Make phenology differntiable and vectorized

[SCiarella]

Targets issue #40.

Notes:

• the variables VERNFAC and VERNR are not part of the 'State variables' so I have excluded them from the output consistency assertion, since the engine does not store them
• not yet vectorized over CROP_START_TYPE and CROP_END_TYPE
• _send_signal needs updates now that we could be processing a vector of crops in parallel
• I am currently using a trick to pass the test of the phenology model, by considering the stopping condition only for the last input element (if it is a tensor).
• IDSL is a switch, so I have not made it differentiable
• DVSI is an initial parameter, so I have not made it differentiable

To do:

• VERNSAT, VERNBASE and VERNDVS are not yet differentiable (tests fail)
• wofost integration tests (patch) fail

[SarahAlidoost]

Targets issue #40.

Notes:

• the variables VERNFAC and VERNR are not part of the 'State variables' so I have excluded them from the output consistency assertion, since the engine does not store them

@SCiarella that's correct. I found that there is a phenology config file also in pcse, see here. The OUTPUT_VARS are listed there. Could we add "TSUM","TSUME" to our phenology config as well?

[SarahAlidoost]

Targets issue #40.

Notes:

• the variables VERNFAC and VERNR are not part of the 'State variables' so I have excluded them from the output consistency assertion, since the engine does not store them

This is in line with pcse config file 👍

• not yet vectorized over CROP_START_TYPE and CROP_END_TYPE

We can define them as numpy arrays of string and fix the if conditions. right?
see #60

• _send_signal needs updates now that we could be processing a vector of crops in parallel

can you please explain this in a separate issue that we dont forget it?
see #60

• I am currently using a trick to pass the test of the phenology model, by considering the stopping condition only for the last input element (if it is a tensor).

Need discussion.

• IDSL is a switch, so I have not made it differentiable
• DVSI is an initial parameter, so I have not made it differentiable

nice! 👍 We can add this to the documentation.

To do:

• VERNSAT, VERNBASE and VERNDVS are not yet differentiable (tests fail)

Tha's okay. As we discussed, the parameters of DVS_Phenology will be differentiable

• wofost integration tests (patch) fail

this is now fixed. see my commit.

[SarahAlidoost]

@SCiarella thanks for the nice work! 👍 please see my comments and let's have a chat about some of the changes I made 😄

[sonarqubecloud]

Quality Gate passed

Issues
8 New issues
0 Accepted issues

Measures
0 Security Hotspots
94.6% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[SarahAlidoost]

@SCiarella thanks for addressing the comments 👍