Compressor Map Table Interpolation v1

[WyattGiroux]

Added the NcTblMap and ecTblMap functions for table interpolation of speed and efficiency (+ derivatives). Also added the bilinearBounded.jl file to perform the bilinear interpolation and the tftbl.jl file to generate the const map data structs.

Ran unit tests in test/runtests.jl and got all either passing or broken:

[codecov]

Codecov Report

Attention: Patch coverage is 18.50000% with 163 lines in your changes missing coverage. Please review.

Project coverage is 72.19%. Comparing base (430cfd8) to head (be84947).
Report is 46 commits behind head on main.

Files with missing lines	Patch %	Lines
src/utils/bilinearBounded.jl	0.00%	108 Missing ⚠️
src/engine/tftbl.jl	55.22%	30 Missing ⚠️
src/engine/tfmap.jl	0.00%	25 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #92      +/-   ##
==========================================
- Coverage   73.41%   72.19%   -1.22%     
==========================================
  Files          77       82       +5     
  Lines       13908    13697     -211     
==========================================
- Hits        10211     9889     -322     
- Misses       3697     3808     +111     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

---

🚨 Try these New Features:

• Flaky Tests Detection - Detect and resolve failed and flaky tests

[ngomezve]

@WyattGiroux This looks promising! Are you planning to implement the new compressor maps into tfsize() and tfoper(), the turbofan functions? It looks like right now they are still using the old maps, which explains why all tests are passed (the new functions are not being run).

[WyattGiroux]

@ngomezve, I am planning on integrating them with oper and size, but need to make some unit tests for the new functions before that.

In general, I've now swapped over to the NaturalNeighbours.jl library to do direct interpolation in julia. Now instead of using the bilinear functions I made, the TblMap functions are now calling interpolator functions from this package. The timing is slower (80 microseconds) but the maps are being very closely matched (max[RMSE] = 0.01-ish) using very few gridding sample points (~50^2). The package precompile time is around 1 minute.

Unit tests to come next, but good progress is being made

[WyattGiroux]

We're now faster than the original drela method and have hashing implemented such that the gridded map files only have to generate for 30 minutes once. Derivatives are a bit noisy; will try various samplings and gridding fineness later tonight. If all goes well, I'll add docstrings, unit tests, and TASOPT implementation tonight.

(Original drela version on top; new table interpolation version on bottom)