Export requested source identity from typed measurement provenance (#731)

openhcs/interop/cellprofiler/analyst_export.py

@@ -69,7 +69,7 @@
     SourceImageProvenance,
 )
 from openhcs.core.source_bindings import CompiledSourceBindingPlan
-from openhcs.core.source_metadata import SourceMetadataFields, SourceMetadataScalar
+from openhcs.core.source_metadata import SourceMetadataScalar
 from openhcs.core.source_workspace_projection import (
     VirtualWorkspacePathLookup,
     VirtualWorkspaceSourceProjectionAuthority,
@@ -87,7 +87,6 @@
     CELLPROFILER_MEASUREMENT_DIALECT,
 )
 from .module_declarations import CellProfilerModule
-from .source_metadata import CellProfilerSourceMetadataField
 
 
 class CellProfilerObjectTableMode(str, Enum):
@@ -920,11 +919,8 @@ def collect_image_provenance(
             values: dict[str, Any] = {}
             source_path = self._resolved_source_path(plane_provenance.source_path)
             if source_path is not None:
-                values[self.dialect.source_image_path_field(source_image_name).name] = (
-                    str(source_path.parent)
-                )
-                values[self.dialect.source_image_file_field(source_image_name).name] = (
-                    source_path.name
+                values.update(
+                    self.dialect.source_image_file_values(source_path, source_image_name)
                 )
                 if source_path.is_file():
                     values.update(
@@ -948,14 +944,9 @@ def collect_image_provenance(
                 field_name: value
                 for field_name, value in self.dialect.source_metadata_defaults().items()
             }
-            if source_path is not None:
-                metadata_items[
-                    CellProfilerSourceMetadataField.FILE_LOCATION.field_name
-                ] = source_path.as_uri()
-            if component_metadata is not None:
-                metadata_items.update(
-                    SourceMetadataFields.original_items(component_metadata)
-                )
+            metadata_items.update(
+                self.dialect.source_metadata_values(component_metadata, source_path)
+            )
             image_number = self.image_set_numbering.for_source_slice(
                 scope=scope,
                 provenance=provenance,

openhcs/interop/cellprofiler/database_column_dialect.py

@@ -6,6 +6,7 @@
 from dataclasses import dataclass, field, replace
 from enum import Enum
 from hashlib import md5
+from pathlib import Path
 import re
 from types import MappingProxyType
 from typing import Any, ClassVar
@@ -15,6 +16,7 @@
     ObjectCoreMeasurementFeature,
 )
 from openhcs.core.runtime_tabular_values import FieldSpec
+from openhcs.core.source_metadata import SourceMetadataFields, SourceMetadataMapping
 from openhcs.interop.cellprofiler.source_metadata import (
     CellProfilerSourceMetadataField,
 )
@@ -417,6 +419,54 @@ def source_metadata_defaults() -> dict[str, object]:
 
         return CellProfilerSourceMetadataField.static_defaults()
 
+    @staticmethod
+    def source_metadata_values(
+        metadata: SourceMetadataMapping | None,
+        source_path: Path | None,
+    ) -> dict[str, object]:
+        """Project original extraction fields and the declared source path."""
+        values: dict[str, object] = {}
+        if metadata is not None:
+            values.update(SourceMetadataFields.original_items(metadata))
+        if source_path is not None:
+            values.setdefault(
+                CellProfilerSourceMetadataField.FILE_LOCATION.field_name,
+                source_path.as_uri() if source_path.is_absolute() else str(source_path),
+            )
+        return values
+
+    @staticmethod
+    def source_acquisition_values(
+        metadata: SourceMetadataMapping | None,
+    ) -> dict[str, object]:
+        """Project declared acquisition components, never absent axes.
+
+        A represented multi-value domain is not a single acquisition coordinate.
+        This is distinct from the original extraction-field namespace.
+        """
+        if metadata is None:
+            return {}
+        return {
+            component.value: (
+                SourceMetadataFields.canonical_component_value(component, domain[0])
+                if len(domain) == 1 else None
+            )
+            for component, domain in SourceMetadataFields.component_domains(metadata).items()
+        }
+
+    def source_image_file_values(
+        self,
+        source_path: Path,
+        source_image_name: str,
+    ) -> dict[str, str]:
+        """Project a named source identity without reading its image pixels."""
+        return {
+            self.source_image_path_field(source_image_name).name: str(
+                source_path.parent
+            ),
+            self.source_image_file_field(source_image_name).name: source_path.name,
+        }
+
     def image_table_name(self) -> str:
         return f"{self.table_prefix}Per_Image"
 

openhcs/interop/cellprofiler/image_set_numbering.py

@@ -106,7 +106,7 @@ def project_measurement_rows(
         image_numbers_by_slice = self.for_source_slices(
             scope=scope,
             provenance=table.source_provenance,
-            slice_indices=projection.present_axis_values(slice_axis.value),
+            slice_indices=self.source_slices_for_measurement_table(table),
             owner=table.name,
         )
         axisless_image_number = None
@@ -116,12 +116,7 @@ def project_measurement_rows(
             ) or (0,)
             source_image_numbers = tuple(
                 dict.fromkeys(
-                    self.for_source_slices(
-                        scope=scope,
-                        provenance=table.source_provenance,
-                        slice_indices=source_plane_indices,
-                        owner=table.name,
-                    ).values()
+                    image_numbers_by_slice[index] for index in source_plane_indices
                 )
             )
             if (
@@ -145,6 +140,26 @@ def project_measurement_rows(
             axisless_value=axisless_image_number,
         )
 
+    @staticmethod
+    def source_slices_for_measurement_table(
+        table: MeasurementTable,
+    ) -> tuple[int, ...]:
+        """Return exactly the slices represented by this producer's row scope.
+
+        Axisless rows consume their declared source stack, not a guessed plate
+        grid. Preserve explicit row order before additional source planes.
+        """
+        projection = MeasurementRowsAxisProjection.from_rows(table.rows)
+        indices = projection.present_axis_values(
+            MeasurementRowAxisField.SLICE_INDEX.value
+        )
+        if not projection.has_axisless_rows(MeasurementRowAxisField.SLICE_INDEX):
+            return indices
+        source_indices = (
+            tuple(range(table.source_provenance.source_plane_count)) or (0,)
+        )
+        return tuple(dict.fromkeys((*indices, *source_indices)))
+
     def _source_identity(
         self,
         provenance: SourceImageProvenance,