arviz.from_emcee#
- arviz.from_emcee(sampler=None, var_names=None, slices=None, arg_names=None, arg_groups=None, blob_names=None, blob_groups=None, index_origin=None, coords=None, dims=None)[source]#
Convert emcee data into an InferenceData object.
For a usage example read Converting emcee objects to InferenceData
- Parameters
- sampler
emcee.EnsembleSampler
Fitted sampler from emcee.
- var_names
list
ofstr
, optional A list of names for variables in the sampler
- slices
list
of array_like orslice
, optional A list containing the indexes of each variable. Should only be used for multidimensional variables.
- arg_names
list
ofstr
, optional A list of names for args in the sampler
- arg_groups
list
ofstr
, optional A list of the group names (either
observed_data
orconstant_data
) where args in the sampler are stored. If None, all args will be stored in observed data group.- blob_names
list
ofstr
, optional A list of names for blobs in the sampler. When None, blobs are omitted, independently of them being present in the sampler or not.
- blob_groups
list
ofstr
, optional A list of the groups where blob_names variables should be assigned respectively. If blob_names!=None and blob_groups is None, all variables are assigned to log_likelihood group
- coords
dict
of {str
array_like}, optional Map of dimensions to coordinates
- dims
dict
of {str
list
ofstr
}, optional Map variable names to their coordinates
- sampler
- Returns