file, verbose)

Load a reference model or a model for every pixel for synthesis/inversion

Parameters:
  • model_file (str) – String with the name of the file. Extensions can currently be “1d” or “h5”

  • verbose (bool) – verbosity flag

Return type:

None

nodes_to_model()

Transform from nodes to model

Parameters:

None

Return type:

None

print_parameters(first=False, error=False)
select_coordinate_system()
set_parameters(pars, ff)

Set the parameters of this model chromosphere

Parameters:
  • pars (list of float) – This list contains the following parameters in order: Bx, By, Bz, tau, v, delta, beta, a

  • ff (float) – Filling factor

Return type:

None

synthesize(stokes=None, returnRF=False, nlte=None)

Carry out the synthesis and returns the Stokes parameters

Parameters:

stokes (float) – An array of size [4 x nLambda] with the input Stokes parameter.

Returns:

stokes

Stokes parameters, with the first index containing the wavelength displacement and the remaining

containing I, Q, U and V. Size (4,nLambda)

Return type:

float

9.4. Parametric

class hazel.parametric.Parametric_atmosphere(working_mode, name='')

Bases: General_atmosphere

Methods

add_active_line(spectrum, wvl_range)

Add an active lines in this atmosphere

allocate_info_cycles(n_cycles)

Set the appropriate variables to store per-cycle models

init_reference([check_borders])

Initialize the reference atmosphere to the values of the parameters, doing the inverse transformation if in inversion mode

load_reference_model(model_file, verbose)

Load a reference model or a model for every pixel for synthesis/inversion

nodes_to_model()

Transform from nodes to model

reset_reference()

Reset reference model to the original reference loaded from file

set_parameters(pars, ff)

Set parameters of the model

set_reference([cycle])

Set reference model to that of the current parameters

synthesize([stokes, nlte])

Carry out the synthesis and returns the Stokes parameters

to_physical()

Transform the atmospheric parameters from transformed domain to physical domain given the ranges.

to_transformed()

Transform the atmospheric parameters from transformed domain to physical domain given the ranges.

print_parameters

add_active_line(spectrum, wvl_range)

Add an active lines in this atmosphere

Parameters:

None

Return type:

None

load_reference_model(model_file, verbose)

Load a reference model or a model for every pixel for synthesis/inversion

Parameters:
  • model_file (str) – String with the name of the file. Extensions can currently be “1d” or “h5”

  • verbose (bool) – verbosity flag

Return type:

None

nodes_to_model()

Transform from nodes to model

Parameters:

None

Return type:

None

print_parameters(first=False, error=False)
set_parameters(pars, ff)

Set parameters of the model

Parameters:
  • pars (list of floats) – Values of the parameters: lambda0, sigma, depth, a

  • ff (float) – Filling factor

Return type:

None

synthesize(stokes=None, nlte=None)

Carry out the synthesis and returns the Stokes parameters

Parameters:

stokes (float) – An array of size [4 x nLambda] with the input Stokes parameter.

Returns:

stokes

Stokes parameters, with the first index containing the wavelength displacement and the remaining

containing I, Q, U and V. Size (4,nLambda)

Return type:

float

9.5. Multiprocess

class hazel.multiprocess.Iterator(use_mpi=False)

Bases: object

Methods

mpi_parent_work([start, end])

MPI parent work

mpi_workers_work()

MPI workers work

nonmpi_work([start, end])

Do the synthesis/inversion for all pixels in the models

run_all_pixels([start, end])

Run synthesis/inversion for all pixels

cleanup

get_rank

use_model

cleanup(*args)
get_rank(n_workers=0)
mpi_parent_work(start=0, end=None)

MPI parent work

Parameters:
  • start (int) – Initial pixel, by default 0

  • end (int) – Final pixel, by default the number of pixels in the observations

Return type:

None

mpi_workers_work()

MPI workers work

Parameters:

None

Return type:

None

nonmpi_work(start=0, end=None)

Do the synthesis/inversion for all pixels in the models

Parameters:

model (model) – Model to be synthesized

Return type:

None

run_all_pixels(start=0, end=None)

Run synthesis/inversion for all pixels

Parameters:
  • start (int, optional) – Initial pixel, by default 0

  • end (int) – Final pixel, by default the number of pixels in the observations

use_model(model=None)
class hazel.multiprocess.tags(value)

Bases: IntEnum

An enumeration.

