Preserve source-bearing fractional-Z points in live projection (partial #152)

benchmark/matched_cellprofiler_batch.py

@@ -181,14 +181,46 @@ def _require_compared_output_inventory(
     compared_file_report: RuntimeEquivalenceReport = RuntimeEquivalenceReport(()),
 ) -> None:
     """Reject unqualified output formats and missing scientific output files."""
+    # Validate all explicit edges against their object rows, then compare the
+    # cross-output correlations before using redundant scalar table inventory.
+    from openhcs.core.equivalence.relationships import (
+        ExportedRelationshipMeasurementSemantics,
+    )
+    from openhcs.core.runtime_equivalence import RuntimeMeasurementSnapshot
+
+    measurements = tuple(
+        RuntimeMeasurementSnapshot.from_output_snapshot(snapshot)
+        for snapshot in (reference_snapshot, candidate_snapshot)
+    )
+    correlations = tuple(
+        measurement.required_relationship_correlations() for measurement in measurements
+    )
+    if correlations[0] != correlations[1]:
+        raise RuntimeError("Matched saved output relationship correlations differ.")
+    comparable_tables = tuple(
+        tuple(
+            table
+            for table in snapshot.tables
+            if table.participates_in_comparison
+            and not ExportedRelationshipMeasurementSemantics.supports_table(table)
+        )
+        for snapshot in (reference_snapshot, candidate_snapshot)
+    )
     counts = []
-    for files, exports, snapshot, managed_files in (
-        (reference_files, reference_exports, reference_snapshot, frozenset()),
+    for files, exports, snapshot, managed_files, tables in (
+        (
+            reference_files,
+            reference_exports,
+            reference_snapshot,
+            frozenset(),
+            comparable_tables[0],
+        ),
         (
             candidate_files,
             candidate_exports,
             candidate_snapshot,
             candidate_managed_files,
+            comparable_tables[1],
         ),
     ):
         files = files - managed_files
@@ -208,7 +240,7 @@ def _require_compared_output_inventory(
             - len(snapshot.tables)
             - len(exports.image_outputs)
             + len(snapshot.images)
-            + sum(table.participates_in_comparison for table in snapshot.tables)
+            + len(tables)
         )
         if count < 1:
             raise RuntimeError("Matched batch has no compared scientific output files.")
@@ -230,10 +262,9 @@ def _require_compared_output_inventory(
                 ),
                 len(table.rows),
             )
-            for table in snapshot.tables
-            if table.participates_in_comparison
+            for table in tables
         )
-        for snapshot in (reference_snapshot, candidate_snapshot)
+        for tables in comparable_tables
     )
     if table_shapes[0] != table_shapes[1]:
         raise RuntimeError(

benchmark/results/perf_schema_source_readiness_20261003/README.rst

@@ -0,0 +1,25 @@
+Registered source readiness: ordinary Illumination ABBA
+====================================================
+
+``receipt.json`` contains public observations, the exact variant hashes and full
+science-qualification links. The committed recipes preserve the actual local
+experiment; they are evidence, not a turnkey portable benchmark command. Their
+paths name the owned checkout and immutable maintenance directory, and the
+ordinary helper/bootstrap and physical dataset/native references must already
+exist. Do not rerun into a completed output directory.
+
+Only three source files vary. Their baseline bytes are derived from the recorded
+Git revisions in ``variants.json``. Candidate registry bytes are available at
+OpenHCS commit ``7258826d7198f2bb7010763276dba02320d62142``; candidate dependency
+bytes are available at python-introspect merge
+``6f5ac0b79d6cef163df6a8389e72a889a91ed548`` and ObjectState merge
+``55e71df39a05aa08d00bc1bb06c2be174ad2c522``. The original maintenance directory
+retains all six variant files, source freezes and four immutable output trees.
+No dependency source, benchmark images or environments are duplicated here.
+
+Mean compilation is 0.962152 → 0.587605 seconds; total is
+1.693765 → 1.363694 seconds, excluding mandatory server READY/startup and
+shutdown. Execution has no demonstrated improvement. All four complete authored
+image inventories pass exact byte parity and the retained native physical input
+qualification; no new native CP timing is claimed. Other workloads and scaling
+remain unmeasured for this change.

