Allow custom gradient and Hessian in OptimizationFunction

[sebapersson]

See #96

For the uni- and bivariate case I have updated construct_fixed_optimisation_function to handle a custom gradient and Hessian. I have also added a test (Custom_hessian.jl) where results are compared for the Rosenbrock model with and without a custom gradient and Hessian.

I do not know if this passes all the tests, as when I run the tests locally the tests are killed after some time due to excessive RAM consumption.

[codecov]

Codecov Report

Attention: 2 lines in your changes are missing coverage. Please review.

Comparison is base (2268df9) 90.41% compared to head (6501492) 91.28%.

Files	Patch %	Lines
src/problem_updates.jl	96.77%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #97      +/-   ##
==========================================
+ Coverage   90.41%   91.28%   +0.87%     
==========================================
  Files          16       16              
  Lines        1575     1629      +54     
==========================================
+ Hits         1424     1487      +63     
+ Misses        151      142       -9     

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

[sebapersson]

Due to new syntax for remaking I should also have updated the test for shift, hopefully all tests should now pass.

[DanielVandH]

Thanks for this. I'll take another close look tomorrow just to be sure, but looks good!

[DanielVandH]

Just a couple of comments. Thanks so much again. If you don't have the time to add the tests I've mentioned, we can just leave it for now - the code all looks good to me.

[DanielVandH]

Awesome job, thanks!