scipy 1.15.2__cp313-cp313t-musllinux_1_2_aarch64.whl → 1.16.0__cp313-cp313t-musllinux_1_2_aarch64.whl

scipy/sparse/linalg/_eigen/arpack/arpack.py

@@ -46,7 +46,6 @@ from scipy.sparse._sputils import (
 from scipy.sparse.linalg import gmres, splu
 from scipy._lib._util import _aligned_zeros
 from scipy._lib._threadsafety import ReentrancyLock
-
 from . import _arpack
 arpack_int = _arpack.timing.nbx.dtype
 
@@ -277,9 +276,13 @@ class ArpackError(RuntimeError):
     ARPACK error
     """
 
-    def __init__(self, info, infodict=_NAUPD_ERRORS):
+    def __init__(self, info, infodict=None):
+        if infodict is None:
+            infodict = _NAUPD_ERRORS
+
         msg = infodict.get(info, "Unknown error")
-        RuntimeError.__init__(self, "ARPACK error %d: %s" % (info, msg))
+        super().__init__(f"ARPACK error {info}: {msg}")
+
 
 
 class ArpackNoConvergence(ArpackError):
@@ -313,12 +316,12 @@ class _ArpackParams:
     def __init__(self, n, k, tp, mode=1, sigma=None,
                  ncv=None, v0=None, maxiter=None, which="LM", tol=0):
         if k <= 0:
-            raise ValueError("k must be positive, k=%d" % k)
+            raise ValueError(f"k must be positive, k={k}")
 
         if maxiter is None:
             maxiter = n * 10
         if maxiter <= 0:
-            raise ValueError("maxiter must be positive, maxiter=%d" % maxiter)
+            raise ValueError(f"maxiter must be positive, maxiter={maxiter}")
 
         if tp not in 'fdFD':
             # Use `float64` libraries from integer dtypes.
@@ -508,12 +511,12 @@ class _SymmetricArpackParams(_ArpackParams):
                 self.B = M_matvec
                 self.bmat = 'G'
         else:
-            raise ValueError("mode=%i not implemented" % mode)
+            raise ValueError(f"mode={mode} not implemented")
 
         if which not in _SEUPD_WHICH:
             raise ValueError(f"which must be one of {' '.join(_SEUPD_WHICH)}")
         if k >= n:
-            raise ValueError("k must be less than ndim(A), k=%d" % k)
+            raise ValueError(f"k must be less than ndim(A), k={k}")
 
         _ArpackParams.__init__(self, n, k, tp, mode, sigma,
                                ncv, v0, maxiter, which, tol)
@@ -690,12 +693,12 @@ class _UnsymmetricArpackParams(_ArpackParams):
                 self.bmat = 'G'
                 self.OP = lambda x: self.OPa(M_matvec(x))
         else:
-            raise ValueError("mode=%i not implemented" % mode)
+            raise ValueError(f"mode={mode} not implemented")
 
         if which not in _NEUPD_WHICH:
             raise ValueError(f"Parameter which must be one of {' '.join(_NEUPD_WHICH)}")
         if k >= n - 1:
-            raise ValueError("k must be less than ndim(A)-1, k=%d" % k)
+            raise ValueError(f"k must be less than ndim(A)-1, k={k}")
 
         _ArpackParams.__init__(self, n, k, tp, mode, sigma,
                                ncv, v0, maxiter, which, tol)
@@ -977,9 +980,10 @@ class IterInv(LinearOperator):
     def _matvec(self, x):
         b, info = self.ifunc(self.M, x, tol=self.tol)
         if info != 0:
-            raise ValueError("Error in inverting M: function "
-                             "%s did not converge (info = %i)."
-                             % (self.ifunc.__name__, info))
+            raise ValueError(
+                f"Error in inverting M: function {self.ifunc.__name__} "
+                f"did not converge (info = {info})."
+            )
         return b
 
 
@@ -1024,9 +1028,10 @@ class IterOpInv(LinearOperator):
     def _matvec(self, x):
         b, info = self.ifunc(self.OP, x, tol=self.tol)
         if info != 0:
-            raise ValueError("Error in inverting [A-sigma*M]: function "
-                             "%s did not converge (info = %i)."
-                             % (self.ifunc.__name__, info))
+            raise ValueError(
+                f"Error in inverting [A-sigma*M]: function {self.ifunc.__name__} "
+                f"did not converge (info = {info})."
+            )
         return b
 