benchmark/results/perf_schema_source_readiness_20261003/receipt.json

@@ -0,0 +1,261 @@
+{
+  "status": "ALL4_ORDINARY_AUTHORED_IMAGES_EXACT_AND_RETAINED_NATIVE_INPUT_PARITY_PASS",
+  "case": "ExampleIlluminationCorrection_Example3",
+  "mode": "1w_1t",
+  "cpu": 5,
+  "candidate_openhcs_commit": "7258826d7198f2bb7010763276dba02320d62142",
+  "variants": [
+    {
+      "path": "/home/ts/code/projects/openhcs-shared-runtime-plumbing/openhcs/processing/backends/lib_registry/registry_service.py",
+      "baseline_revision": "4eaacf17e5de3dba59ec217f133fa6e1821c8641",
+      "repository": "/home/ts/code/projects/openhcs-shared-runtime-plumbing",
+      "relative": "openhcs/processing/backends/lib_registry/registry_service.py",
+      "A": "/home/ts/.local/state/openhcs-maintenance/20261003/schema-source-readiness-ordinary-abba-v1/0-A.source",
+      "B": "/home/ts/.local/state/openhcs-maintenance/20261003/schema-source-readiness-ordinary-abba-v1/0-B.source",
+      "A_sha256": "73e0d41f98d7ea5bb3fe076b97b97d9980b621c54f2d23b808550e00bae481d4",
+      "B_sha256": "d9e6e30b627fae7b01e08872c58ad26ee060919b6b6bd9fd9174223f2ec5a95b"
+    },
+    {
+      "path": "/home/ts/code/projects/openhcs-shared-runtime-plumbing/external/python-introspect/src/python_introspect/signature_analyzer.py",
+      "baseline_revision": "13f6ff3b43aaa430f5e771848f08d99faf941158",
+      "repository": "/home/ts/code/projects/openhcs-shared-runtime-plumbing/external/python-introspect",
+      "relative": "src/python_introspect/signature_analyzer.py",
+      "A": "/home/ts/.local/state/openhcs-maintenance/20261003/schema-source-readiness-ordinary-abba-v1/1-A.source",
+      "B": "/home/ts/.local/state/openhcs-maintenance/20261003/schema-source-readiness-ordinary-abba-v1/1-B.source",
+      "A_sha256": "51f4729861062ab857153c57d64b01bd7b0f47daa6f18fc4443819fb75c77b45",
+      "B_sha256": "7e25a0355159c86d361a21a7afc449ec758cbf8a32e9b72d06de66a924f04159"
+    },
+    {
+      "path": "/home/ts/code/projects/openhcs-shared-runtime-plumbing/external/ObjectState/src/objectstate/lazy_factory.py",
+      "baseline_revision": "b0004daf3223326ac78689ba7b2e68abc18ae298",
+      "repository": "/home/ts/code/projects/openhcs-shared-runtime-plumbing/external/ObjectState",
+      "relative": "src/objectstate/lazy_factory.py",
+      "A": "/home/ts/.local/state/openhcs-maintenance/20261003/schema-source-readiness-ordinary-abba-v1/2-A.source",
+      "B": "/home/ts/.local/state/openhcs-maintenance/20261003/schema-source-readiness-ordinary-abba-v1/2-B.source",
+      "A_sha256": "f81ce722695dbab98fd02658f2015e5cb93c1e1e69547d24249e62d896e2a828",
+      "B_sha256": "90637bdad398bae1348168e62533d69618ec965341b1b27057292e709c2a1f85"
+    }
+  ],
+  "means": {
+    "A": {
+      "compile_seconds": 0.9621515274047852,
+      "execute_seconds": 0.38265538215637207,
+      "total_seconds": 1.693764767376706
+    },
+    "B": {
+      "compile_seconds": 0.5876049995422363,
+      "execute_seconds": 0.40058863162994385,
+      "total_seconds": 1.3636944827158004
+    }
+  },
+  "saved_seconds": {
+    "compile_seconds": 0.37454652786254883,
+    "execute_seconds": -0.017933249473571777,
+    "total_seconds": 0.3300702846609056
+  },
+  "runs": [
+    {
+      "label": "A",
+      "public_rows": [
+        {
+          "case_name": "ExampleIlluminationCorrection_Example3",
+          "mode_name": "1w_1t",
+          "worker_count": "1",
