jit_scorers

corsort.jit_scorers.jit_scorer_delta(n, downs, ups)

Estimates scores of nodes by the difference between the numbers of descendants and ascendants. The rational is to consider that an item should be in average halfway between its highest and lowest possible values.

Parameters:

• n (int) – Number of items to sort

• downs (ndarray) – Indices of the low value of a sequence of performed comparisons

• ups (ndarray) – Indices of the high value of a sequence of performed comparisons

Returns:

An array that represents, for each new performed comparison, some item scores that estimate the final position of items.

Return type:

ndarray

Examples

>>> my_n = 5
>>> my_downs = np.array([0, 1, 2])
>>> my_ups = np.array([2, 2, 3])
>>> jit_scorer_delta(my_n, my_downs, my_ups)
array([[ 0,  0,  0,  0,  0],
       [-1,  0,  1,  0,  0],
       [-1, -1,  2,  0,  0],
       [-2, -2,  1,  3,  0]])

corsort.jit_scorers.jit_scorer_rho(n, downs, ups)

Estimates scores of nodes by dividing the number of the descendants by the size of the family plus one. The rational is to consider that the family should be in average evenly spaced in the sorted result.

Parameters:

• n (int) – Number of items to sort

• downs (ndarray) – Indices of the low value of a sequence of performed comparisons

• ups (ndarray) – Indices of the high value of a sequence of performed comparisons

Returns:

An array that represents, for each new performed comparison, some item scores that estimate the final position of items.

Return type:

ndarray

Examples

>>> my_n = 5
>>> my_downs = np.array([0, 1, 2])
>>> my_ups = np.array([2, 2, 3])
>>> jit_scorer_rho(my_n, my_downs, my_ups)
array([[0.5       , 0.5       , 0.5       , 0.5       , 0.5       ],
       [0.33333333, 0.5       , 0.66666667, 0.5       , 0.5       ],
       [0.33333333, 0.33333333, 0.75      , 0.5       , 0.5       ],
       [0.25      , 0.25      , 0.6       , 0.8       , 0.5       ]])