     @property
@@ -1266,7 +1271,7 @@ def eigs(A, k=6, M=None, sigma=None, which='LM', v0=None,
     n = A.shape[0]
 
     if k <= 0:
-        raise ValueError("k=%d must be greater than 0." % k)
+        raise ValueError(f"k={k} must be greater than 0.")
 
     if k >= n - 1:
         warnings.warn("k >= N - 1 for N * N square matrix. "
@@ -1422,16 +1427,19 @@ def eigsh(A, k=6, M=None, sigma=None, which='LM', v0=None,
         unspecified.  This is computed internally via a (sparse) LU
         decomposition for explicit matrices A & M, or via an iterative
         solver if either A or M is a general linear operator.
-        Alternatively, the user can supply the matrix or operator OPinv,
+        Alternatively, the user can supply the matrix or operator `OPinv`,
         which gives ``x = OPinv @ b = [A - sigma * M]^-1 @ b``.
+        Regardless of the selected mode (normal, cayley, or buckling),
+        `OPinv` should always be supplied as ``OPinv = [A - sigma * M]^-1``.
+
         Note that when sigma is specified, the keyword 'which' refers to
         the shifted eigenvalues ``w'[i]`` where:
 
-            if mode == 'normal', ``w'[i] = 1 / (w[i] - sigma)``.
+            if ``mode == 'normal'``: ``w'[i] = 1 / (w[i] - sigma)``.
 
-            if mode == 'cayley', ``w'[i] = (w[i] + sigma) / (w[i] - sigma)``.
+            if ``mode == 'cayley'``:  ``w'[i] = (w[i] + sigma) / (w[i] - sigma)``.
 
-            if mode == 'buckling', ``w'[i] = w[i] / (w[i] - sigma)``.
+            if ``mode == 'buckling'``: ``w'[i] = w[i] / (w[i] - sigma)``.
 
         (see further discussion in 'mode' below)
     v0 : ndarray, optional

scipy/sparse/linalg/_eigen/lobpcg/lobpcg.py

@@ -307,7 +307,7 @@ def lobpcg(
     ``A x = lambda x`` without constraints or preconditioning.
 
     >>> import numpy as np
-    >>> from scipy.sparse import spdiags
+    >>> from scipy.sparse import diags_array
     >>> from scipy.sparse.linalg import LinearOperator, aslinearoperator
     >>> from scipy.sparse.linalg import lobpcg
 
@@ -323,7 +323,7 @@ def lobpcg(
     the sparse diagonal matrix `A`
     of the eigenvalue problem ``A x = lambda x`` to solve.
 
-    >>> A = spdiags(vals, 0, n, n)
+    >>> A = diags_array(vals, offsets=0, shape=(n, n))
     >>> A = A.astype(np.int16)
     >>> A.toarray()
     array([[  1,   0,   0, ...,   0,   0,   0],
@@ -507,15 +507,15 @@ def lobpcg(
         if M is None:
             aux += "out"
         aux += " preconditioning\n\n"
-        aux += "matrix size %d\n" % n
-        aux += "block size %d\n\n" % sizeX
+        aux += f"matrix size {n}\n"
+        aux += f"block size {sizeX}\n\n"
         if blockVectorY is None:
             aux += "No constraints\n\n"
         else:
             if sizeY > 1:
-                aux += "%d constraints\n\n" % sizeY
+                aux += f"{sizeY} constraints\n\n"
             else:
-                aux += "%d constraint\n\n" % sizeY
+                aux += f"{sizeY} constraint\n\n"
         print(aux)
 
     if (n - sizeY) < (5 * sizeX):

scipy/sparse/linalg/_expm_multiply.py

@@ -115,7 +115,7 @@ def expm_multiply(A, B, start=None, stop=None, num=None,
     ----------
     A : transposable linear operator
         The operator whose exponential is of interest.
-    B : ndarray
+    B : ndarray, sparse array
         The matrix or vector to be multiplied by the matrix exponential of A.
     start : scalar, optional
         The starting time point of the sequence.
@@ -443,7 +443,7 @@ class LazyOperatorNormInfo:
 
     def d(self, p):
         """
-        Lazily estimate :math:`d_p(A) ~= || A^p ||^(1/p)` 
+        Lazily estimate :math:`d_p(A) ~= || A^p ||^(1/p)`
         where :math:`||.||` is the 1-norm.
         """
         if p not in self._d:
@@ -702,7 +702,12 @@ def _expm_multiply_interval(A, B, start=None, stop=None, num=None,
         m_star, s = _fragment_3_1(norm_info, n0, tol, ell=ell)
 
     # Compute the expm action up to the initial time point.
-    X[0] = _expm_multiply_simple_core(A, B, t_0, mu, m_star, s)
+    action_t0 = _expm_multiply_simple_core(A, B, t_0, mu, m_star, s)
+    if scipy.sparse.issparse(action_t0):
+        action_t0 = action_t0.toarray()
+    elif is_pydata_spmatrix(action_t0):
+        action_t0 = action_t0.todense()
+    X[0] = action_t0
 
     # Compute the expm action at the rest of the time points.
     if q <= s:

scipy/sparse/linalg/_interface.py

@@ -55,14 +55,14 @@ __all__ = ['LinearOperator', 'aslinearoperator']
 class LinearOperator:
     """Common interface for performing matrix vector products
 
-    Many iterative methods (e.g. cg, gmres) do not need to know the
-    individual entries of a matrix to solve a linear system A@x=b.
+    Many iterative methods (e.g. `cg`, `gmres`) do not need to know the
+    individual entries of a matrix to solve a linear system ``A@x = b``.
     Such solvers only require the computation of matrix vector
-    products, A@v where v is a dense vector.  This class serves as
+    products, ``A@v`` where ``v`` is a dense vector.  This class serves as
     an abstract interface between iterative solvers and matrix-like
     objects.
 
-    To construct a concrete LinearOperator, either pass appropriate
+    To construct a concrete `LinearOperator`, either pass appropriate
     callables to the constructor of this class, or subclass it.
 
     A subclass must implement either one of the methods ``_matvec``
@@ -82,17 +82,17 @@ class LinearOperator:
     Parameters
     ----------
     shape : tuple
-        Matrix dimensions (M, N).
+        Matrix dimensions ``(M, N)``.
     matvec : callable f(v)
-        Returns returns A @ v.
+        Returns returns ``A @ v``.
     rmatvec : callable f(v)
-        Returns A^H @ v, where A^H is the conjugate transpose of A.
+        Returns ``A^H @ v``, where ``A^H`` is the conjugate transpose of ``A``.
     matmat : callable f(V)
-        Returns A @ V, where V is a dense matrix with dimensions (N, K).
+        Returns ``A @ V``, where ``V`` is a dense matrix with dimensions ``(N, K)``.
     dtype : dtype
         Data type of the matrix.
     rmatmat : callable f(V)
-        Returns A^H @ V, where V is a dense matrix with dimensions (M, K).
+        Returns ``A^H @ V``, where ``V`` is a dense matrix with dimensions ``(M, K)``.
 
     Attributes
     ----------
@@ -108,17 +108,17 @@ class LinearOperator:
 
     Notes
     -----
-    The user-defined matvec() function must properly handle the case
-    where v has shape (N,) as well as the (N,1) case.  The shape of
-    the return type is handled internally by LinearOperator.
+    The user-defined `matvec` function must properly handle the case
+    where ``v`` has shape ``(N,)`` as well as the ``(N,1)`` case.  The shape of
+    the return type is handled internally by `LinearOperator`.
 
     It is highly recommended to explicitly specify the `dtype`, otherwise
     it is determined automatically at the cost of a single matvec application
-    on `int8` zero vector using the promoted `dtype` of the output.
-    Python `int` could be difficult to automatically cast to numpy integers
+    on ``int8`` zero vector using the promoted `dtype` of the output.
+    Python ``int`` could be difficult to automatically cast to numpy integers
     in the definition of the `matvec` so the determination may be inaccurate.
     It is assumed that `matmat`, `rmatvec`, and `rmatmat` would result in
-    the same dtype of the output given an `int8` input as `matvec`.
+    the same dtype of the output given an ``int8`` input as `matvec`.
 