+          "well_count": "1",
+          "compile_seconds": "0.9799041748046875",
+          "prepare_seconds": "0.0",
+          "execute_seconds": "0.38725757598876953",
+          "total_seconds": "1.7173671890050173",
+          "wells_per_second": "2.5822606502835725",
+          "successful_wells": "1",
+          "native_single_sample_execution_seconds": "",
+          "projected_native_execution_seconds": "",
+          "projected_execution_speedup": "",
+          "peak_memory_mb": "1011.84765625",
+          "status": "success",
+          "memory_limit_mb": "",
+          "error_message": "",
+          "execution_route": "ordinary-zmq-outcomes-v1",
+          "server_lifecycle": "reused-per-sweep",
+          "run_input_sha256": "a37695e774a71b04c3e2a390df76245c973458380fca25a9aac513dbf5e8be8e"
+        }
+      ],
+      "receipt_sha256": "6805ee89e2872aea6199e959e53b956ef144e2ff82516fbf6ac68042450d6fd2",
+      "source_freeze_sha256": "d71dcaef08a9eb2ac12a00b5e004aedfcd9b9606d860e6a8bca5da73baa9900b"
+    },
+    {
+      "label": "B",
+      "public_rows": [
+        {
+          "case_name": "ExampleIlluminationCorrection_Example3",
+          "mode_name": "1w_1t",
+          "worker_count": "1",
+          "well_count": "1",
+          "compile_seconds": "0.593247652053833",
+          "prepare_seconds": "0.0",
+          "execute_seconds": "0.42087531089782715",
+          "total_seconds": "1.4185294113121927",
+          "wells_per_second": "2.3760006208650304",
+          "successful_wells": "1",
+          "native_single_sample_execution_seconds": "",
+          "projected_native_execution_seconds": "",
+          "projected_execution_speedup": "",
+          "peak_memory_mb": "1043.54296875",
+          "status": "success",
+          "memory_limit_mb": "",
+          "error_message": "",
+          "execution_route": "ordinary-zmq-outcomes-v1",
+          "server_lifecycle": "reused-per-sweep",
+          "run_input_sha256": "3c135c9fa72e1a8ec100dca796e96c7f6721cccfab82316b3836236b6b4e75b0"
+        }
+      ],
+      "receipt_sha256": "c8f76ce4b6acaa60f5bbfe13c5a5ca924e299d5c8834e249f757dd53adc434c7",
+      "source_freeze_sha256": "ac0a8d0ad5938927dbddf01b937ef15387a15c6701d87d292ed486fbcbec7870"
+    },
+    {
+      "label": "B",
+      "public_rows": [
+        {
+          "case_name": "ExampleIlluminationCorrection_Example3",
+          "mode_name": "1w_1t",
+          "worker_count": "1",
+          "well_count": "1",
+          "compile_seconds": "0.5819623470306396",
+          "prepare_seconds": "0.0",
+          "execute_seconds": "0.38030195236206055",
+          "total_seconds": "1.3088595541194081",
+          "wells_per_second": "2.629489524807818",
+          "successful_wells": "1",
+          "native_single_sample_execution_seconds": "",
+          "projected_native_execution_seconds": "",
+          "projected_execution_speedup": "",
+          "peak_memory_mb": "1043.19140625",
+          "status": "success",
+          "memory_limit_mb": "",
+          "error_message": "",
+          "execution_route": "ordinary-zmq-outcomes-v1",
+          "server_lifecycle": "reused-per-sweep",
+          "run_input_sha256": "3c135c9fa72e1a8ec100dca796e96c7f6721cccfab82316b3836236b6b4e75b0"
+        }
+      ],
+      "receipt_sha256": "8a7f0ecb3274b47312e8ee1431b4bd75f9aaba713fd76d3e9ec28ab038aa93da",
