pflm.utils.FlattenFunctionalData

class FlattenFunctionalData(y: ndarray, t: ndarray, w: ndarray, tid: ndarray, unique_tid: ndarray, inverse_tid_idx: ndarray, sid: ndarray, unique_sid: ndarray, sid_cnt: ndarray)

Bases: object

Flattened functional dataset in 1D arrays with indexing helpers.

Attributes:

ynp.ndarray of shape (M,)

Flattened responses after removing NaN.

tnp.ndarray of shape (M,)

Flattened time points aligned to y.

wnp.ndarray of shape (M,)

Per-observation weights expanded from sample-level weights.

tidnp.ndarray of shape (M,)

Integer indices mapping each t to its position in unique_tid.

unique_tidnp.ndarray of shape (nt,)

Sorted unique time points (observation grid).

inverse_tid_idxnp.ndarray of shape (M,)

Inverse mapping indices from t back to unique_tid.

sidnp.ndarray of shape (M,)

Sample id for each observation (0-based).

unique_sidnp.ndarray of shape (n_samples,)

Unique sample ids present in the data.

sid_cntnp.ndarray of shape (n_samples,)

Number of observations for each sample id.

Examples

>>> import numpy as np
>>> from pflm.utils.utility import flatten_and_sort_data_matrices
>>> y = [np.array([1.0, 2.0]), np.array([3.0, 4.0, 5.0])]
>>> t = [np.array([0.0, 1.0]), np.array([0.0, 0.5, 1.0])]
>>> ffd = flatten_and_sort_data_matrices(y, t)
>>> ffd.y.shape
(5,)
>>> ffd.unique_sid.shape
(2,)

inverse_tid_idx: ndarray

sid: ndarray

sid_cnt: ndarray

t: ndarray

tid: ndarray

unique_sid: ndarray

unique_tid: ndarray

w: ndarray

y: ndarray