drpangloss.grid_fit

Functions

likelihood_grid(data_obj, model_class, samples_dict)

Function to vmap a likelihood function over a grid of parameter values provided in a dictionary.

Parameters:

Name	Type	Description	Default
data_obj	OIData	Object containing the data to be fitted.	required
model_class	class	Model class to be fitted to the data.	required
samples_dict		Dictionary of parameter names and values to be fitted to the data.	required

Returns:

Type	Description
array - like	Log-likelihood values over the grid of parameter values.

optimized_likelihood_grid(data_obj, model_class, samples_dict)

Function to optimize the contrast of a model over a grid of parameter values provided in a dictionary.

Parameters:

Name	Type	Description	Default
data_obj	OIData	Object containing the data to be fitted.	required
model_class	class	Model class to be fitted to the data.	required
samples_dict		Dictionary of parameter names and values to be fitted to the data.	required

Returns:

Type	Description
array - like	Optimized contrast values over the grid of parameter values.

optimized_contrast_grid(data_obj, model_class, samples_dict)

Function to optimize the contrast of a model over a grid of parameter values provided in a dictionary.

Parameters:

Name	Type	Description	Default
data_obj	OIData	Object containing the data to be fitted.	required
model_class	class	Model class to be fitted to the data.	required
samples_dict		Dictionary of parameter names and values to be fitted to the data.	required

Returns:

Type	Description
array - like	Optimized contrast values over the grid of parameter values.

laplace_contrast_uncertainty_grid(best_contrast_indices, data_obj, model_class, samples_dict)

Calculate the uncertainty with the Laplace method over a grid of parameters, for an optimized fit between a model and data object.

Parameters:

Name	Type	Description	Default
best_contrast_indices	array - like	Indices of the best contrast values in a grid calculated with likelihood_grid.	required
data_obj	OIData	Object containing the data to be fitted.	required
model_class	class	Model class to be fitted to the data.	required
samples_dict		Dictionary of parameter names and values to be fitted to the data.	required

Returns:

Type	Description
array - like	Uncertainty in the contrast.

ruffio_upperlimit(mean, sigma, percentile)

Calculate the upper limit of a distribution given the mean and standard deviation. This is a vectorized JAX implementation of Ruffio et al. (2018).

Parameters:

Name	Type	Description	Default
mean	array - like	Mean of the distribution.	required
sigma	array - like	Standard deviation of the distribution.	required
percentile	float	Percentile value for the upper limit.	required

Returns:

Type	Description
array - like	Upper limit of the distribution.

azimuthalAverage(image, center=None, stddev=False, returnradii=False, return_nr=False, binsize=0.5, weights=None, steps=False, interpnan=False, left=None, right=None, return_max=False)

Calculate an azimuthally averaged radial profile for a 2D image.

Parameters:

Name	Type	Description	Default
image	array - like	Two-dimensional image.	required
center	tuple[float, float]	Pixel coordinates (x, y) of the radial center. If omitted, the geometric image center is used.	None
stddev	bool	If True, return the azimuthal standard deviation instead of the weighted mean.	False
returnradii	bool	If True, return (radii, profile).	False
return_nr	bool	If True, return (n_per_bin, radii, profile).	False
binsize	float	Radial bin width in pixel units.	0.5
weights	array - like	Per-pixel weights. Must match image.shape.	None
steps	bool	If True, return step-ready (x, y) arrays.	False
interpnan	bool	If True, interpolate over bins with NaN profile values.	False
left	float	Left extrapolation value passed to numpy.interp when interpnan=True.	None
right	float	Right extrapolation value passed to numpy.interp when interpnan=True.	None
return_max	bool	If True, return the maximum value per radial bin.	False

Returns:

Type	Description
array - like or tuple	Radial profile array, or a tuple depending on returnradii, return_nr, or steps.

Notes

Empty bins are returned as NaN unless interpolated.

absil_limits(samples_dict, data_obj, model_class, sigma)

Using Jax for optimization, calculate the detection limits for a given model class and data object. This is by finding the contrast at which the detection significance is equal to the sigma value. For example, if we set sigma = 3, we optimize in each coordinate cell to find the contrast at which we would be 3 sigma confident that the companion is detected.

Parameters:

Name	Type	Description	Default
samples_dict		Dictionary of parameter names and values to be fitted to the data, eg dra and ddec grids.	required
data_obj		Observational data in the format of an OIData object.	required
model_class		Model class to be fitted to the data.	required
sigma		Detection significance.	required

Returns:

Name	Type	Description
res	float	Maximum relative flux of companion.