+      "source_freeze_sha256": "d6608b629bb37ecb89d30797258d674d816857671e77fcf580bec3d62910a7a2"
+    },
+    {
+      "label": "A",
+      "public_rows": [
+        {
+          "case_name": "ExampleIlluminationCorrection_Example3",
+          "mode_name": "1w_1t",
+          "worker_count": "1",
+          "well_count": "1",
+          "compile_seconds": "0.9443988800048828",
+          "prepare_seconds": "0.0",
+          "execute_seconds": "0.3780531883239746",
+          "total_seconds": "1.6701623457483947",
+          "wells_per_second": "2.6451304495891326",
+          "successful_wells": "1",
+          "native_single_sample_execution_seconds": "",
+          "projected_native_execution_seconds": "",
+          "projected_execution_speedup": "",
+          "peak_memory_mb": "1012.59375",
+          "status": "success",
+          "memory_limit_mb": "",
+          "error_message": "",
+          "execution_route": "ordinary-zmq-outcomes-v1",
+          "server_lifecycle": "reused-per-sweep",
+          "run_input_sha256": "a37695e774a71b04c3e2a390df76245c973458380fca25a9aac513dbf5e8be8e"
+        }
+      ],
+      "receipt_sha256": "4a804982debefcd6cd075ac16520e85036281b3eddd78cabaf86f41f2841b4c6",
+      "source_freeze_sha256": "6b40a27bf359bb67febc0af967221829a7719cac0d965704609c0d36991f608e"
+    }
+  ],
+  "tests": {
+    "python_introspect": 136,
+    "objectstate_lazy_factory": 25,
+    "openhcs_readiness": 119
+  },
+  "upstream_prs": [
+    "https://github.com/OpenHCSDev/python-introspect/pull/6",
+    "https://github.com/OpenHCSDev/ObjectState/pull/9"
+  ],
+  "science_runs": [
+    {
+      "label": "A",
+      "directory": "/home/ts/.local/state/openhcs-maintenance/20261003/schema-source-readiness-ordinary-abba-v1/0-A",
+      "receipt_sha256": "6805ee89e2872aea6199e959e53b956ef144e2ff82516fbf6ac68042450d6fd2",
+      "scientific_npy_bytes": {
+        "results/reference_image__PolynomialCorrected.npy": "8875da63a072c8f2de4f5cb707d779a064b663bf69aa0f304fe6fa74d4ff57a6",
+        "results/reference_image__ConvexHullCorrWorm.npy": "9ba1259e4ca007c06f14be6946e32a86307ec2dfb9952b73fa2cc253ccae8b3a"
+      },
+      "physical_selected_inputs": {
+        "/home/ts/.cache/openhcs/cellprofiler_examples/ExampleIlluminationCorrection/images/ADSAStaphInfection2_A01_w2247376DD-6ADD-442D-AE47-F54A05F3EA94.tif": "5a57b7557bb0da5e241b48851b82d936c7ec36ede7aaca5e77b28ac4bfebf502"
+      },
+      "image_differences": [],
+      "table_differences": [],
+      "full_physical_inventory": "PASS"
+    },
+    {
+      "label": "B",
+      "directory": "/home/ts/.local/state/openhcs-maintenance/20261003/schema-source-readiness-ordinary-abba-v1/1-B",
+      "receipt_sha256": "c8f76ce4b6acaa60f5bbfe13c5a5ca924e299d5c8834e249f757dd53adc434c7",
+      "scientific_npy_bytes": {
+        "results/reference_image__ConvexHullCorrWorm.npy": "9ba1259e4ca007c06f14be6946e32a86307ec2dfb9952b73fa2cc253ccae8b3a",
+        "results/reference_image__PolynomialCorrected.npy": "8875da63a072c8f2de4f5cb707d779a064b663bf69aa0f304fe6fa74d4ff57a6"
+      },
+      "physical_selected_inputs": {
+        "/home/ts/.cache/openhcs/cellprofiler_examples/ExampleIlluminationCorrection/images/ADSAStaphInfection2_A01_w2247376DD-6ADD-442D-AE47-F54A05F3EA94.tif": "5a57b7557bb0da5e241b48851b82d936c7ec36ede7aaca5e77b28ac4bfebf502"
+      },
+      "image_differences": [],
+      "table_differences": [],
+      "full_physical_inventory": "PASS"
+    },