     LinearOperator instances can also be multiplied, added with each
     other and exponentiated, all lazily: the result of these operations
@@ -322,6 +322,10 @@ class LinearOperator:
         """Default implementation of _rmatvec; defers to adjoint."""
         if type(self)._adjoint == LinearOperator._adjoint:
             # _adjoint not overridden, prevent infinite recursion
+            if (hasattr(self, "_rmatmat")
+                    and type(self)._rmatmat != LinearOperator._rmatmat):
+                # Try to use _rmatmat as a fallback
+                return self._rmatmat(x.reshape(-1, 1)).reshape(-1)
             raise NotImplementedError
         else:
             return self.H.matvec(x)
@@ -399,8 +403,7 @@ class LinearOperator:
             X = np.asanyarray(X)
 
         if X.ndim != 2:
-            raise ValueError('expected 2-d ndarray or matrix, not %d-d'
-                             % X.ndim)
+            raise ValueError(f'expected 2-d ndarray or matrix, not {X.ndim}-d')
 
         if X.shape[0] != self.shape[0]:
             raise ValueError(f'dimension mismatch: {self.shape}, {X.shape}')
@@ -548,7 +551,7 @@ class LinearOperator:
         else:
             dt = 'dtype=' + str(self.dtype)
 
-        return '<%dx%d %s with %s>' % (M, N, self.__class__.__name__, dt)
+        return f'<{M}x{N} {self.__class__.__name__} with {dt}>'
 
     def adjoint(self):
         """Hermitian adjoint.
@@ -822,22 +825,22 @@ class MatrixLinearOperator(LinearOperator):
 
     def _adjoint(self):
         if self.__adj is None:
-            self.__adj = _AdjointMatrixOperator(self)
+            self.__adj = _AdjointMatrixOperator(self.A)
         return self.__adj
 
+
 class _AdjointMatrixOperator(MatrixLinearOperator):
-    def __init__(self, adjoint):
-        self.A = adjoint.A.T.conj()
-        self.__adjoint = adjoint
-        self.args = (adjoint,)
-        self.shape = adjoint.shape[1], adjoint.shape[0]
+    def __init__(self, adjoint_array):
+        self.A = adjoint_array.T.conj()
+        self.args = (adjoint_array,)
+        self.shape = adjoint_array.shape[1], adjoint_array.shape[0]
 
     @property
     def dtype(self):
-        return self.__adjoint.dtype
+        return self.args[0].dtype
 
     def _adjoint(self):
-        return self.__adjoint
+        return MatrixLinearOperator(self.args[0])
 
 
 class IdentityOperator(LinearOperator):

scipy/sparse/linalg/_isolve/_gcrotmk.py

@@ -184,7 +184,7 @@ def _fgmres(matvec, v0, m, atol, lpsolve=None, rpsolve=None, cs=(), outer_v=(),
 def gcrotmk(A, b, x0=None, *, rtol=1e-5, atol=0., maxiter=1000, M=None, callback=None,
             m=20, k=None, CU=None, discard_C=False, truncate='oldest'):
     """
-    Solve a matrix equation using flexible GCROT(m,k) algorithm.
+    Solve ``Ax = b`` with the flexible GCROT(m,k) algorithm.
 
     Parameters
     ----------
@@ -273,7 +273,7 @@ def gcrotmk(A, b, x0=None, *, rtol=1e-5, atol=0., maxiter=1000, M=None, callback
     True
 
     """
-    A,M,x,b,postprocess = make_system(A,M,x0,b)
+    A,M,x,b = make_system(A,M,x0,b)
 
     if not np.isfinite(b).all():
         raise ValueError("RHS must contain only finite numbers")
@@ -306,7 +306,7 @@ def gcrotmk(A, b, x0=None, *, rtol=1e-5, atol=0., maxiter=1000, M=None, callback
 
     if b_norm == 0:
         x = b
-        return (postprocess(x), 0)
+        return (x, 0)
 
     if discard_C:
         CU[:] = [(None, u) for c, u in CU]
@@ -433,7 +433,8 @@ def gcrotmk(A, b, x0=None, *, rtol=1e-5, atol=0., maxiter=1000, M=None, callback
             ux = axpy(u, ux, ux.shape[0], -byc)  # ux -= u*byc
 
         # cx := V H y
-        hy = Q.dot(R.dot(y))
+        with np.errstate(invalid="ignore"):
+            hy = Q.dot(R.dot(y))
         cx = vs[0] * hy[0]
         for v, hyc in zip(vs[1:], hy[1:]):
             cx = axpy(v, cx, cx.shape[0], hyc)  # cx += v*hyc
@@ -499,4 +500,4 @@ def gcrotmk(A, b, x0=None, *, rtol=1e-5, atol=0., maxiter=1000, M=None, callback
     if discard_C:
         CU[:] = [(None, uz) for cz, uz in CU]
 
-    return postprocess(x), j_outer + 1
+    return x, j_outer + 1