Hello, I recently added wrappers for Lapack pstrf and psft2 with the purpose of adding full pivot capabilities to the cholesky decomposition because it allows the use of positive semidefinite matrices. R has this feature and it is useful in many areas. All feedback is welcome! :)Tuxcell P.S.: my proposed changes ("is_hermitian" is currently in PR Matrix identification methods and test by tuxcell · Pull Request #10639 · scipy/scipy) | | | | | | | | | | | Matrix identification methods and test by tuxcell · Pull Request #10639 ... Adds several methods for checking matrices It implements these checks: is_hermitian, is_symmetric, is_skew_symm... | | | def _cholesky(a, lower=False, overwrite_a=False, clean=True,              check_finite=True, full_pivot=False, pivot_tol=-1):    """Common code for cholesky() and cho_factor()."""     a1 = asarray_chkfinite(a) if check_finite else asarray(a)    a1 = atleast_2d(a1)     # Dimension check    if a1.ndim != 2:        raise ValueError('Input array needs to be 2 dimensional but received '                         'a {}d-array.'.format(a1.ndim))    # Squareness check    if a1.shape[0] != a1.shape[1]:        raise ValueError('Input array is expected to be square but has '                         'the shape: {}.'.format(a1.shape))     # Quick return for square empty array    if a1.size == 0:        return a1.copy(), lower     if not is_hermitian():        raise LinAlgError("Expected symmetric or hermitian matrix")     overwrite_a = overwrite_a or _datacopied(a1, a)     # if the pivot flag is false, return the result    if not full_pivot:        potrf, = get_lapack_funcs(('potrf',), (a1,))        c, info = potrf(a1, lower=lower, overwrite_a=overwrite_a, clean=clean)        if info > 0:            raise LinAlgError("%d-th leading minor of the array is not positive "                              "definite" % info)        if info < 0:            raise ValueError('LAPACK reported an illegal value in {}-th argument'                             'on entry to "POTRF".'.format(-info))        return c, lower    else: # if the pivot flag is true, return the result plus rank and pivot         pstrf, = get_lapack_funcs(('pstrf',), (a1,))        c, pivot, rank, info=pstrf(a1, lower=False, overwrite_a=False, clean=True, tol=pivot_tol)         if info > 0:            if rank == 0:                raise LinAlgError("%d-th leading minor of the array is not positive "                              "semidefinite" % info)            else:                raise LinAlgError("The array is rank deficient with "                              "computed rank %d" % info)         if info < 0:            raise ValueError('LAPACK reported an illegal value in {}-th argument'                         'on entry to "PSTRF".'.format(-info))        return c, lower, rank, pivot def cholesky(a, lower=False, overwrite_a=False, check_finite=True, full_pivot=False, pivot_tol=-1):    """    Compute the Cholesky decomposition of a matrix.     Returns the Cholesky decomposition, :math:`A = L L^*` or    :math:`A = U^* U` of a Hermitian positive-definite matrix A.     Parameters    ----------    a : (M, M) array_like        Matrix to be decomposed    lower : bool, optional        Whether to compute the upper or lower triangular Cholesky        factorization.  Default is upper-triangular.    overwrite_a : bool, optional        Whether to overwrite data in `a` (may improve performance).    check_finite : bool, optional        Whether to check that the input matrix contains only finite numbers.        Disabling may give a performance gain, but may result in problems        (crashes, non-termination) if the inputs do contain infinities or NaNs.    full_pivot : bool, optional        Whether to use full pivot or not    pivot_tol : float, optional        Tolerance for the pivot, if < 0 then tolerance = N*U*MAX( A(M,M) )     Returns    -------    c : (M, M) ndarray        Upper- or lower-triangular Cholesky factor of `a`.     Raises    ------    LinAlgError : if decomposition fails.     Examples    --------    >>> from scipy.linalg import cholesky    >>> a = np.array([[1,-2j],[2j,5]])    >>> L = cholesky(a, lower=True)    >>> L    array([[ 1.+0.j,  0.+0.j],           [ 0.+2.j,  1.+0.j]])    >>> L @ L.T.conj()    array([[ 1.+0.j,  0.-2.j],           [ 0.+2.j,  5.+0.j]])     """    if not full_pivot:        c,lower = _cholesky(a, lower=lower, overwrite_a=overwrite_a, clean=True, check_finite=check_finite)        return c    else:        c,lower, rank_bn, piv = _cholesky(a, lower=lower, overwrite_a=overwrite_a, clean=True, check_finite=check_finite, full_pivot=full_pivot, pivot_tol=pivot_tol)        return c, rank_bn, piv