coconut_tools.magnetogram.NLD_implicit_method

Preprocess magnetograms with nonlinear diffusion filtering.

This module shares the magnetogram download, reading, temporal interpolation, effective-time handling, Stonyhurst rotation, flux correction, plotting, and COCONUT boundary writing utilities from sph_filtering. Its specific processing step applies optional Gaussian smoothing followed by the Perona-Malik nonlinear diffusion filter implemented in nonlinear_diffusion_filter.

Author: Jose Murteira Cleaned and modularized by: Luis

Functions

filter_radial_field(Br, phi, theta[, ...])	Apply optional Gaussian smoothing and nonlinear diffusion to Br.
process_config(config[, method_used])	Process all target times described by one nonlinear diffusion config.
process_magnetogram_date(config, target_date)	Process one target time with the nonlinear diffusion pipeline.

coconut_tools.magnetogram.NLD_implicit_method.filter_radial_field(Br, phi, theta, iterations=3, apply_gaussian=True, gaussian_sigma=1.0, dx_override=1.0, dy_override=1.0, tau=1.0)

Apply optional Gaussian smoothing and nonlinear diffusion to Br.

phi and theta are kept in the signature for consistency with the other filters; the numerical spacing used by the nonlinear diffusion solver comes from dx_override and dy_override.

Parameters:

• Br (np.ndarray) – 2D array of the radial magnetic field.

• phi (np.ndarray) – 1D longitude grid in radians.

• theta (np.ndarray) – 1D colatitude grid in radians.

• iterations (int) – Number of iterations for nonlinear diffusion.

• apply_gaussian (bool) – Whether to apply Gaussian filtering before diffusion.

• gaussian_sigma (float) – Sigma used in Gaussian smoothing.

• dx_override (float) – Spatial resolution in x-direction (default: 1.0).

• dy_override (float) – Spatial resolution in y-direction (default: 1.0).

• tau (float) – Time step for the nonlinear diffusion.

Returns:

Filtered Br field and final time step.

Return type:

tuple[np.ndarray, float]

coconut_tools.magnetogram.NLD_implicit_method.process_config(config, method_used='NLD')

Process all target times described by one nonlinear diffusion config.

With only date set, one target time is processed. With cadence_hours and total_hours, the function builds a time sequence and processes each target independently. When output_path_fig is omitted each figure is named from the effective magnetogram time.

Parameters:

• config (dict[str, Any]) – Processing configuration.

• method_used (str) – Method label used in output filenames.

Returns:

Per-date processing results.

Return type:

list[dict[str, Any]]

coconut_tools.magnetogram.NLD_implicit_method.process_magnetogram_date(config, target_date, method_used='NLD', output_path_fig=None)

Process one target time with the nonlinear diffusion pipeline.

The function downloads or reuses a magnetogram, optionally interpolates a four-map stencil, computes and logs the effective magnetogram time, optionally rotates to Stonyhurst, optionally balances net flux, applies the nonlinear diffusion filter, writes the boundary file, and optionally saves a diagnostic figure.

Parameters:

• config (dict[str, Any]) – Processing configuration. Common keys are map_type, output_dir, download_dir, r_st, amp, adapt_map, write_map, show_map, visu_type, interpolation_order, interpolation, rotate_to_stonyhurst, flux_correct, flux_correction_method, drms_email or jsoc_email, tau, iterations, apply_gaussian, gaussian_sigma, dx_override, and dy_override.

• target_date – Requested processing time.

• method_used (str) – Method label used in output filenames.

• output_path_fig (str | None) – Explicit diagnostic figure path. If omitted, the figure name is built from the effective time.

Returns:

Processing metadata, including target date, effective_date, output paths, selected local file or interpolation stencil, optional Br_linear, final diffusion timestep, and rotation angle.

Return type:

dict[str, Any]