drpangloss.models Functions

cvis_binary_angular(u, v, sep, pa, contrast)

Compute complex visibilities for an angular-parameterized binary model.

Parameters:

Name	Type	Description	Default
u	array - like	Baseline u coordinates in wavelength units.	required
v	array - like	Baseline v coordinates in wavelength units.	required
sep	float or array - like	Separation in milliarcseconds.	required
pa	float or array - like	Position angle in degrees.	required
contrast	float or array - like	Contrast ratio star/companion.	required

Returns:

Type	Description
array - like	Complex visibility samples.

cvis_binary(u, v, ddec, dra, planet)

Compute complex visibilities for a Cartesian-parameterized binary model.

Parameters:

Name	Type	Description	Default
u	array - like	Baseline u coordinates in wavelength units.	required
v	array - like	Baseline v coordinates in wavelength units.	required
ddec	float or array - like	Declination offset in milliarcseconds.	required
dra	float or array - like	Right-ascension offset in milliarcseconds.	required
planet	float or array - like	Flux ratio of the companion.	required

Returns:

Type	Description
array - like	Complex visibility samples.

loglike(values, params, data_obj, model_class)

Abstract log-likelihood function for a given model class and data object, assuming Gaussian errors.

Parameters:

Name	Type	Description	Default
values	array - like	Values of the model parameters.	required
params	list	List of parameter names.	required
data_obj	OIData	Object containing the data to be fitted.	required
model_class	class	Model class to be fitted to the data.	required

Returns:

Type	Description
float	Log-likelihood value.

loglike_nosignal(values, params, data_obj, model_class)

Abstract null log-likelihood function for a given model class and data object, assuming Gaussian errors.

Parameters:

Name	Type	Description	Default
values	array - like	Values of the model parameters.	required
params	list	List of parameter names.	required
data_obj	OIData	Object containing the data to be fitted.	required
model_class	class	Model class to be fitted to the data.	required

Returns:

Type	Description
float	Log-likelihood value.

laplace_cov(values, params, data_obj, model_class)

Compute the full Laplace covariance matrix for all model parameters jointly.

Computes the inverse of the Hessian of the negative log-likelihood with respect to all parameters in params simultaneously, returning an N x N covariance matrix (where N = len(params)).

.. note:: This function returns the full covariance matrix over all N parameters. To obtain only the marginal flux uncertainty at a fixed position, use :func:laplace_contrast_uncertainty instead.

Parameters:

Name	Type	Description	Default
values	array - like	Values of the model parameters.	required
params	list	List of parameter names.	required
data_obj	OIData	Object containing the data to be fitted.	required
model_class	class	Model class to be fitted to the data.	required

Returns:

Type	Description
array - like	N x N covariance matrix, where N = len(params).

laplace_contrast_uncertainty(flux, dra, ddec, data_obj, model_class, params=None)

Compute the Laplace uncertainty in flux at a fixed sky position.

Unlike :func:laplace_cov, which inverts the full N-parameter Hessian, this function fixes dra and ddec and computes only the scalar curvature of the negative log-likelihood along the flux axis alone:

.. math::

\sigma_f = \left(\frac{\partial^2 (-\log L)}{\partial f^2}\right)^{-1/2}

This is a 1-D (scalar) second derivative, not a matrix inversion. It is appropriate when the position is held fixed (e.g. on a detection grid) and only the contrast uncertainty at that grid point is needed. For the joint uncertainty over all parameters, use :func:laplace_cov instead.

Parameters:

Name	Type	Description	Default
flux	float	Flux ratio value at which the local Laplace uncertainty is evaluated.	required
dra	float	Right ascension offset in mas (held fixed).	required
ddec	float	Declination offset in mas (held fixed).	required
data_obj	OIData	Object containing the data to be fitted.	required
model_class	class	Model class to be fitted to the data.	required
params	list[str] or tuple[str, str, str]	Parameter names corresponding to (dra, ddec, flux). Defaults to ["dra", "ddec", "flux"].	None

Returns:

Type	Description
float	Scalar uncertainty in the contrast (standard deviation along flux axis).

fisher(values, params, data_obj, model_class, ridge=0.0)

Approximate the local Fisher matrix at a parameter point.

Parameters:

Name	Type	Description	Default
values	array - like	Parameter vector at which to evaluate the local curvature.	required
params	list[str]	Parameter names corresponding to values.	required
data_obj	OIData	Observational data object.	required
model_class	class	Model class used to evaluate the likelihood.	required
ridge	float	Diagonal regularization term.	0.0

Returns:

Type	Description
array - like	Fisher information matrix.

chi2ppf(p, df)

Percentile function for chi-square.

For df=1 (the path used in nsigma), use the closed-form identity based on the standard normal quantile, i.e. square norm.ppf((p+1)/2). This remains JAX-native, differentiable, and fast.

For df != 1, this falls back to numpyro's gammaincinv backend when available.

Parameters:

Name	Type	Description	Default
p	array - like	Percentile value	required
df	array - like	Degrees of freedom	required

Returns:

Type	Description
array - like	Corresponding chi2 value to the percentile

nsigma(chi2r_test, chi2r_true, ndof)

Parameters:

Name	Type	Description	Default
chi2r_test		Reduced chi-squared of test model.	required
chi2r_true		Reduced chi-squared of true model.	required
ndof		Number of degrees of freedom.	required

Returns:

Name	Type	Description
nsigma	float	Detection significance.

closure_phases(cvis, index_cps1, index_cps2, index_cps3)

Calculate closure phases from complex visibilities.

Parameters:

Name	Type	Description	Default
cvis	array - like	Complex visibilities.	required
index_cps1	array - like	First baseline indices for each closure triangle.	required
index_cps2	array - like	Second baseline indices for each closure triangle.	required
index_cps3	array - like	Third baseline indices for each closure triangle.	required

Returns:

Type	Description
array - like	Closure phases in degrees.

cp_indices(vis_sta_index, cp_sta_index)

Map closure-triangle station indices to baseline indices.

Parameters:

Name	Type	Description	Default
vis_sta_index	array - like	Baseline station index pairs from visibility data.	required
cp_sta_index	array - like	Triangle station index triplets from closure-phase data.	required

Returns:

Type	Description
tuple[ndarray, ndarray, ndarray]	Arrays (i_cps1, i_cps2, i_cps3) identifying the three baselines composing each closure phase.