Missing numpy.hanning, mod, searchsorted, rollaxis, unwrap, sinc, frexp

Hi everyone,

Firstly, thank you for this awesome math library.

Secondly, during development I have across several functions I am missing coming from NumPy. I have seen that you have list of functions to implement (https://github.com/QuantStack/xtensor/projects/1), but couldn’t find some of the functions I am missing. After a brief talk on Gitter, I was suggested to write down an issue listing these missing functions:

numpy.hanning
https://docs.scipy.org/doc/numpy/reference/generated/numpy.hanning.html

Example of usage:

numpy.hanning(12)
array([0.        , 0.07937323, 0.29229249, 0.57115742, 0.82743037,
       0.97974649, 0.97974649, 0.82743037, 0.57115742, 0.29229249,
       0.07937323, 0.        ])

numpy.mod
https://docs.scipy.org/doc/numpy/reference/generated/numpy.mod.html

a = numpy.array([[1.0, 2.1, 3.9, 4.0],[-5.0, -6.2, -7.8, -8.5]])
numpy.mod(a, 0.5)
array([[0. , 0.1, 0.4, 0. ],
       [0. , 0.3, 0.2, 0. ]])

numpy.mod(a, -0.5)
array([[-0. , -0.4, -0.1, -0. ],
       [-0. , -0.2, -0.3, -0. ]])

numpy.searchsorted
https://docs.scipy.org/doc/numpy/reference/generated/numpy.searchsorted.html

Example of usage:

np.searchsorted([1,2,3,4,5], 3)
2

np.searchsorted([1,2,3,4,5], 3, side='right')
3

np.searchsorted([1,2,3,4,5], [-10, 10, 2, 3])
array([0, 5, 1, 2])

numpy.rollaxis
https://docs.scipy.org/doc/numpy/reference/generated/numpy.rollaxis.html

Example of usage:

a = np.ones((3,4,5,6))
np.rollaxis(a, 3, 1).shape
(3, 6, 4, 5)

numpy.unwrap
https://docs.scipy.org/doc/numpy/reference/generated/numpy.unwrap.html

phase = np.linspace(0, np.pi, num=5)
phase[3:] += np.pi
phase
array([ 0.        ,  0.78539816,  1.57079633,  5.49778714,  6.28318531])

np.unwrap(phase)
array([ 0.        ,  0.78539816,  1.57079633, -0.78539816,  0.        ])

numpy.sinc
https://docs.scipy.org/doc/numpy/reference/generated/numpy.sinc.html
Example of usage:

x = np.linspace(-4, 4, 41)
np.sinc(x)

array([ -3.89804309e-17, -4.92362781e-02, -8.40918587e-02,
-8.90384387e-02, -5.84680802e-02, 3.89804309e-17,
6.68206631e-02, 1.16434881e-01, 1.26137788e-01,
8.50444803e-02, -3.89804309e-17, -1.03943254e-01,
-1.89206682e-01, -2.16236208e-01, -1.55914881e-01,
3.89804309e-17, 2.33872321e-01, 5.04551152e-01,
7.56826729e-01, 9.35489284e-01, 1.00000000e+00,
9.35489284e-01, 7.56826729e-01, 5.04551152e-01,
2.33872321e-01, 3.89804309e-17, -1.55914881e-01,
-2.16236208e-01, -1.89206682e-01, -1.03943254e-01,
-3.89804309e-17, 8.50444803e-02, 1.26137788e-01,
1.16434881e-01, 6.68206631e-02, 3.89804309e-17,
-5.84680802e-02, -8.90384387e-02, -8.40918587e-02,
-4.92362781e-02, -3.89804309e-17])

numpy.frexp
https://docs.scipy.org/doc/numpy/reference/generated/numpy.frexp.html

x = np.arange(9)
y1, y2 = np.frexp(x)

y1
array([ 0. , 0.5 , 0.5 , 0.75 , 0.5 , 0.625, 0.75 , 0.875,
0.5 ])

numpy.asscalar
https://docs.scipy.org/doc/numpy/reference/generated/numpy.asscalar.html

Also, there was a discussion about adding hstack and vstack Is the decission to omit it definitive or could you reconsider the decision?

Thank you.
Best regards,
Sob3kx

Hi, thanks for the list.

numpy mod is implementd with the % operator, and xt::fmod.

I think hanning is simple to implement as a function on top of an arange/linspace.

frexp should be straightforward, too, since it exists in C++ (look at xmath to see how to define your own xfunction, or even in the docs we have a xfunction from lambda https://en.cppreference.com/w/cpp/numeric/math/frexp

https://xtensor.readthedocs.io/en/latest/api/xfunction.html#_CPPv4I0DpEN2xt21make_lambda_xfunctionEDaRR1FDpRR1E

sinc will be harder, as i can only find an implementation in boost: https://www.boost.org/doc/libs/1_51_0/boost/math/special_functions/sinc.hpp

regarding hstack & vstack, you can achieve the same functionality by passing the axis argument. it would be simple to write a wrapper if you really need it.

numpy.unwrap should be straightforward again (implement a xfunction, e.g. with the lambda method i linked above).

numpy.searchsorted
numpy.rollaxis

numpy asscalar is deprecated apparently. this is also quite simple to write as a function

template <class E>
auto asscalar(const xexpression<E>& expr)
{
    XTENSOR_ASSERT(expr.derived_cast().size() == 1);
    return xt::xscalar<typename E::value_type>(expr.derived_cast()());
}

I am not sure we want this in xtensor though (I also don't really see the point).

Ah, I wanted to write: searchsorted and rollaxis are probably a little harder to implement.

It would be really cool if you try your hand at this!

I am looking forward xt::searchsorted too!
I was playing with std::lower_bound without success. :(

xt::searchsorted seems to exist now thanks to:

Adding "digitize", "searchsorted", and "bin_items" #2037 (first appeard in 0.21.6)

Navigation Menu

Repository navigation

Missing numpy.hanning, mod, searchsorted, rollaxis, unwrap, sinc, frexp #1608

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Missing numpy.hanning, mod, searchsorted, rollaxis, unwrap, sinc, frexp #1608

Description

Activity

wolfv commented on May 21, 2019

wolfv commented on May 21, 2019

victorgabr commented on Sep 26, 2019

doronbehar commented on Oct 26, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions