baobab.bnn_priors package

baobab.bnn_priors.base_bnn_prior module

class baobab.bnn_priors.base_bnn_prior.BaseBNNPrior(bnn_omega, components)

Bases: abc.ABC

Abstract base class equipped with PDF evaluation and sampling utility functions for various lens/source macromodels

eval_param_pdf(eval_at, hyperparams)

Assigns and evaluates the PDF

sample()

Gets kwargs of sampled parameters to be passed to lenstronomy

Overridden by subclasses.

sample_param(hyperparams)

Assigns a sampling distribution

set_comps_qphi_to_e1e2()

set_params_list(params_to_exclude)

Set the list of tuples, each tuple specifying the component and parameter name, to be realized independently as well as the list of tuples to be converted from the q, phi convention to the e1, e2 convention

baobab.bnn_priors.diagonal_bnn_prior module

class baobab.bnn_priors.diagonal_bnn_prior.DiagonalBNNPrior(bnn_omega, components)

Bases: baobab.bnn_priors.base_bnn_prior.BaseBNNPrior

BNN prior with independent parameters

Note

This BNNPrior is cosmology-agnostic. For a version that’s useful for H0 inference, see DiagonalCosmoBNNPrior.

sample()

Gets kwargs of sampled parameters to be passed to lenstronomy

Returns:dictionary of config-specified components (e.g. lens mass), itself a dictionary of sampled parameters corresponding to the config-specified profile of that component Return type:dict

baobab.bnn_priors.cov_bnn_prior module

class baobab.bnn_priors.cov_bnn_prior.CovBNNPrior(bnn_omega, components)

Bases: baobab.bnn_priors.base_bnn_prior.BaseBNNPrior

BNN prior with marginally covariant parameters

Note

This BNNPrior is cosmology-agnostic. For a version that’s useful for H0 inference, see CovCosmoBNNPrior.

sample()

Gets kwargs of sampled parameters to be passed to lenstronomy

Returns:dictionary of config-specified components (e.g. lens mass), itself a dictionary of sampled parameters corresponding to the config-specified profile of that component Return type:dict

baobab.bnn_priors.empirical_bnn_prior module

class baobab.bnn_priors.empirical_bnn_prior.EmpiricalBNNPrior(bnn_omega, components)

Bases: baobab.bnn_priors.base_bnn_prior.BaseBNNPrior, baobab.bnn_priors.base_cosmo_bnn_prior.BaseCosmoBNNPrior

BNN prior that encodes physical correlations between parameters

get_agn_absolute_magnitude(z_src)

Get the AGN absolute magnitude at 1450A, sampled from the luminosity function for its redshift bin

Parameters:z_src (float) – the AGN redshift Returns:AGN absolute magnitude at 1450A Return type:float

get_lens_absolute_magnitude(vel_disp)

Get the lens absolute magnitude from the Faber-Jackson relation given the realized velocity dispersion, with some scatter

Parameters:vel_disp (float) – the velocity dispersion in km/s Returns:the V-band absolute magnitude Return type:float

get_lens_apparent_magnitude(M_lens, z_lens)

Get the lens apparent magnitude from the Faber-Jackson relation given the realized velocity dispersion, with some scatter

Parameters:

• M_lens (float) – the V-band absolute magnitude of lens
• z_lens (float) – the lens redshift

Note

Does not account for peculiar velocity or dust. K-correction is approximate and implicit, as the absolute magnitude is in the V-band (480nm ~ 650nm) and, for z ~ 2-3, this portion of the SED roughly lands in the IR.

Returns:the apparent magnitude in the IR Return type:float

get_lens_size(vel_disp, z_lens, m_V)

Get the lens V-band efefctive radius from the Fundamental Plane relation given the realized velocity dispersion and apparent magnitude, with some scatter

Parameters:

• vel_disp (float) – the velocity dispersion in km/s
• z_lens (float) – redshift
• m_V (float) – V-band apparent magnitude

Returns:

the effective radius in kpc and arcsec

Return type:

tuple

get_src_absolute_magnitude(z_src)

Sample the UV absolute magnitude from the luminosity function for the given redshift and convert into apparent magnitude

Parameters:z_src (float) – the source redshift Returns:the absolute magnitude at 1500A Return type:float

get_src_apparent_magnitude(M_src, z_src)

Convert the souce absolute magnitude into apparent magnitude

Parameters:

• M_src (float) – the source absolute magnitude
• z_src (float) – the source redshift

Note

Does not account for peculiar velocity or dust. K-correction is approximate and implicit, as the absolute magnitude is at 150nm and, for z ~ 5-9, this portion of the SED roughly lands in the IR.

