Promote CellProfiler parity and accelerated backends

[trissim]

Summary

This PR promotes the CellProfiler compatibility path from benchmark-only smoke execution into a broader OpenHCS runtime compatibility layer with explicit runtime semantics, output equivalence, accelerated backends, parity tracking, and expanded generated-pipeline coverage.

Audit

Runtime and equivalence semantics

• Adds CellProfiler measurement dialect normalization and equivalence policy wiring.
• Expands runtime output equivalence for images, tables, measurements, reference artifacts, and cache-aware OpenHCS adapter comparisons.
• Promotes runtime artifact query/export/value semantics so measurements, object labels, relationships, and image outputs are compared through typed runtime records instead of ad hoc file checks.
• Adds source-schema workspace and source matching improvements so CellProfiler setup-module semantics map into OpenHCS components without filename smell heuristics.
• Extends execution validation and runtime cache behavior to separate OpenHCS execution from external equivalence/materialization work.

CellProfiler module compatibility

• Expands generated .cppipe conversion and binding coverage for alignment, color conversion, image math, mask objects, smooth, enhance edges, expand/shrink, filter objects, classify objects, illumination, object measurements, object relationships, and tracking.
• Adds typed settings modules for newly covered converter surfaces.
• Improves module execution handling for object-measurement rows, source-image provenance, object-label domains, stack/slice alignment, and relationship payloads.
• Fixes per-object measurement semantics where aggregate rows should not be padded as object rows, while preserving dense object-domain completion where required.

Explicit accelerated backends

• Adds openhcs.processing.backends.cellprofiler modules for alignment, classification, colocalization, illumination, image quality, intensity, intensity distribution, morphology, neighbors, outlines, region properties, relationships, secondary objects, shape, texture, thresholding, tracking, watershed, and Zernike calculations.
• Adds shared analysis-region-properties backend support.
• Makes numba a required dependency and declares CellProfiler compatibility extras separately.
• Keeps backend selection explicit through OpenHCS registry patterns; no silent backend fallback is introduced.

Benchmark tracking and tests

• Adds a parity runner script for CellProfiler .cppipe suites.
• Adds tracker docs for full-pass status, parity evidence, and backend optimization work.
• Expands unit coverage for CP adapters, generated pipeline execution, library loading, module execution, runtime adapter behavior, source schema, symbol table/settings binding, runtime equivalence, artifact queries, runtime values, materialization, morphology, track objects, expand/shrink, measure granularity, intensity distribution, and backend strategy registries.

Verification

Ran with the sibling OpenHCS virtualenv:

../openhcs/.venv/bin/python -m pytest \
  tests/unit/test_cellprofiler_adapter.py \
  tests/unit/test_cellprofiler_generated_pipeline_execution.py \
  tests/unit/test_cellprofiler_library_loading.py \
  tests/unit/test_cellprofiler_module_execution.py \
  tests/unit/test_cellprofiler_runtime_adapter.py \
  tests/unit/test_cellprofiler_source_schema.py \
  tests/unit/test_cellprofiler_symbol_table.py \
  tests/unit/test_cppipe_execution_validation.py \
  tests/unit/test_cppipe_parser.py \
  tests/unit/test_image_file_serialization.py \
  tests/unit/test_materialization_core.py \
  tests/unit/test_runner_cellprofiler_compatibility.py \
  tests/unit/test_runtime_artifact_queries.py \
  tests/unit/test_runtime_equivalence.py \
  tests/unit/test_runtime_execution_validation.py \
  tests/unit/test_runtime_exports.py \
  tests/unit/test_runtime_values.py \
  tests/unit/test_settings_binder.py \
  tests/unit/test_source_matching.py \
  tests/unit/test_source_schema_workspace.py \
  tests/unit/test_cellprofiler_morphology.py \
  tests/unit/test_cellprofiler_processing_backend.py \
  tests/unit/test_cellprofiler_strategy_registries.py \
  tests/unit/test_cellprofiler_trackobjects.py \
  tests/unit/test_expandorshrinkobjects.py \
  tests/unit/test_measuregranularity.py \
  tests/unit/test_measureobjectintensitydistribution.py \
  tests/unit/test_runtime_semantics.py \
  -q

Result:

645 passed, 1 warning in 21.73s

Follow-up Work

• Extract the .cppipe dialect compiler out of benchmark ownership into a product-facing OpenHCS interop/dialect package so benchmark, CLI, and PyQt consume the same canonical conversion service.
• Integrate PyQt Code mode with .cppipe loading alongside .py: compile .cppipe into a normal OpenHCS Pipeline, show generated OpenHCS code/form state, and expose source-schema/provenance mapping instead of a black-box CellProfiler runner.
• Add phase timing records for startup, setup, compile, execute, validation, equivalence, and cache/materialization phases, with graphable long-table output for OpenHCS vs native CellProfiler comparisons.
• Continue replacing centrosome/scikit-image hot paths with explicit numba/CuPy-capable backend strategies while preserving CellProfiler semantics and avoiding silent fallbacks.

[continue]

Learn more

---

All Green is an AI agent that automatically:

✅ Addresses code review comments

✅ Fixes failing CI checks

✅ Resolves merge conflicts

---

Unsubscribe from All Green comments