ENH: Function Validation Rework & Swap np.searchsorted to bisect_left

[MateusStano]

Pull request type

• Code changes (bugfix, features)
• Code maintenance (refactoring, formatting, tests)

Checklist

• Docs have been reviewed and added / updated
• Lint (black rocketpy/ tests/) has passed locally
• All tests (pytest tests -m slow --runslow) have passed locally

Current behavior

1. The input parameters of the Funciton class are all checked in a very disorganized way. This would lead to redundant checks and overall slowing down of the class
2. all interpolations used in get_value_opt were using np.searchsorted, which was slow

New behavior

1. Validation has been reworked as to follow the order: source -> interpolation and extrapolation -> inputs, outputs and title. The get_value_opt function is now modularized and is defined according to the current interpolation and extrapolation
2. bisect_left is now used instead of np.searchsorted. This cuts the execution time by half in a u_dot_generalized call.

The Function class should behave exactly the same as before.

Breaking change

• Yes
• No

Additional information

• There is an attribute called __img_dim__ that was never used and was always equal to 1. I changed it so its defined at the __init__ for backward compatibility purposes, but it is still never used
• Further optimizations can be done in __mul__, __sub__, __truediv__ and other operations by avoiding going through all the validation checks when re-initializing the Functions

[Gui-FernandesBR]

@MateusStano I made all the commentaries that I found important to be done in this first review.
Overall, code is much better and intuitive. Let's follow the discussions in the next days so we can merge this PR really soon.

[Gui-FernandesBR]

Also, @MateusStano can we make all the new methods completely private (i.e. use def __your_private_method) instead of using a single underscore?

These new methods won't be used out of the Function class, plus using double score seems better, it allows python to mangle the methods.

[codecov]

Codecov Report

Attention: Patch coverage is 88.77005% with 21 lines in your changes are missing coverage. Please review.

❗ No coverage uploaded for pull request base (mnt/flight-simulation-speed-up@939d294). Click here to learn what that means.

❗ Current head 29bb5fa differs from pull request most recent head 00a3c1f. Consider uploading reports for the commit 00a3c1f to get more accurate results

Files	Patch %	Lines
rocketpy/mathutils/function.py	88.77%	21 Missing ⚠️

Additional details and impacted files

@@                        Coverage Diff                        @@
##             mnt/flight-simulation-speed-up     #582   +/-   ##
=================================================================
  Coverage                                  ?   72.86%           
=================================================================
  Files                                     ?       59           
  Lines                                     ?     9547           
  Branches                                  ?        0           
=================================================================
  Hits                                      ?     6956           
  Misses                                    ?     2591           
  Partials                                  ?        0           

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

[Gui-FernandesBR]

@MateusStano good work solving the majority of commentaries!

Could you add tests for the lines missing coverage?

[Gui-FernandesBR]

Also, @MateusStano could you please add your PR to the CHANGELOG?