def wrapper_distribute_and_combine_likelihood(df, base_draws_est, *args,
optim_paras, options):
dense_columns = create_dense_state_space_columns(optim_paras)
# Duplicate the DataFrame for each type.
if dense_columns:
n_obs = df.shape[0]
n_types = optim_paras["n_types"]
# Number each state to split the shocks later. This is necessary to keep the
# regression tests from failing.
df["__id"] = np.arange(n_obs)
# Each row of indices corresponds to a state whereas the columns refer to
# different types.
indices = np.arange(n_obs * n_types).reshape(n_obs, n_types)
df_ = pd.concat([df.copy().assign(type=i) for i in range(n_types)])
splitted_df = _split_dataframe(df_, dense_columns)
splitted_shocks = _split_shocks(base_draws_est, splitted_df,
indices, optim_paras)
else:
splitted_df = df
splitted_shocks = base_draws_est
out = func(splitted_df, splitted_shocks, *args, optim_paras, options)
out = pd.concat(out.values()).sort_index() if isinstance(out,
dict) else out
return out
def _create_conversion_dictionaries(self):
"""Create mappings between state space location indices and properties.
See :ref:`state space location indices <state_space_location_indices>`.
"""
self.dense_key_to_complex = {
i: k
for i, k in enumerate(self.dense_period_cores)
}
self.dense_key_to_core_key = {
i: self.dense_period_cores[self.dense_key_to_complex[i]]
for i in self.dense_key_to_complex
}
self.dense_key_to_choice_set = {
i: self.dense_key_to_complex[i][1]
for i in self.dense_key_to_complex
}
self.dense_key_to_core_indices = {
i: np.array(
self.core_key_to_core_indices[self.dense_key_to_core_key[i]])
for i in self.dense_key_to_complex
}
self.core_key_and_dense_index_to_dense_key = Dict.empty(
key_type=nb.types.UniTuple(nb.types.int64, 2),
value_type=nb.types.int64,
)
for i in self.dense_key_to_complex:
self.core_key_and_dense_index_to_dense_key[return_core_dense_key(
self.dense_key_to_core_key[i],
*self.dense_key_to_complex[i][2:],
)] = i
if self.dense is False:
self.dense_covariates_to_dense_index = {}
self.dense_key_to_dense_covariates = {
i: {}
for i in self.dense_key_to_complex
}
else:
n_dense = len(create_dense_state_space_columns(self.optim_paras))
self.dense_covariates_to_dense_index = Dict.empty(
key_type=nb.types.UniTuple(nb.types.int64, n_dense),
value_type=nb.types.int64,
)
for i, k in enumerate(self.dense):
self.dense_covariates_to_dense_index[k] = i
self.dense_key_to_dense_covariates = {
i: list(self.dense.keys())[self.dense_key_to_complex[i][2]]
for i in self.dense_key_to_complex
}
def wrapper_distribute_and_combine_df(df, *args, optim_paras,
**kwargs):
dense_columns = create_dense_state_space_columns(optim_paras)
if remove_type:
dense_columns.remove("type")
splitted_df = _split_dataframe(
df, dense_columns) if dense_columns else df
out = func(splitted_df, *args, optim_paras, **kwargs)
df = pd.concat(out.values()).sort_index() if isinstance(
out, dict) else out
return df
def _create_dense_state_space_covariates(dense_grid, optim_paras, options):
if dense_grid:
columns = create_dense_state_space_columns(optim_paras)
df = pd.DataFrame(data=dense_grid,
columns=columns).set_index(columns, drop=False)
covariates = compute_covariates(df, options["covariates_dense"])
covariates = covariates.apply(downcast_to_smallest_dtype)
covariates = covariates.to_dict(orient="index")
covariates = convert_dictionary_keys_to_dense_indices(covariates)
else:
covariates = False
return covariates