- relax pythran requirement

* BUG: stats.zipf: incorrect pmf values
* Several scipy.sparse array API improvements, including sparse.sparray,
a new public base class distinct from the older sparse.spmatrix class,
proper 64-bit index support, and numerous deprecations paving the way
* scipy.stats added tools for survival analysis, multiple hypothesis
* A new function was added for quasi-Monte Carlo integration, and linear
* An axes argument was added broadly to ndimage functions, facilitating
- Test in parallel (pytest-xdist)
python -Wd and check for `DeprecationWarning`s).
All convolution-based filters also now accept complex-valued inputs
* `scipy.spatial.SphericalVoronoi` now supports n-dimensional input,
* `scipy.fft` is a new submodule that supersedes the `scipy.fftpack` submodule.
or the most part, this is a drop-in replacement for ``numpy.fft`` and
* uses NumPy's conventions for real transforms (``rfft``). This means the
his is different from the output of ``fftpack`` which returned a real array
* the inverse real to real transforms (``idct`` and ``idst``) are normalized
or ``norm=None`` in thesame way as ``ifft``. This means the identity
* This submodule is based on the ``pypocketfft`` library, developed by the
* Note that `scipy.fftpack` has not been deprecated and will continue to be
aintained but is now considered legacy. New code is recommended to use
* `scipy.integrate.solve_ivp` can now return a ``y_events`` attribute
unge-Kutta method originally implemented in Fortran. Now we provide a pure
* `scipy.integrate.quad` provides better user feedback when break points are
* New boolean keyword argument ``check_finite`` for `scipy.linalg.norm`; whether
o check that the input matrix contains only finite numbers. Disabling may
* It is now possible to use linear and non-linear constraints with
* `scipy.optimize.linear_sum_assignment` has been re-written in C++ to improve
* The implementation of ``choose_conv_method`` has been updated to reflect the