MulensModel.fitdata module

class MulensModel.fitdata.FitData(model=None, dataset=None, fix_blend_flux=False, fix_source_flux=False, fix_source_flux_ratio=False)

Bases: object

Performs a least squares linear fit for given dataset and model to determine the source flux(es) and (optionally) blend flux. After creating the object, you must run update() to perform the linear fit for the fluxes and calculate the chi2. To perform the linear fit without calculating chi2, you can run fit_fluxes(). If you change anything in the object, e.g. the model parameters, you must re-run update() or fit_fluxes().

Arguments :

model: Model object

The model to fit to the data.

dataset: MulensData object

A single photometric dataset to be fitted.

fix_blend_flux: False or float, optional

Default is False, i.e. allow the blend flux to be a free parameter. If set to a float, it will fix the blend value to that value.

fix_source_flux: False, float, or list, optional

Default is False, i.e. allow the source flux to be a free parameter. If set to a float, it will fix the source value to that value. For binary source models, a list should be used to set the fluxes of the individual sources or fix one and not the other, e.g. [2.3, False] would fix source_flux_0 to 2.3 but allow a free fit to source_flux_1.

fix_source_flux_ratio: False or float, optional

For binary source models, source_flux_ratio is the flux ratio between two components, i.e., source_flux_ratio = source_flux_1 / source_flux_0 Default is False, i.e. allow the source flux to be a free parameter. If set to a float, it will fix the source value to that value.

update(bad=False)

Calculate the best-fit source and blend fluxes as well as the chi2.

Parameters :

bad: bool

Default is False. If True recalculates the data magnification for each point to ensure that there are values even for bad datapoints.

No returns.

fit_fluxes()

Execute the linear least squares fit to determine the fitted fluxes. Sets the values of source_fluxes, blend_flux, and (if applicable) source_flux.

Does not calculate chi2. To fit for the fluxes and calculate chi2, run update().

No parameters.

No returns.

get_data_magnification(bad=False)

Calculates the model magnification for each data point.

Parameters :

bad: boolean

If True, calculates the magnification for all points. If False, only calculates the magnification for good data points. Values for bad data points are set to 0. Default is False.

Returns :

data_magnification: np.ndarray

The model magnification evaluated for each datapoint. If there is more than one source, the magnification of each source is reported separately.

get_model_fluxes(bad=False)

Calculate model in flux space.

Parameters :

bad: bool

Default is False. If True recalculates the data magnification for each point to ensure that the values for bad datapoints are calculated (otherwise, they are set to the magnitude of the blend).

Returns :

model_flux: np.ndarray

The model flux evaluated for each datapoint.

get_model_magnitudes(**kwargs)

Calculate model in magnitude space

Parameters :

**kwargs:

see get_model_fluxes()

Returns :

model_mag: np.ndarray

The model magnitude evaluated for each datapoint.

scale_fluxes(source_flux, blend_flux)

Rescale the data fluxes to an arbitrary flux scale:

flux = source_flux_0 * (data.flux - blend_flux) / source_flux flux += blend_flux_0 err_flux = source_flux_0 * data.err_flux / source_flux

Parameters :

source_flux: float, list, np.array

Flux of the source in the desired system. If n_sources > 1 and source_flux has more than one element, the elements are summed to produce the overall scaling flux.

blend_flux: float

Flux of the blend in the desired system

Returns :

flux: np.ndarray

Fluxes from the data rescaled to the desired system.

err_flux: np.ndarray

Uncertainties of fluxes from the data rescaled to the desired system.

get_residuals(phot_fmt=None, source_flux=None, blend_flux=None, bad=False, type=None)

Calculate the residuals for each datapoint relative to the model.

Parameters :

phot_fmt: str, optional

specify whether the residuals should be returned in magnitudes (‘mag’) or in flux (‘flux’). Default is ‘mag’. If ‘scaled’, will return the residuals in magnitudes scaled to source_flux, blend_flux.

source_flux, blend_flux: float

reference source and blend fluxes for scaling the residuals

bad: bool

Default is False. If True recalculates the data magnification for each point to ensure that there are values even for bad datapoints.

type:

DEPRECATED, see “phot_fmt” above.

Returns :

residuals: np.ndarray

the residuals for the corresponding dataset.

errorbars: np.ndarray

the scaled errorbars for each point. For plotting errorbars for the residuals.

get_chi2_gradient(parameters)

Fits fluxes and calculates chi^2 gradient (also called Jacobian), i.e., \(d chi^2/d parameter\).

