leniax.utils package

leniax.utils.get_container(omegaConf, main_path)[source]

Retrieve the populated container from an omegaConf

This functions is used to: - Populate missing mandatory value in Leniax configuration - Handle versioning update if necessary

# Use with Hydra as follow
@hydra.main(config_path=config_path, config_name=config_name)
def run(omegaConf: DictConfig) -> None:
    config = leniax_utils.get_container(omegaConf, config_path)

Parameters

omegaConf (omegaconf.dictconfig.DictConfig) – Hydra’s Omega configuration
main_path (str) – Absolute path of the configuration directory

Returns

A populated python dictionnary

Return type

Dict

leniax.utils.update_config_v1_v2(config)[source]

Update a configuration from version 1 to version 2

Parameters: config (Dict) – A v1 configuration
Returns: A v2 configuration
Return type: Dict

leniax.utils.get_param(dic, key_string)[source]

Retrieve a parameter in a dictionnary using a string

Parameters

dic (Dict) – Dictionnary
key_string (str) – String representing the parameter path in the dicitonnary

Returns

The parameter

Return type

Any

leniax.utils.set_param(dic, key_string, value)[source]

Set a parameter in a dictionnary using a string

Parameters

dic (Dict) – Dictionnary
key_string (str) – String representing the parameter path in the dicitonnary
value (Any) – Value to be set

leniax.utils.st2fracs2float(st)[source]

Convert a string of fraction into an list of floats

String example: st = '1,2/3,6.7'

Parameters: st (str) – String of fractions.
Returns: An list of float
Return type: List[float]

leniax.utils.merge_cells(cells, other_cells, offset=None)[source]

Merge cells together using addition

Parameters

cells (jax._src.numpy.ndarray.ndarray) – Base cells
other_cells (jax._src.numpy.ndarray.ndarray) – Cells to be merged into the base cells
offset (Optional[List[int]]) – Offsets in each dimensions

Returns

Resulting cells

Return type

jax._src.numpy.ndarray.ndarray

leniax.utils.center_world(cells, field, potential, shift_idx, axes)[source]

Vectorized version of center_world. Takes similar arguments as center_world but with additional array axes over which center_world is mapped.

Original documentation:

Rool cells, field and potential

Args:
cells: Lenia state of shape [bs, other_axes...] field: Lenia field [bs, other_axes...] potential: Lenia potential [bs, other_axes...] shift_idx: Amount to roll of shape [bs, nb_axes]` axes: Axes to roll

Returns:
Updated cells, field and potential

Parameters

cells (jax._src.numpy.ndarray.ndarray) –
field (jax._src.numpy.ndarray.ndarray) –
potential (jax._src.numpy.ndarray.ndarray) –
shift_idx (jax._src.numpy.ndarray.ndarray) –
axes (Tuple[int, ...]) –

Return type

Tuple[jax._src.numpy.ndarray.ndarray, jax._src.numpy.ndarray.ndarray, jax._src.numpy.ndarray.ndarray]

leniax.utils.crop_zero(kernels)[source]

Crop zero values out for 3 and 4 dimensions array

Parameters: kernels (jax._src.numpy.ndarray.ndarray) – A 3 or 4 dimension kernels
Returns: The croppped kernels
Return type: jax._src.numpy.ndarray.ndarray

leniax.utils.auto_center(cells)[source]

Automatically center cells on its total mass centroid

Parameters: cells (jax._src.numpy.ndarray.ndarray) – Lenia state
Returns: The mass centered Lenia state
Return type: jax._src.numpy.ndarray.ndarray

leniax.utils.center_and_crop(cells)[source]

Automatically center cells on its total mass centroid and crop zeros values

Parameters: cells (jax._src.numpy.ndarray.ndarray) – Lenia state
Returns: The mass centered cropped Lenia state
Return type: jax._src.numpy.ndarray.ndarray

leniax.utils.get_image(cells_buffer, pixel_size, colormap)[source]

Convert a numpy array into a PIL image

Parameters

cells_buffer (numpy.ndarray) – A Lenia state of shape [C, world_dims...]
pixel_size (int) – Size of each state pixels in the image
colormap – A matplotlib compatible colormap

Returns

A PIL image

Return type

leniax.utils.plot_stats(save_dir, stats_dict)[source]

Plot Leniax statistics

Parameters

save_dir (str) – directory used to save assets.
stats_dict (Dict) – Statistics dictionnary

leniax.utils.generate_beta_faces(nb_betas, denominator)[source]

Generate a grid of all the valid beta values given a maximum number: of beta values.

This function makes sense only if we normalize our kernels.

Parameters

nb_betas (int) – Maximum number of betas values
denominator (int) – Denominator to construct the betas grid

Returns

The list of possible list of beta values

Return type

List[List[List[float]]]

leniax.utils.check_dir(dir)[source]

Ensure a directory exist and is not a file

Parameters: dir (str) – Checked directory

leniax.utils.save_config(save_dir, config)[source]

Save a configuration file

Parameters

save_dir (str) – directory used to save assets.
config (Dict) – Leniax configuration

leniax.utils.print_config(config)[source]

Pretty print a Leniax configuration

Parameters: config (Dict) – Leniax configuration

leniax.utils.load_img(fullpath, resize)[source]

Load an image as a np.array

Parameters

fullpath (str) – Absolute image path
resize (Tuple[int, int]) – Resize factors for the image dimensions

Returns

The image as a np.array

Return type

jax._src.numpy.ndarray.ndarray

leniax.utils.set_log_level(config)[source]

Set the python logging root level

Parameters: config (Dict) – Leniax configuration

leniax.utils.seed_everything(seed)[source]

Seed all the depedencies used by Leniax

Parameters: seed (int) – A seed integer
Returns: A JAX PRNG key
Return type: jax._src.prng.PRNGKeyArray

leniax.utils.get_needed_memory(config, nb_sols=1)[source]

Compute an approximate of the needed memory by different kind of simulations

Parameters

config (Dict) – Leniax configuration
nb_sols (int) – How many solutions will be simulated at the same time

Returns

A dictionnary witth different memory requirements