Patch for v1.0.6

[JoeWaddell]

Context

This PR contains 3 patches:

1. Prevents error generated in farmgrown feed emissions methods for the edge case when simulating fields without crops being grown.
2. Prevents a divide by zero error in soil water content methods that is already fixed on dev.
3. Prevents repro Enum state error during initialization of cows entering the herd in a pregnant state. See PR Pregnant cows entering simulation not flagged as pregnant in ReproStateEnum #3263
4. Removes the deepdiff comparison of results from e2e test simulation runs from old expected results in the step prior to overwriting those results as the new expected e2e results. Implementing on dev branch via [E2E Testing] Fix update e2e bottleneck #3275.
5. Fixes an issue where the tolerance set for E2E testing results comparison was not being properly implemented. Implemented on dev branch in [E2E] E2E sensitivity patch #3026.

Test plan

• Test on edge case farms that produced the error
• Test on subset of evaluation farms to ensure behavior did not change.
• Test that e2e expected results can be updated without exorbitant runtime.
• Test that e2e testing is passing.

Input Changes

None.

Output Changes

None.

Filter

[morrowcj]

LGTM, but we should make sure to add unit tests when we add this bug fix to dev.

[KFosterReed]

I tested this on farms 2, 5, 6, and 7 and, as expected, results did not change beyond those due to random variation.

[KFosterReed]

Works as expected!

[allisterakun]

Changes look good, but some e2e tests are failing:

Summary of e2e results:

freestall_e2e results:
  Animal: Failing
  CropAndSoil: Passing
  Manure: Passing
  Feed: Passing

open_lot_e2e results:
  Animal: Failing
  CropAndSoil: Failing
  Manure: Failing
  Feed: Failing

no_animal_e2e results:
  Animal: n/a
  CropAndSoil: Passing
  Manure: Passing
  Feed: Passing

I have tried on both my Mac and Linux machine, all failing at the same modules