Parameters :

parameters: str or list, required

Parameters with respect to which gradient is calculated. Currently accepted parameters are: t_0, u_0, t_eff, t_E, pi_E_N, and pi_E_E. The parameters for which you request gradient must be defined in py:attr:~model.

Returns :

gradient: float or np.ndarray

chi^2 gradient

calculate_chi2_gradient(parameters)

Calculates chi^2 gradient (also called Jacobian), i.e., \(d chi^2/d parameter\) WITHOUT refitting for the fluxes. Saves computations if, e.g., you want to retrieve both py:attr:~chi2 and py:attr:~chi2_gradient.

Parameters :

parameters: str or list, required

Parameters with respect to which gradient is calculated. Currently accepted parameters are: t_0, u_0, t_eff, t_E, pi_E_N, and pi_E_E. The parameters for which you request gradient must be defined in py:attr:~model.

Returns :

gradient: float or np.ndarray

chi^2 gradient

get_d_A_d_params_for_point_lens_model(parameters)

Calculate d A / d parameters for a point lens model.

Parameters :

parameters: list

List of the parameters to take derivatives with respect to.

Returns :

dA_dparam: dict

Keys are parameter names from parameters argument above. Values are the partial derivatives for that parameter evaluated at each data point.

get_dataset_trajectory()

Retrieve a Trajectory object. If the dataset has an ephemerides_file, apply it to the Trajectory, even if it is not part of the model.

No parameters.

Returns :

trajectory: Trajectory

Trajectory for given dataset.

get_d_A_d_u_for_PSPL_model()

Calculate dA/du for PSPL point-source–point-lens model.

No parameters.

Returns :

dA_du: np.ndarray

Derivative dA/du.

get_d_A_d_u_for_point_lens_model()

Calculate dA/du for point-source–point-lens model. For finite source models see FSPLDerivs or py:func:get_d_A_d_params_for_point_lens_model()

Returns :

dA_du: np.ndarray

Derivative dA/du.

get_d_A_d_rho()

Calculate the derivative of the magnification with respect to rho.

Returns:

d_A_d_rho: np.ndarray

derivative

property chi2_gradient

float or np.ndarray

Previously calculated chi^2 gradient (also called Jacobian), i.e., \(d chi^2/d parameter\). See get_chi2_gradient() and calculate_chi2_gradient().

Gives None if the chi2 gradient was not previously calculated using one of the functions mentioned above.

property chi2

float The total chi2 for the fitted dataset. Good points only. See good.

If None, you need to run update() to execute the linear fit and calculate the chi2.

property chi2_per_point

np.ndarray

The chi^2 contribution from each data point, e.g., chi2_per_point[k] returns the chi2 contribution from the k-th point of dataset. Includes bad datapoints.

If None, you need to run update() to execute the linear fit and calculate the chi2.

property source_flux

float

The fitted source flux. Only defined for models with a single source. See also source_fluxes

If None, you need to run fit_fluxes() or update() to execute the linear fit.

property source_fluxes

np.array

The fitted source flux(es).

If None, you need to run fit_fluxes() or update() to execute the linear fit.

property blend_flux

float

The fitted blend flux or the value set by fix_blend_flux (see Keywords).

If None, you need to run fit_fluxes() or update() to execute the linear fit.

property source_flux_ratio

float

source_flux_ratio = source_flux_1 / source_flux_0

i.e., the ratio of the fitted source fluxes or the value set by fix_source_flux_ratio (see Keywords).

If None, you need to run fit_fluxes() or update() to execute the linear fit.

property dataset

MulensData

A single photometric dataset to be fitted.

property model

Model

The model to fit to the data.

property data_magnification

Returns previously calculated magnifications. To calculate the magnifications (e.g., if something changed in the model), use get_data_magnification().

Returns :

data_magnification: np.ndarray

The model magnification evaluated for each datapoint. If there is more than one source, the magnification of each source is reported separately.

property gamma

float

Limb-darkening coefficient for this fit. Set by bandpass and get_limb_coeff_gamma().

class FSPL_Derivatives(fit)

Bases: object

Calculates derivatives of an FSPL model for each datapoint.

Arguments :

fit: py:class:FitData object

get_gradient(parameters)

Return the gradient of the magnification with respect to the FSPL parameters.

get_d_A_d_rho()

Return the derivative of the magnification with respect to rho.