Merge pull request #35 from festim-dev/meshtags

extract and create meshtags

Author: jhdark

pyproject.toml

@@ -13,7 +13,7 @@ description = "Convert OpenFOAM files to dolfinx functions"
 readme = "README/md"
 requires-python = ">=3.9"
 license = { file = "LICENSE" }
-dependencies = ["fenics-dolfinx>=0.9.0", "pyvista"]
+dependencies = ["fenics-dolfinx>=0.9.0", "pyvista", "scipy"]
 classifiers = [
     "Natural Language :: English",
     "Topic :: Scientific/Engineering",

src/foam2dolfinx/helpers.py

@@ -0,0 +1,42 @@
+import numpy as np
+from scipy.spatial import cKDTree
+
+
+def tag_boundary_patch(
+    dolfinx_mesh,
+    patch_dataset,
+    patch_id,
+    tol=1e-6,
+    *,
+    tree: cKDTree,
+    facet_indices: np.ndarray,
+    facet_vertices: np.ndarray,
+):
+    """Tags the facets of a dolfinx mesh that belong to a given OpenFOAM boundary patch.
+
+    Args:
+        dolfinx_mesh: the dolfinx mesh
+        patch_dataset: the pyvista dataset for the boundary patch
+        patch_id: integer tag to assign to matched facets
+        tol: spatial tolerance for matching patch points to mesh vertices
+        tree: pre-built cKDTree on mesh geometry points
+        facet_indices: pre-computed exterior facet indices
+        facet_vertices: pre-computed 2D array (n_facets, n_verts_per_facet)
+            of vertex indices for each exterior facet
+
+    Returns:
+        tuple of (matched_facet_indices, tags) as int32 arrays
+    """
+    matched = tree.query_ball_point(patch_dataset.points, tol)
+    matched_idx = np.unique(
+        np.fromiter((i for sub in matched for i in sub), dtype=np.intp)
+    )
+
+    vertex_matched = np.zeros(len(dolfinx_mesh.geometry.x), dtype=bool)
+    if len(matched_idx):
+        vertex_matched[matched_idx] = True
+
+    facet_mask = vertex_matched[facet_vertices].all(axis=1)
+    matched_facets = facet_indices[facet_mask]
+
+    return matched_facets, np.full(len(matched_facets), patch_id, dtype=np.int32)