"""Unique-related operations.
The :mod:`skrough.unique` module delivers helper functions for unique-related
computations. Currently all operations are simple wrappers around :func:`numpy.unique`
but they are here to provide interfaces that the rest of the code uses.
"""
from typing import Tuple
import numpy as np
[docs]def get_rows_nunique(x: np.ndarray) -> int:
"""Compute the number of unique rows.
Compute the number of unique rows. Degenerated tables are handled accordingly,
i.e., a table with no columns has 1 unique rows if only it has at least one row,
otherwise it is ``0``.
Args:
x: Input data table.
Returns:
Number of unique rows.
"""
return np.unique(x, axis=0).shape[0]
[docs]def get_uniques(values: np.ndarray) -> np.ndarray:
"""Get unique values.
Get unique elements (reported in ascending order) of the input array ``values``
along with the positions/indices in the input array for which unique values appear
for the first time.
Args:
values: Input array.
Returns:
Unique values (reported in ascending order) of the input array
Examples:
>>> get_uniques(np.array([2, 7, 1, 8, 2, 8, 1]))
array([1, 2, 7, 8])
>>> get_uniques(np.array([3, 3, 2, 1, 3]))
array([1, 2, 3])
>>> get_uniques(np.array([1, 3, 3, 2]))
array([1, 2, 3])
"""
return np.unique(values)
[docs]def get_uniques_positions(values: np.ndarray) -> np.ndarray:
"""Get positions of first occurrences of unique values.
Get positions/indices for which unique values in the input array appear for the
first time. The indices are reported in the order corresponding to the ascending
order of unique values, i.e., the first index indicates the first occurrence of the
lowest unique value, the second index indicates the first occurrence of the second
lowest unique value, etc.
Args:
values: Input array.
Returns:
The positions/indices of the input array for which unique values (reported in
ascending order) appear for the first time.
Examples:
>>> get_uniques_index(np.array([1, 2, 3]))
array([0, 1, 2])
>>> get_uniques_index(np.array([3, 2, 1]))
array([2, 1, 0])
>>> get_uniques_index(np.array([1, 1, 1]))
array([0])
>>> get_uniques_index(np.array([1, 1, 2, 1]))
array([0, 2])
>>> get_uniques_index(np.array([2, 2, 1, 2]))
array([2, 0])
>>> get_uniques_index(np.array([]))
array([])
"""
_, idx = get_uniques_and_positions(values)
return idx
[docs]def get_uniques_and_positions(values: np.ndarray) -> Tuple[np.ndarray, np.ndarray]:
"""Get unique elements of an array and indices of first occurrences of these values.
Get unique elements (reported in ascending order) of the input array ``values``
along with the positions/indices in the input array for which unique values appear
for the first time.
Args:
values: Input array.
Returns:
Result is a 2-element tuple consisted of the following elements
- unique values (reported in ascending order) of the input array
- position/indices in the input array for which unique values appear for the
first time
Examples:
>>> get_uniques_and_positions(np.array([2, 7, 1, 8, 2, 8, 1]))
(array([1, 2, 7, 8]), array([2, 0, 1, 3]))
>>> get_uniques_and_positions(np.array([3, 3, 2, 1, 3]))
(array([1, 2, 3]), array([3, 2, 0]))
>>> get_uniques_and_positions(np.array([1, 3, 3, 2]))
(array([1, 2, 3]), array([0, 3, 1]))
"""
uniques, uniques_index = np.unique(values, return_index=True)
return uniques, uniques_index
[docs]def get_uniques_and_compacted(values: np.ndarray) -> Tuple[np.ndarray, np.ndarray]:
"""Get unique values and compacted version of an input array.
Get unique elements (reported in ascending order) together with a compacted version
of the input array ``values``. The compacted array is consisted of indices of the
unique values that can be used to reconstruct the original array.
Args:
values: Input array.
Returns:
Result is a 2-element tuple consisted of the following elements
- unique values (reported in ascending order) of the input array
- compacted version of the input array in the form of indices of unique values
that can be used to reconstruct the original input array
Examples:
>>> get_uniques_and_compacted(np.array([1, 2, 3]))
(array([1, 2, 3]), array([0, 1, 2]))
>>> get_uniques_and_compacted(np.array([3, 2, 1]))
(array([1, 2, 3]), array([2, 1, 0]))
>>> get_uniques_and_compacted(np.array([33, 11, 33, 33, 22]))
(array([11, 22, 33]), array([2, 0, 2, 2, 1]))
"""
uniques, uniques_inverse = np.unique(values, return_inverse=True)
return uniques, uniques_inverse
