py21cmmc.cosmoHammer.HDFStorage

class py21cmmc.cosmoHammer.HDFStorage(filename, name)

A HDF Storage utility, based on the HDFBackend from emcee v3.0.0.

Methods

__init__(filename, name)
get_blobs(**kwargs)	Get the chain of blobs for each sample in the chain.
get_chain(**kwargs)	Get the stored chain of MCMC samples.
get_last_sample()	Access the most recent sample in the chain.
get_log_prob(**kwargs)	Get the chain of log probabilities evaluated at the MCMC samples.
get_trialled_log_prob(**kwargs)	Get the chain of log probabilities evaluated as trials of the MCMC.
get_trials(**kwargs)	Get the stored chain of trials.
get_value(name[, flat, thin, discard])	Get a particular kind of entry from the backend file.
grow(ngrow, blobs)	Expand the storage space by some number of samples.
open([mode])	Open the backend file.
reset(nwalkers, params)	Clear the state of the chain and empty the backend.
save_step(coords, log_prob, blobs, truepos, ...)	Save a step to the file.

Attributes

accepted	Number of acceptances for each walker.
accepted_array	An array of bools representing whether parameter proposals were accepted.
blob_names	Names for each of the arbitrary blobs.
has_blobs	Whether this file has blobs in it.
initialized	Whether the file object has been initialized.
iteration	The iteration the chain is currently at.
param_guess	An initial guess for the parameters.
param_names	The parameter names.
random_state	The defining random state of the process.
shape	Tuple of (nwalkers, ndim).
size	The length of the chain.