Returns:the apparent magnitude in the IR Return type:float

get_src_size(z_src, M_V_src)

Get the effective radius of the source from its empirical relation with V-band absolute magnitude and redshift

Parameters:

• M_V_src (float) – V-band absolute magnitude of the source
• z_src (float) – source redshift

Returns:

tuple of the effective radius in kpc and arcsec

Return type:

tuple

sample()

Gets kwargs of sampled parameters to be passed to lenstronomy

Returns:dictionary of config-specified components (e.g. lens mass), itself a dictionary of sampled parameters corresponding to the config-specified profile of that component Return type:dict

sample_vel_disp(vel_disp_cfg)

Sample velocity dispersion from the config-specified model, on a grid with the range and resolution specified in the config

Parameters:vel_disp_cfg (dict) – Copy of cfg.bnn_omega.kinematics.vel_disp Returns:a realization of velocity dispersion Return type:float

baobab.bnn_priors.kinematics_models module

baobab.bnn_priors.kinematics_models.vel_disp_function_CPV2007(vel_disp_grid)

Evaluate the velocity dispersion function from the fit on SDSS DR6 by [1]_ on a provided grid and normalizes the result to unity, so it can be used as a PMF from which to draw the velocity dispersion.

Parameters:vel_disp_grid (array-like) – a grid of velocity dispersion values in km/s

Note

The returned array is normalized to unity and we treat it as a PMF from which to sample the velocity dispersion. We also use the exact fit values also used in LensPop ([2]_).

References

[1]	Choi, Yun-Young, Changbom Park, and Michael S. Vogeley. “Internal and collective properties of galaxies in the Sloan Digital Sky Survey.” The Astrophysical Journal 658.2 (2007): 884.

[2]	Collett, Thomas E. “The population of galaxy–galaxy strong lenses in forthcoming optical imaging surveys.” The Astrophysical Journal 811.1 (2015): 20.

Returns:the velocity dispersion function evaluated at vel_disp_grid Return type:array-like, same shape as vel_disp_grid

baobab.bnn_priors.parameter_models module

baobab.bnn_priors.parameter_models.approximate_theta_E_for_SIS(vel_disp_iso, z_lens, z_src, cosmo)

Compute the Einstein radius for a given isotropic velocity dispersion assuming a singular isothermal sphere (SIS) mass profile

Parameters:

• vel_disp_iso (float) – isotropic velocity dispersion, or an approximation to it, in km/s
• z_lens (float) – the lens redshift
• z_src (float) – the source redshift
• cosmo (astropy.cosmology object) – the cosmology

Note

The computation is purely analytic.

Returns:the Einstein radius for an SIS in arcsec Return type:float

class baobab.bnn_priors.parameter_models.FaberJackson(slope=None, intercept=None, fit_data=None)

Bases: object

Represents the Faber-Jackson (FJ) relation between velocity dispersion and luminosity of elliptical galaxies.

FJ is a projection of the Fundamental Plane (FP) relation.

get_luminosity(vel_disp)

Evaluate the V-band luminosity L_V expected from the FJ relation for a given velocity dispersion

Parameters:vel_disp (float) – the velocity dispersion in km/s Returns:log(L_V/L_solar) Return type:float

class baobab.bnn_priors.parameter_models.FundamentalPlane(a=None, b=None, c=None, intrinsic_scatter=0.0, fit_data=None)

Bases: object

Represents the Fundamental Plane (FP) relation between the velocity dispersion, luminosity, and effective radius for elliptical galaxies

Luminosity is expressed as apparent magnitude in this form.

get_effective_radius(vel_disp, m_V)

Evaluate the size expected from the FP relation for a given velocity dispersion and V-band apparent magnitude

Parameters:

• vel_disp (float) – the velocity dispersion in km/s
• m_V (float) – the apparent V-band magnitude

Returns:

the effective radius in kpc

Return type:

float

class baobab.bnn_priors.parameter_models.FundamentalMassHyperplane(a=None, b=None, intrinsic_scatter=0.0, fit_data=None)

Bases: object

Represents bivariate relations (projections) within the Fundamental Mass Hyperplane (FMHP) relation between the stellar mass, stellar mass density, effective radius, and velocity dispersion of massive ETGs.

