[feature] Create benchmarking cases

Summary

Create benchmarking cases to compare different versions of PASE.

Current behavior

There is no benchmarking.

Desired behavior

We should create several benchmarking cases to test how modifications to the code influence results.

For example, issue #117 changes how diffuse irradiance is computed ; we should build benchmarking cases to analyze and validate changes made to diffuse.

The test cases should reflect either :

• configurations where we can evaluate the expected result exactly or at least have a precise idea of the expected result ;
• experimental measurements

To do:

• identify how to build a test case (input solar irradiance data, custom geometry, etc.) --> Joran had an interesting approach for this with the diffusers (see line 123 in main.py at this commit)
• identify how to store results of the benchmarking (table with result for each version, one table per version, ... ? )
• build benchmarking cases for :
  • sun position ?
  • diffuse irradiance
  • direct irradiance
  • crop models
    • SIMPLE yield
    • GrasSim yield
    • STICS yield
  • PV production

Sources

n/a

Linked features or branches

Symbiosyst deliverable 2.2 - see section 4 (pp. 37 et seq.)