Fix resolved configuration edge cases

[j-rausch]

What does this PR do?

Type of change: Bug fix

This is a focused follow-up to #2163, which consolidated Puzzletron's resolved configuration. It fixes edge cases in the compatibility renderer: stage-owned pruning and replacement-scoring batch sizes now update the shared data sections consumed by existing runtime paths. Resolved field records reuse their normalized serialization, and redundant validation branches are removed.

The PyYAML imports also carry the annotations required by the repository's type-checking configuration.

Testing

Adds focused unit coverage for mirroring stage batch sizes into the shared data configuration. Existing resolved-configuration tests cover snapshot serialization, compatibility rendering, and bundle validation.

Summary by CodeRabbit

• Bug Fixes
  • Experiment configuration rendering now correctly applies stage batch settings to calibration and replacement-scoring sections.
  • Configuration resolution more reliably preserves complete state information when determining effective values.
  • Improved dataset validation for Nemotron-VLM configurations by accepting valid subset selections without requiring unrelated acquisition details.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 97d85726-7da0-4d70-81f9-92a443a12adf

📥 Commits

Reviewing files that changed from the base of the PR and between 96ad385 and 6d9d617.

📒 Files selected for processing (2)

• puzzletron_setup/v2/resolved.py
• tests/unit/torch/puzzletron/test_setup_v2_resolved_config.py

🚧 Files skipped from review as they are similar to previous changes (2)

• puzzletron_setup/v2/resolved.py
• tests/unit/torch/puzzletron/test_setup_v2_resolved_config.py

---

📝 Walkthrough

Walkthrough

The update mirrors stage batch values into legacy experiment paths, serializes resolved records through to_dict(), simplifies Nemotron-VLM validation, and adds type annotations and regression coverage.

Changes

Configuration compatibility

Layer / File(s)	Summary
Experiment rendering and resolved records puzzletron_setup/v2/bundle.py, puzzletron_setup/v2/defaults.py, puzzletron_setup/v2/resolved.py, tests/unit/torch/puzzletron/test_setup_v2_resolved_config.py	Stage batch values populate legacy calibration and replacement-scoring paths. Resolved field records use to_dict(). Tests cover the mirrored values. YAML imports and test callbacks include diagnostic-suppression or return-type annotations.
Nemotron-VLM dataset validation puzzletron_setup/v2/wizard.py	The subset-selection guard no longer checks acquisition or asserts subset_selection after the guard. It still requires selectable subsets and subset metadata.

Estimated code review effort: 2 (Simple) | ~10 minutes

Possibly related PRs

• NVIDIA/Model-Optimizer#2117: Changes related experiment rendering and stage batch-value propagation.
• NVIDIA/Model-Optimizer#2163: Updates related resolved-configuration, bundle-rendering, and validation code.
• NVIDIA/Model-Optimizer#2166: Changes related Puzzletron experiment rendering and batch-value synchronization.

Suggested labels: puzzletron_v2

🚥 Pre-merge checks | ✅ 6

✅ Passed checks (6 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly describes the pull request's main change: fixing edge cases in resolved configuration handling.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Security Anti-Patterns	✅ Passed	The PR delta adds no torch.load, allow_pickle=True, trust_remote_code=True, external eval/exec, # nosec, unsafe YAML, or non-permissive dependency changes.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch jrausch/puzzletron-resolved-config-edge-cases

---

_{Comment @coderabbitai help to get the list of available commands.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 53.04%. Comparing base (f5c8138) to head (6d9d617).
⚠️ Report is 2 commits behind head on feature/puzzletron_v2.

Additional details and impacted files

@@                  Coverage Diff                   @@
##           feature/puzzletron_v2    #2169   +/-   ##
======================================================
  Coverage                  53.03%   53.04%           
======================================================
  Files                        710      710           
  Lines                      92198    92198           
======================================================
+ Hits                       48896    48902    +6     
+ Misses                     43302    43296    -6     

Flag	Coverage Δ
unit	29.32% <ø> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[github-actions]

PR Preview Action v1.8.1
Preview removed because the pull request was closed.
2026-08-12 13:14 UTC