Encapsulate spatial searches into class

[javijv4]

Problem

Follows the discussion in #609
Right now, there are several instances in the code performing spatial search to match vertices to vertices, vertices to faces, or faces to faces between different meshes. Ideally, all these spatial search are encapsulated into classes that the different applications call. This will make the code easier to read, easier to optimize, and make it easier to implement features that require this type of search.

Solution

Create a PointLocator class to centralize all spatial searches.

From looking at the code, I believe the following are all the instances where a spatial search is performed

Query type	Current implementations
Nearest neighbor	match_faces, match_nodes (read_msh.cpp), face_match (read_files.cpp), distrn, remesh face loop (remesh.cpp), find_vtp_point_index (BoundaryCondition.cpp)
Closest centroid	uris_find_closest_face_centroid, find_closest_element_centroid (uris.cpp)
Containing element	uris_find_tetra (uris.cpp), find_n (remesh.cpp)
Nodes within distance	contact neighbor search (contact.cpp)
Point in region	uris_point_in_bbox (uris.cpp), SpatialBounds::inside_* (CepMod.cpp)

I am not sure whether all this code is currently maintained (like the functions in contact.cpp), but these applications seem easy to implement, so replacing them shouldn't be complicated.

Additional context

In #609, alternatives for using specialized packages to accelerate spatial queries were proposed. Depending on implementation complexity, we can implement them now or leave optimization of these routines for a future issue.

Code of Conduct

• I agree to follow this project's Code of Conduct and Contributing Guidelines