Cross-elements coupling for DiscontinuousLagrange sparsity patterns.

[AbdAlazezAhmed]

It iterates over all cells twice, once to get the length of I,J and another to fill them. There is probably a better way to estimate their length.
Also, codegen here looks a bit complicated, I'm still learning how to use it.
I think explicit tests for sparsity patterns are missing (they're used inside other tests, but they themselves don't have a standalone test)

Manual tests:

• Test with a single DiscontinuousLagrange
• Test with multiple DiscontinuousLagrange
• Test with older tests to make sure nothing breaks
• Test with DiscontinuousLagrange and Lagrange
• Test coupling

Tasks:

• Write tests
• Code works fine and doesn't break anything. I hope
• Update docs

Things to check:

• push! vs size precalculation
• L2 trait

• I think I should stop removing the extra whitespaces :)

[codecov-commenter]

Codecov Report

Patch coverage: 100.00% and project coverage change: +0.14 🎉

Comparison is base (3cff601) 91.77% compared to head (1130709) 91.92%.

❗ Current head 1130709 differs from pull request most recent head 7aaada1. Consider uploading reports for the commit 7aaada1 to get more accurate results

❗ Your organization is not using the GitHub App Integration. As a result you may experience degraded service beginning May 15th. Please install the Github App Integration for your organization. Read more.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #710      +/-   ##
==========================================
+ Coverage   91.77%   91.92%   +0.14%     
==========================================
  Files          32       32              
  Lines        4718     4766      +48     
==========================================
+ Hits         4330     4381      +51     
+ Misses        388      385       -3     

Impacted Files	Coverage Δ
src/Grid/grid.jl	92.08% <ø> (ø)
src/Dofs/sparsity_pattern.jl	100.00% <100.00%> (+0.73%)	⬆️
src/interpolations.jl	96.75% <100.00%> (+0.21%)	⬆️

... and 1 file with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[fredrikekre]

Nice. Didn't look at the details, but the code is surprisingly short.

It iterates over all cells twice, once to get the length of I,J and another to fill them.

Probably you can just push! then instead? Would be interesting to compare the two approaches. push! is generally pretty fast, but if you push into an already large vector it might have to be reallocated, which is expensive. That is why we currently make an estimate.

[fredrikekre]

Some comments.

[termi-official]

Can you test and clarify the case when we use discontinuous interpolations with shared dofs? (e.g. Crouzeix-Raviart)