fix: preserve zero max wait duration

[hugosmoreira]

Summary

• preserve an explicit max_wait_duration=0 when building custom-training scheduling inputs
• continue treating None as an omitted duration
• add regression coverage for omission, zero with Flex Start, zero without another scheduling option, and a positive duration

Why

The documented value 0 is a sentinel that requests an indefinite Dynamic Workload Scheduler wait. Request preparation used truthiness checks, causing explicit zero to be converted to None and silently omitted from the outgoing training pipeline.

Using explicit is not None checks keeps the API distinction between omission and zero without changing positive-duration behavior.

Fixes #7067.

Validation

• focused regression test: 4 passed, 236 deselected
• complete tests/unit/aiplatform/test_training_jobs.py module: 240 passed
• negative control: restoring the original truthiness checks makes both zero cases fail with None != "0s"
• Black 24.8.0 check passes for the changed production file
• Flake8 6.1.0 passes for both changed files
• git diff --check passes

A broader local Windows package run was not completed because the Python process encountered a native memory-access failure; no full-suite result is claimed. GitHub's Linux CI remains the broader verification environment.

[hugosmoreira]

I revalidated this on Python 3.13 in an isolated, offline WSL-backed Docker environment: all four regression cases and all 240 tests in the related module pass, both at the PR head and with the change applied to current main. Restoring the original code makes exactly the two zero cases fail. The installed-wheel regression, Black, changed-file lint and package builds also pass.

I am keeping the PR in draft while resolving the wider gates. The current [testing] dependency resolution selects Protobuf 5.29.0rc3 because of the Protobuf cap and exclusions in the resolved databricks-sdk; I did not install that set or override the constraints. Full Flake8 under Python 3.13 also reports the same five findings on the original code, PR head and current main.

Could you point me to the expected stable CI dependency set and lint environment, and help approve the pending workflows? I would like to finish those checks without broadening this fix into unrelated changes.