[Cross Validation]Implement cross validation rules for daily spread inputs

[matthew7838]

Adds cross-validation rules for the manure schedule daily_spread block so configurations that turn daily spreading on but silently apply no manure are rejected at input validation instead of running to completion with no manure applied.

Context

Issue(s) closed by this pull request: closes #3097

What

• Eight new blocks in input/metadata/cross_validation/crop_and_soil_cross_validation.json, four rules written once for manure_schedule_1 and once for manure_schedule_2, matching how the existing field_1/field_2 and soil_1/soil_2 rules in that file are duplicated:
  • Daily spreading requires a DailySpread storage. When a schedule has daily_spread.is_daily_spreading: true, manure_management.storage must contain at least one entry with processor_type: "DailySpread".
  • spread_all_available_manure requires the master switch. When spread_all_available_manure is true, is_daily_spreading must also be true.
  • A targeted daily spread must request something. When is_daily_spreading is true and spread_all_available_manure is false, nitrogen_spread_amount + phosphorus_spread_amount must be greater than zero.
• Daily spread manure must be solid. When is_daily_spreading is true, manure_type must be "solid".
• Sets manure_type to "solid" in input/data/manure_schedule/example_daily_spread_manure_schedule.json, which previously demonstrated the exact configuration the manure type rule now rejects.

Why

Each of the four is a real misconfiguration that passes per-field schema validation today and produces a silent no-op, which is difficult to diagnose from the outputs because the run completes normally and simply shows no manure applied.

How

Each block pairs an apply_when clause that narrows it to the configuration under test with a single asserting condition, except the DailySpread storage check, which puts the constant on the left and a for_each filter on the right under is_in because the engine has no list-length operator.

Test plan

• Go into the issue and modify the inputs to match the four cases and see if errors were thrown.

Input Changes

• N/A

Output Changes

• N/A

Filter

[github-actions]

Current Coverage: 99%

Mypy errors on daily-spread-cross-validation-3097 branch: 1136
Mypy errors on dev branch: 1136
No difference in error counts

[github-actions]

Current Coverage: 99%

Mypy errors on daily-spread-cross-validation-3097 branch: 1136
Mypy errors on dev branch: 1136
No difference in error counts

[github-actions]

Current Coverage: 99%

Mypy errors on daily-spread-cross-validation-3097 branch: 1136
Mypy errors on dev branch: 1136
No difference in error counts

[github-actions]

Current Coverage: 99%

Mypy errors on daily-spread-cross-validation-3097 branch: 1136
Mypy errors on dev branch: 1136
No difference in error counts

[github-actions]

Current Coverage: 99%

Mypy errors on daily-spread-cross-validation-3097 branch: 1135
Mypy errors on dev branch: 1135
No difference in error counts

[elle-andreen]

Is this the information reported to the user in the errors when the inputs violate the CV rules?

[matthew7838]

yes

[elle-andreen]

Is this an actual rule in the code? Will the simulation fail if the user sets the manure type in the daily spread inputs to liquid? If so, it's a bit confusing that we offer an input for manure type but don't actually accept any user input for it i.e. require manure type to be solid

[matthew7838]

Well, the type is actually hard-coded in the code anyway. I am thinking that maybe we don't even have to specify manure type for daily spread.

[elle-andreen]

Looking back in our Slack conversations, Kristan and Clay's original intention was to let the user specify which type of manure the daily spread manure is (liquid or solid). Either option is ostensibly possible in reality and I believe the liquid vs. solid manure designation has some impacts on manure nutrient outcomes in C&S. I know in every other manure processor type, the user does not make this liquid vs. solid specification though, it is set based on what type of processor the manure came from, and I see DailySpread is hardcoded to ManureType.SOLID in ManureManager. I think this might've been something we missed in the original implementation that we should open a small, separate issue for (assuming it's mechanistically possible to set ManureType via user input). I can open the issue. The good news is it shouldn't hold up this PR - I think there is no CV required on the manure_type input for daily spread given that it can be either liquid or solid.

[github-actions]

Current Coverage: %

Mypy errors on daily-spread-cross-validation-3097 branch: 1135
Mypy errors on dev branch: 1135
No difference in error counts

[github-actions]

🚨 Some tests have failed.