Skip to content

VALIDATION.md

Latest commit

 

History

History
43 lines (33 loc) · 4.7 KB

VALIDATION.md

File metadata and controls

43 lines (33 loc) · 4.7 KB
Raw

zReg Validation Record

Backfilled 2026-05-13 as part of Phase 12 (CARRY-05). Each row records the validation state at the end of the named phase: the completion date, the passing test count at that point, and a brief description of what was validated. Covers Phases 6 through Phase 11.1 (all v1.1 milestone phases); Phases 1–5 (v1.0 Consolidation) are out of scope per D-06.

Phase Name Completed Tests Passing Summary
6 Python 3.12 Migration 2026-04-14 all passing Raised the Python version floor to 3.12, converted all Union/Optional type hints to PEP 604 X | Y syntax, replaced typing-module generics with built-in equivalents, and confirmed full-codebase AST validity under Python 3.12.
7 CPD Deep Restructure 2026-04-20 all passing Extracted the CPD implementation from a monolithic module into a structured package (cpd/) with an abstract base class, typed result namedtuples, kernel utilities, and four concrete variant classes (Rigid, Affine, NonRigid, ConstrainedNonRigid); UAT confirmed 5/5 CPD acceptance tests passed.
8 DTW Deep Restructure 2026-04-23 all passing Migrated DynamicTimeWarping from a flat module into a dtw/ package with core.py, result.py, and constraints.py; the original dtw.py was removed, backwards-compatible import paths were verified, and compose_constraints() was covered by 4 dedicated tests.
9 Distance & Transform Restructure 2026-04-27 all passing Replaced star imports in distances/__init__.py with explicit __all__, created the DistanceMetric Protocol for internal typing, and added comprehensive NumPy-style docstrings to BaseWD, _sanitize_pairwise_distance_matrix, and all kernel-based transform classes.
10 Code Quality & Verification 2026-04-29 all passing Fixed three silent-failure bugs (minkowski p parameter, normalize zero-range axis, CPD use_color guard), replaced import-time PyTorch global side effects with an opt-in configure_pytorch() function, and inventoried all remaining TODO(deferred) markers for v1.2+ planning.
11 Validate Refactoring and Fix Coverage 2026-05-12 365 passing Removed the zero-caller _preserve_labels dead function, added 10 targeted tests covering 8 previously-uncovered branches in downsampling.py, and raised downsampling.py coverage from 85% to 93%; full regression suite confirmed 365 passing tests, 0 failures.
11.1 Close DTW-02: consistent metric variant interface 2026-05-13 107 passing Promoted DistanceMetric Protocol to the public API (from zreg.distances import DistanceMetric), wired a callable() early-continue pass-through into _sanitize_pairwise_distance_matrix, widened type annotations in DynamicTimeWarping and create_pairwise_distance_matrix, and verified 4 new end-to-end callable-metric tests pass alongside 107 total tests.

Data Sources

Each row is sourced from the corresponding plan SUMMARY.md file under .planning/phases/:

  • Phase 6: .planning/phases/06-python-3-12-migration/06-01-SUMMARY.md, 06-02-SUMMARY.md
  • Phase 7: .planning/phases/07-cpd-deep-restructure/07-01-SUMMARY.md, 07-02-SUMMARY.md, 07-03-SUMMARY.md
  • Phase 8: .planning/phases/08-dtw-deep-restructure/08-01-SUMMARY.md, 08-02-SUMMARY.md
  • Phase 9: .planning/phases/09-distance-transform-restructure/09-01-SUMMARY.md, 09-02-SUMMARY.md
  • Phase 10: .planning/phases/10-code-quality-verification/10-01-SUMMARY.md, 10-02-SUMMARY.md, 10-03-SUMMARY.md
  • Phase 11: .planning/phases/11-validate-refactoring-and-fix-coverage/11-01-SUMMARY.md and 11-VERIFICATION.md
  • Phase 11.1: .planning/phases/11.1-close-dtw-02/11.1-01-SUMMARY.md

Notes

Test counts for Phases 6–10: Runtime test execution was not available during these phases due to the open3d dependency being absent in the executor environment. Verification was performed via Python AST syntax checking (py_compile) and code-inspection pattern greps. "all passing" means all verifiable checks passed; no runtime failures were observed in subsequent phases that ran the full suite. The Phase 11 full regression run (365 passing, 0 failures) retroactively confirms no regressions from Phases 6–10.

Test count for Phase 11: Sourced from 11-VERIFICATION.md criterion SC-1: "365 passed, 0 failures" across the full repository.

Test count for Phase 11.1: Sourced from 11.1-01-SUMMARY.md: "107 passed, 1 skipped (CUDA), 0 failed" in tests/test_dtw.py and tests/test_pairwise_distance_matrix.py (the two modules exercised by this phase).

Date discrepancy — Phase 6: ROADMAP shows 2026-04-13; both SUMMARY files record completed: 2026-04-14. The SUMMARY date is used as it reflects the actual commit timestamp.