DONE = 1
DONOTHING = 4
EXIT = 2
READY = 0
START = 3

9.6. Stray-light

class hazel.stray.Straylight_atmosphere(working_mode, name='')

Bases: General_atmosphere

Methods

add_active_line(spectrum, wvl_range)

Add an active lines in this atmosphere

allocate_info_cycles(n_cycles)

Set the appropriate variables to store per-cycle models

init_reference([check_borders])

Initialize the reference atmosphere to the values of the parameters, doing the inverse transformation if in inversion mode

load_reference_model(model_file, verbose)

Load a reference model or a model for every pixel for synthesis/inversion

nodes_to_model()

Transform from nodes to model

reset_reference()

Reset reference model to the original reference loaded from file

set_reference([cycle])

Set reference model to that of the current parameters

set_straylight(stokes)

Load a reference model or a model for every pixel for synthesis/inversion

synthesize([nlte])

Carry out the synthesis and returns the Stokes parameters

to_physical()

Transform the atmospheric parameters from transformed domain to physical domain given the ranges.

to_transformed()

Transform the atmospheric parameters from transformed domain to physical domain given the ranges.

print_parameters

set_parameters

add_active_line(spectrum, wvl_range)

Add an active lines in this atmosphere

Parameters:

None

Return type:

None

load_reference_model(model_file, verbose)

Load a reference model or a model for every pixel for synthesis/inversion

Parameters:
  • model_file (str) – String with the name of the file. Extensions can currently be “1d” or “h5”

  • verbose (bool) – verbosity flag

Return type:

None

nodes_to_model()

Transform from nodes to model

Parameters:

None

Return type:

None

print_parameters(first=False, error=False)
set_parameters(pars, ff)
set_straylight(stokes)

Load a reference model or a model for every pixel for synthesis/inversion

Parameters:
  • model_file (str) – String with the name of the file. Extensions can currently be “1d” or “h5”

  • verbose (bool) – verbosity flag

Return type:

None

synthesize(nlte=None)

Carry out the synthesis and returns the Stokes parameters

Parameters:

None

Returns:

stokes

Stokes parameters, with the first index containing the wavelength displacement and the remaining

containing I, Q, U and V. Size (4,nLambda)

Return type:

float

9.7. Tools

class hazel.tools.File_chromosphere(file=None, mode='single')

Bases: object

Class that defines a model atmosphere and can be used to easily save observations

Methods

save(file[, default])

Save the curent observation

set_default([n_pixel, default])

Set the atmosphere to one of the default ones available in the code

set_size([n_pixel])

Set the number of pixels of the current atmosphere

save(file, default=None)

Save the curent observation

Parameters:

file (str) – Name of the output files. Extensions will be added to it

Return type:

None

set_default(n_pixel=1, default='disk')

Set the atmosphere to one of the default ones available in the code

Parameters:
  • n_pixel (int (optional, equal to 1 as default)) – Number of pixels of the output

  • default (str ('disk' -> on-disk observations, 'offlimb' -> off-limb observations))

Return type:

None

set_size(n_pixel=1)

Set the number of pixels of the current atmosphere

Parameters:

n_pixel (int (optional, equal to 1 as default)) – Number of pixels of the output

Return type:

None

class hazel.tools.File_observation(file=None, mode='single')

Bases: object

Class that defines an observation. This can be used to easily save observations in the appropriate format

Methods

save(file)

Save the curent observation

set_size(n_lambda[, n_pixel])

Set the number of wavelengths and number of pixels of the current observation

save(file)

Save the curent observation

Parameters:

file (str) – Name of the output files. Extensions will be added to it

Return type:

None

set_size(n_lambda, n_pixel=1)

Set the number of wavelengths and number of pixels of the current observation

Parameters:
  • n_lambda (int) – Number of wavelength points

  • n_pixel (int (optional, equal to 1 as default)) – Number of pixels of the output

Return type:

None

class hazel.tools.File_photosphere(file=None, mode='single')

Bases: object

Class that defines a model photosphere and can be used to easily save observations

Methods

save(file[, default])

Save the curent model

set_default([n_pixel, default])

Set the atmosphere to one of the default ones available in the code

set_size(nz[, n_pixel])

Set the number of depth points and number of pixels of the current atmosphere

list_models

list_models()
save(file, default=None)

Save the curent model

Parameters:

file (str) – Name of the output files. Extensions will be added to it

Return type:

None

set_default(n_pixel=1, default='hsra')

Set the atmosphere to one of the default ones available in the code

Parameters:
  • n_pixel (int (optional, equal to 1 as default)) – Number of pixels of the output

  • default (str ('hsra' -> Harvard-Smithsonian Reference Atmosphere))

Return type:

None

set_size(nz, n_pixel=1)

Set the number of depth points and number of pixels of the current atmosphere

Parameters:
  • nz (int) – Number of depth points of the atmosphere

  • n_pixel (int (optional, equal to 1 as default)) – Number of pixels of the output

Return type:

None