Interpolation point error

[oriolcg]

Fixing issue #727, closing #727 .

[codecov-commenter]

Codecov Report

Merging #729 (6838ed5) into master (bf394f5) will increase coverage by 0.32%.
The diff coverage is 98.03%.

@@            Coverage Diff             @@
##           master     #729      +/-   ##
==========================================
+ Coverage   88.07%   88.40%   +0.32%     
==========================================
  Files         132      132              
  Lines       15038    15459     +421     
==========================================
+ Hits        13245    13666     +421     
  Misses       1793     1793              

Impacted Files	Coverage Δ
src/CellData/CellFields.jl	88.19% <98.03%> (+0.19%)	⬆️
src/Arrays/CachedArrays.jl	73.17% <0.00%> (-1.19%)	⬇️
src/FESpaces/FESpacesWithConstantFixed.jl	75.58% <0.00%> (-0.67%)	⬇️
src/Geometry/Grids.jl	87.50% <0.00%> (-0.27%)	⬇️
src/Arrays/Tables.jl	94.97% <0.00%> (-0.27%)	⬇️
src/Arrays/Autodiff.jl	100.00% <0.00%> (ø)
src/Arrays/SubVectors.jl	100.00% <0.00%> (ø)
src/Arrays/AlgebraMaps.jl	100.00% <0.00%> (ø)
src/Helpers/GridapTypes.jl	100.00% <0.00%> (ø)
src/Fields/InverseFields.jl	100.00% <0.00%> (ø)
... and 80 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update bf394f5...6838ed5. Read the comment docs.

[fverdugo]

@Balaje are you happy with this?

[oriolcg]

@fverdugo, I did an update trying to go in the direction that @Balaje suggested. I now use the searchmethod in Interpolable to define an optional argument with the number of nearest vertices (default = 1). This is much cleaner, this optional argument can be changed directly from the user driver (if needed) as follows:

sm=KDTreeSearch(num_nearest_vertices=10)
uₕ = interpolate_everywhere(Interpolable(uh_0;searchmethod=sm),U)

[Balaje]

@oriolcg @fverdugo Yes, this is what I had in mind. It gives the user an option to provide a way to set the search level in case the algorithm misses points like this.

[fverdugo]

@oriolcg Can you fabricate a simple example and add it that reproduces the bug that this PR wants to fix?

[oriolcg]

@fverdugo, I added the test in 6838ed5. The PR is ready