+    {
+      "label": "B",
+      "directory": "/home/ts/.local/state/openhcs-maintenance/20261003/schema-source-readiness-ordinary-abba-v1/2-B",
+      "receipt_sha256": "8a7f0ecb3274b47312e8ee1431b4bd75f9aaba713fd76d3e9ec28ab038aa93da",
+      "scientific_npy_bytes": {
+        "results/reference_image__PolynomialCorrected.npy": "8875da63a072c8f2de4f5cb707d779a064b663bf69aa0f304fe6fa74d4ff57a6",
+        "results/reference_image__ConvexHullCorrWorm.npy": "9ba1259e4ca007c06f14be6946e32a86307ec2dfb9952b73fa2cc253ccae8b3a"
+      },
+      "physical_selected_inputs": {
+        "/home/ts/.cache/openhcs/cellprofiler_examples/ExampleIlluminationCorrection/images/ADSAStaphInfection2_A01_w2247376DD-6ADD-442D-AE47-F54A05F3EA94.tif": "5a57b7557bb0da5e241b48851b82d936c7ec36ede7aaca5e77b28ac4bfebf502"
+      },
+      "image_differences": [],
+      "table_differences": [],
+      "full_physical_inventory": "PASS"
+    },
+    {
+      "label": "A",
+      "directory": "/home/ts/.local/state/openhcs-maintenance/20261003/schema-source-readiness-ordinary-abba-v1/3-A",
+      "receipt_sha256": "4a804982debefcd6cd075ac16520e85036281b3eddd78cabaf86f41f2841b4c6",
+      "scientific_npy_bytes": {
+        "results/reference_image__ConvexHullCorrWorm.npy": "9ba1259e4ca007c06f14be6946e32a86307ec2dfb9952b73fa2cc253ccae8b3a",
+        "results/reference_image__PolynomialCorrected.npy": "8875da63a072c8f2de4f5cb707d779a064b663bf69aa0f304fe6fa74d4ff57a6"
+      },
+      "physical_selected_inputs": {
+        "/home/ts/.cache/openhcs/cellprofiler_examples/ExampleIlluminationCorrection/images/ADSAStaphInfection2_A01_w2247376DD-6ADD-442D-AE47-F54A05F3EA94.tif": "5a57b7557bb0da5e241b48851b82d936c7ec36ede7aaca5e77b28ac4bfebf502"
+      },
+      "image_differences": [],
+      "table_differences": [],
+      "full_physical_inventory": "PASS"
+    }
+  ],
+  "scope": "Four ordinary uninstrumented public Illumination runs ABBA, CPU5 1w_1t INLINE, default OUTCOMES and memory observer. Server READY preparation/startup and shutdown excluded. A is exact prior source bytes for three files; dependency Git HEADs identify containing checkout, file hashes identify actual baseline contents. Only source lookup, existing registered type projection and registry readiness differ. No scientific processing or numerical changes. No new native CP execution.",
+  "limits": [
+    "Two observations per variant; no execution speedup is claimed.",
+    "Authored workload exports two images and no tables.",
+    "Native comparison is retained matched CP4.2.8.1 evidence, not a fresh native run.",
+    "Ordinary pipeline total excludes mandatory server startup and shutdown."
+  ],
+  "artifact_hashes": {
+    "freeze.json": "2c5d9272f1693e1da821ad45c6db3190cd046f12f446142f9c263611f8a81da8",
+    "observations.json": "aa72dc885608831f3bbc21efc62f2739d86ec74d64f5ae56d0ee27cd129792fd",
+    "science.json": "918c36612ec20738ad2d2760da85c9e9be07e81d4cd4bf0f60ec91dfcf906488",
+    "controller.py": "6fcf1cb1008ac32ee3ceaff8fd61f6db1fdba87544e1ffee1e4b3d484ce30b90",
+    "validate_science.py": "86fd34f0ca552309a453c5c169a97b949304aa37fb642a66972f4d12f9940684"
+  },
+  "authoritative_artifacts": "/home/ts/.local/state/openhcs-maintenance/20261003/schema-source-readiness-ordinary-abba-v1",
+  "remaining_work": "No execution speedup claimed. Runtime plumbing, other workloads, fresh native/scaling sweeps and figures remain outstanding."
+}