Only the relation between the power-law mass slope (gamma) and effective radius is currently supported.

get_gamma_from_R_eff(R_eff)

Evaluate the power-law slope of the mass profile from its power-law relation with effective radius

Parameters:R_eff (float) – the effective radius in kpc Returns:the power-law slope, gamma Return type:float

get_gamma_from_vel_disp(vel_disp)

Evaluate the power-law slope of the mass profile from its power-law relation with effective radius

Parameters:vel_disp (float) – the velocity dispersion in km/s Returns:the power-law slope, gamma Return type:float

class baobab.bnn_priors.parameter_models.AxisRatioRayleigh(a=None, b=None, lower=0.2, fit_data=None)

Bases: object

Represents various scaling relations that the axis ratio can follow with quantities like velocity dispersion, when its PDF is assumed to be a Rayleigh distribution

Only the relation with velocity dispersion is currently supported.

get_axis_ratio(vel_disp)

Sample (one minus) the axis ratio of the lens galaxy from the Rayleigh distribution with scale that depends on velocity dispersion

Parameters:vel_disp (float) – velocity dispersion in km/s Returns:the axis ratio q Return type:float

baobab.bnn_priors.parameter_models.redshift_binned_luminosity_function(z, M_grid)

Sample FUV absolute magnitude from the redshift-binned luminosity function

Parameters:

• z (float) – galaxy redshift
• M_grid (array-like) – grid of FUV absolute magnitudes at which to evaluate luminosity function

Note

For z < 4, we use the Schechter function fits in Table 1 of [1]_ and, for 4 < z < 8, those in Table 4 of [2]_. z > 8 are binned into the z=8 bin. I might add high-redshift models, e.g. from [3]_.

References

[1]	Arnouts, Stephane, et al. “The GALEX VIMOS-VLT Deep Survey* Measurement of the Evolution of the 1500 Å Luminosity Function.” The Astrophysical Journal Letters 619.1 (2005): L43.

[2]	Finkelstein, Steven L., et al. “The evolution of the galaxy rest-frame ultraviolet luminosity function over the first two billion years.” The Astrophysical Journal 810.1 (2015): 71.

[3]	Kawamata, Ryota, et al. “Size–Luminosity Relations and UV Luminosity Functions at z= 6–9 Simultaneously Derived from the Complete Hubble Frontier Fields Data.” The Astrophysical Journal 855.1 (2018): 4.

Returns:unnormalized function of the absolute magnitude at 1500A Return type:array-like

baobab.bnn_priors.parameter_models.size_from_luminosity_and_redshift_relation(z, M_V)

Sample the effective radius of Lyman break galaxies from the relation with luminosity and redshift

Parameters:

• z (float) – galaxy redshift
• M_V (float) – V-band absolute magnitude

Note

The relation and scatter agree with [1]_ and [2]_, which both show that size decreases with higher redshift. They have been used in LensPop ([3]_) for source galaxies.

References

[1]	Mosleh, Moein, et al. “The evolution of mass-size relation for Lyman break galaxies from z= 1 to z= 7.” The Astrophysical Journal Letters 756.1 (2012): L12.

[2]	Huang, Kuang-Han, et al. “The bivariate size-luminosity relations for Lyman break galaxies at z∼ 4-5.” The Astrophysical Journal 765.1 (2013): 68.

[3]	Collett, Thomas E. “The population of galaxy–galaxy strong lenses in forthcoming optical imaging surveys.” The Astrophysical Journal 811.1 (2015): 20.

Returns:a sampled effective radius in kpc Return type:float

class baobab.bnn_priors.parameter_models.AGNLuminosityFunction(M_grid, z_bins=None, alphas=None, betas=None, M_stars=None, fit_data=None)

Bases: object

Redshift-binned AGN luminosity function parameterized as a double power-law

get_double_power_law(alpha, beta, M_star)

Evaluate the double power law at the given grid of absolute magnitudes

Parameters:

• alpha (float) – bright-end slope of the double power-law luminosity function
• beta (float) – faint-end slope of the double power-law luminosity function
• M_star (float) – break magnitude

Note

Returned luminosity function is normalized to unity. See Note under slope of the double power-law luminosity function.

Returns:the luminosity function evaluated at self.M_grid and normalized to unity Return type:array-like

sample_agn_luminosity(z)

Sample the AGN luminosity from the redshift-binned luminosity function

Parameters:z (float) – the AGN redshift Returns:sampled AGN luminosity at 1450A in mag Return type:float