Make leaf differentiable and vectorized

[SCiarella]

This PR addresses #45.

It follows the test structure of #55 to ensure that leaf parameters are differentiable and vectorizable.
Before merging, the issues with the vectorization of drv should be addressed.

[SCiarella]

In order to work with vectorized drv data, this PR implements _ensure_shape_finalized which is called inside calc_rates. The function has the role of reshaping the parameters if multiple temperatures are provided. Similarly to all the other parameters, we first ensure consistency between vectorized shapes, so all the parameters + drv.TEMP can only be scalar or the same vector shape.

This PR also implements the function utils._check_drv_shape that ensures consistency within every drv instance. Once the weather module is updated and vector shape checks are included in the main class, this function can be removed.

[SarahAlidoost]

@SCiarella thanks for implementation and tests 👍 . Please have a look at my comments. The main issue with the current implementation is that the physical models shouldn’t be worry about the shape or format of their inputs. At this stage, we assume the physical model receives correctly shaped data, so it can do vectorized computations and proper gradient tracking. Everything else should be handled by the Engine. The physical model can include basic checks, but it shouldn’t go more than that.

If anything in my comments is unclear, please let me know, we can discuss the details.

[SarahAlidoost]

@SCiarella looks great! thanks for addressing my comments. Just one test needed for weather data as 2d array, as we discussed :)

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
100.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud