Merge pull request #257 from Deltares/scipy_cg_tol

Update pixi lockfile, laplace interpolate tol -> (rtol, atol) for cg solve

docs/changelog.rst

@@ -27,6 +27,15 @@ Added
 
 - Included ``edge_node_connectivity`` in :meth:`xugrid.Ugrid2d.from_meshkernel`, 
   so the ordering of edges is consistent with ``meshkernel``.
+
+Changed
+~~~~~~~
+
+- :meth:`xugrid.UgridDataArrayAccessor.laplace_interpolate` now uses ``rtol``
+  and ``atol`` keywords instead of ``tol``, to match changes in
+  ``scipy.linalg.sparse.cg``.
+
+[0.10.0] 2024-05-01
 
 [0.10.0] 2024-05-01
 -------------------

pyproject.toml

@@ -82,6 +82,7 @@ mapbox_earcut = "*"
 matplotlib-base = "*"
 netcdf4 = "*"
 numba_celltree = "*"
+numpy = "<2.0"
 pip = "*"
 pooch = "*"
 pre-commit = "*"

tests/test_interpolate.py

@@ -22,8 +22,8 @@ def test_ilu0():
     b[0] = 1.0
     b[-1] = 1.0
     M = interpolate.ILU0Preconditioner.from_csr_matrix(A)
-    _, info_cg = sparse.linalg.cg(A, b, tol=1.0e-5, maxiter=10)
-    x_pcg, info_pcg = sparse.linalg.cg(A, b, tol=1.0e-5, maxiter=10, M=M)
+    _, info_cg = sparse.linalg.cg(A, b, maxiter=10)
+    x_pcg, info_pcg = sparse.linalg.cg(A, b, maxiter=10, M=M)
     x_direct = sparse.linalg.spsolve(A, b)
     assert info_cg != 0  # cg does not converge
     assert info_pcg == 0  # preconditioned cg does converge

xugrid/core/dataarray_accessor.py

@@ -573,7 +573,8 @@ def laplace_interpolate(
         direct_solve: bool = False,
         delta=0.0,
         relax=0.0,
-        tol: float = 1.0e-5,
+        rtol=1.0e-5,
+        atol=0.0,
         maxiter: int = 500,
     ):
         """
@@ -606,7 +607,9 @@ def laplace_interpolate(
             ILU0 preconditioner non-diagonally dominant correction.
         relax: float, default 0.0.
             Modified ILU0 preconditioner relaxation factor.
-        tol: float, optional, default 1.0e-5.
+        rtol: float, optional, default 1.0e-5.
+            Convergence tolerance for ``scipy.sparse.linalg.cg``.
+        atol: float, optional, default 0.0.
             Convergence tolerance for ``scipy.sparse.linalg.cg``.
         maxiter: int, default 500.
             Maximum number of iterations for ``scipy.sparse.linalg.cg``.
@@ -631,7 +634,8 @@ def laplace_interpolate(
             direct_solve=direct_solve,
             delta=delta,
             relax=relax,
-            tol=tol,
+            rtol=rtol,
+            atol=atol,
             maxiter=maxiter,
         )
         da_filled = da.copy(data=filled)

xugrid/ugrid/interpolate.py

@@ -207,7 +207,8 @@ def laplace_interpolate(
     direct_solve: bool = False,
     delta=0.0,
     relax=0.0,
-    tol: float = 1.0e-5,
+    atol: float = 0.0,
+    rtol: float = 1.0e-5,
     maxiter: int = 500,
 ):
     """
@@ -235,6 +236,8 @@ def laplace_interpolate(
         ILU0 preconditioner non-diagonally dominant correction.
     relax: float, default 0.0
         Modified ILU0 preconditioner relaxation factor.
+    atol: float, optional, default 0.0
+        Convergence tolerance for ``scipy.sparse.linalg.cg``.
     tol: float, optional, default 1.0e-5.
         Convergence tolerance for ``scipy.sparse.linalg.cg``.
     maxiter: int, default 500.
@@ -301,7 +304,7 @@ def laplace_interpolate(
         # Create preconditioner M
         M = ILU0Preconditioner.from_csr_matrix(A, delta=delta, relax=relax)
         # Call conjugate gradient solver
-        x, info = sparse.linalg.cg(A, rhs, tol=tol, maxiter=maxiter, M=M, atol="legacy")
+        x, info = sparse.linalg.cg(A, rhs, rtol=rtol, atol=atol, maxiter=maxiter, M=M)
         if info < 0:
             raise ValueError("scipy.sparse.linalg.cg: illegal input or breakdown")
         elif info > 0: