convenience function to plot a regular grid of lon lat coordinates

Convenience function that indexes the hyperparameter dict but returns a default value if not defined rather than an error

Calculates the pairwise Haversine distances between each point pair. Args: lon1 (numpy.ndarray): The longitudes of the first set of

assigns the label of lons, lats, to the closest lons_grid, lats_grid point

Args: coord: the deltaX or deltaY cur_freq: the frequency

searches a directory for checkpoints following a pattern (e.g., sphericalharmonics-siren) and returns the one with lowest val_loss. checkpoint

searches a directory for checkpoints following a pattern (e.g., sphericalharmonics-siren) and returns the one with lowest val_loss. checkpoint

samples random points on the sphere returns [N x 2] tensor of lon, lat coordinates in degree

convenience function to plot a scatter plot of values at lonlats

see last equation of https://en.wikipedia.org/wiki/Spherical_harmonics#Real_form

convenience function to store two triangle matrices in one where m can be negative

generate a loss function :param presence_only: if True, calculate presence-only loss as in Cole and al., 2023 :param loss_weight: weight

The simple "assume negative" loss from Cole and al., 2021. This is the standard cross-entropy loss.

MSE on logits.

The "assume negative" loss (same location, different species) from Cole and al., 2023.

The "assume negative" loss (same species, different location) from Cole and al., 2023.