Phonetic.Languages.Filters

A module allows to change the structure of the function output for the functions of elements from RealFrac class. At the moment only the equal intervals are supported. Uses less dependencies than its former analogue package uniqueness-periods-vector-filters.

One interval used

intervalNRealFrac :: (RealFrac b, Integral c) => b -> b -> c -> b -> c

Given the minimum and maximum elements, a quantity of equal intervals, and an element in between the first two arguments (or equal to one of them), finds out the index of the interval, to which the element belongs (starting from 1). The minimum element belongs to the interval with the index 1.

unsafeTransfer1I5 :: RealFrac b => b -> b -> b -> b -> b -> b

Moves (if needed) the given value so that its result divides the new [min..max] interval in the same proportion as the starting one. Is intended to be used for the arguments satisfying some additional constraints, but they are not checked (hence, its name prefix "unsafe"). For example, the second argument must be greater than the first one, the fourth -- than the third one, and the fifth must be located in between the first two. Then the result is also located in between the third and fourth arguments similarly.

transfer1IEq3 :: RealFrac b => b -> b -> b -> b

A variant of the unsafeTransfer1I5 where the lengths of the both intervals (the old and the new ones) are equal.

Several intervals

unsafeRearrangeIG :: (RealFrac b, Integral c) => b -> b -> c -> [(c, c)] -> b -> b

Makes a complex interval-based transformation moving the value from its own interval to the corresponding by the list of tuples second element of the respective pair with the first element being the starting number of the interval (numeration of them begins at 1). The list argument must be sorted by the first argument in the ascending order. Usually, its first elements in the tuples are from the range [1..n]. Number of the intervals are given as the third argument and for many cases should not be greater than 10. There do exist several semantical constraints for the possible accurate arguments, but they are not checked. For example, the first argument must be less than the second one; the fifth argument must be located between the first two ones; the third argument must be greater than zero.

unsafeRearrangeIGV

Arguments

:: (RealFrac b, Integral c)

=> b

-> b

-> c

-> [c]

Must be not empty though this is not checked

-> b

-> b

An unzipped variant of the unsafeRearrangeIG function where the list argument is internally zipped as the second argument with the [1..n]. This allows to shorten the time of the arguments writing.

Some basic usage examples

unsafeSwapIWithMaxI

Arguments

:: (RealFrac b, Integral c)

=> b

-> b

-> c

It is expected to be greater than 0, though this is not checked.

-> c

It is expected to be less than the previous argument, but greater than 0, though this is not checked.

-> b

It is expected to lie between the first two arguments, though this is not checked.

-> b

Swaps the k-th inner interval values with the maximum one's (that is the n-th one) values.

unsafeSwapVecIWithMaxI

Arguments

:: (RealFrac b, Integral c)

=> b

-> b

-> c

It is expected to be greater than 0, though this is not checked.

-> [c]

It is expected the list to be sorted in the ascending order (indices are counted in it starting with 1 opposed to the usual behaviour for lists and are the numbers of the intervals in the range from 1 to n), and besides all the elements to be less than the previous argument, greater than 0 and to be not pairwise equal, though it is not checked.

-> b

It is expected to lie between the first two arguments, though this is not checked.

-> b

Swaps the inner intervals values (given by the list of elements of the data type that has an instance of the Integral class that represent numbers-indices starting from 1 to n) with the maximum one's (that is the n-th one) values. The list must be not empty and sorted in the ascending order, though it is not checked. Be aware that this can significantly change the density of the values and break some other properties for distributions.