feat(provenance): govern source snapshots and data inputs

backend/tests/test_sprint233_operational_completion.py

@@ -1,8 +1,9 @@
 from __future__ import annotations
 
 import json
+from datetime import UTC, datetime
+from hashlib import sha256
 from pathlib import Path
-from types import SimpleNamespace
 from uuid import uuid4
 
 import pytest
@@ -10,12 +11,13 @@ from fastapi.testclient import TestClient
 from pydantic import ValidationError
 
 from app.main import app
-from app.models import Dataset, Export
+from app.models import Dataset, Export, SourceRegistry, SourceSnapshot
 from app.schemas.export import ExportCreateResponse, MapResultExportRequest
 from app.schemas.project import ProjectRead
 from app.services.export_service import ExportService
 from app.services.project_service import ProjectService
 from app.services.storage_service import StorageService
+from app.services.source_registry_service import SourceRegistryService
 from app.services.temporal_analysis_service import TemporalAnalysisService
 from app.services.thematic_raster_analysis_service import ThematicRasterAnalysisService
 
@@ -48,6 +50,71 @@ def bbox_payload() -> dict:
     }
 
 
+def governed_dataset(
+    *,
+    project_id,
+    dataset_id,
+    name: str,
+    dataset_type: str,
+    source_key: str,
+    dataset_role: str = "source",
+) -> Dataset:
+    """Build an in-memory stand-in for a passed governed dataset.
+
+    Map-result export is an operational consumption boundary.  These tests
+    must therefore model the same source registry/snapshot, checksum and
+    passed-contract evidence supplied by a real adapter rather than relying
+    on an old transient Dataset fixture.
+    """
+
+    source = SourceRegistry(
+        id=uuid4(),
+        **SourceRegistryService.definition_for(source_key).as_model_values(),
+    )
+    checksum = sha256(f"{dataset_id}:{source_key}:{dataset_type}".encode("utf-8")).hexdigest()
+    snapshot = SourceSnapshot(
+        id=uuid4(),
+        source_registry_id=source.id,
+        snapshot_key=f"test:{source_key}:{checksum}",
+        checksum_sha256=checksum,
+        fetched_at=datetime.now(UTC),
+        crs="EPSG:31370",
+        units=source.default_units,
+        spatial_resolution_json={"x": 1.0, "y": 1.0, "unit": "m"},
+        temporal_coverage_json={"status": "test-fixture"},
+        geographic_coverage_json={"zone": "Flanders"},
+        observed_schema_json={"dataset_type": dataset_type},
+        freshness_status="current",
+        ingest_status="ingested",
+        known_limitations_json=["In-memory governed fixture used only by this export test."],
+        snapshot_metadata_json={"fixture_mode": True},
+    )
+    return Dataset(
+        id=dataset_id,
+        project_id=project_id,
+        name=name,
+        dataset_type=dataset_type,
+        source="governed test fixture",
+        dataset_role=dataset_role,
+        source_name=source.source_key,
+        source_registry_id=source.id,
+        source_snapshot_id=snapshot.id,
+        source_registry=source,
+        source_snapshot=snapshot,
+        data_contract_key=("geointel.raster.geotiff" if dataset_type == "raster" else "geointel.vector.geojson"),
+        data_contract_version="1.0.0",
+        validation_status="passed",
+        provenance_status="complete",
+        lineage_status="not_applicable",
+        quarantine_status="not_quarantined",
+        checksum_sha256=checksum,
+        metadata_json={"fixture_mode": True},
+        source_metadata={"fixture_mode": True, "source_registry_key": source.source_key},
+        provenance_metadata={"fixture_mode": True, "source_snapshot_id": str(snapshot.id)},
+        status="ready",
+    )
+
+
 def test_map_result_export_request_requires_a_complete_target() -> None:
     with pytest.raises(ValidationError):
         MapResultExportRequest(project_id=uuid4(), mode="current", bbox=bbox_payload())
@@ -67,13 +134,12 @@ def test_current_vector_map_result_uses_authoritative_selection_export(monkeypat
     project_id = uuid4()
     dataset_id = uuid4()
     area_id = uuid4()
-    dataset = Dataset(
-        id=dataset_id,
+    dataset = governed_dataset(
         project_id=project_id,
+        dataset_id=dataset_id,
         name="buildings.geojson",
         dataset_type="vector",
-        source="fixture",
-        status="ready",
+        source_key="grb",
     )
     db = FakeSession({(Dataset, dataset_id): dataset})
     expected = ExportCreateResponse(
@@ -109,14 +175,12 @@ def test_current_vector_map_result_uses_authoritative_selection_export(monkeypat
 def test_partitioned_vector_map_result_uses_governed_partition_export(monkeypatch) -> None:
     project_id = uuid4()
     dataset_id = uuid4()
-    dataset = Dataset(
-        id=dataset_id,
+    dataset = governed_dataset(
         project_id=project_id,
+        dataset_id=dataset_id,
         name="vha-municipality.geojson",
         dataset_type="vector",
-        source="VHA",
-        source_name="vmm_vha_bathymetry_profiles",
-        status="ready",
+        source_key="vmm_vha_bathymetry_profiles",
     )
     db = FakeSession({(Dataset, dataset_id): dataset})
     expected = ExportCreateResponse(
@@ -159,14 +223,12 @@ def test_partitioned_vector_map_result_uses_governed_partition_export(monkeypatc
 def test_raster_map_result_is_recomputed_and_persisted(tmp_path, monkeypatch) -> None:
     project_id = uuid4()
     dataset_id = uuid4()
-    dataset = Dataset(
-        id=dataset_id,
+    dataset = governed_dataset(
         project_id=project_id,
+        dataset_id=dataset_id,
         name="space-occupation.tif",
         dataset_type="raster",
-        source="official",
-        source_name="department_omgeving_thematic_raster",
-        status="ready",
+        source_key="department_omgeving_thematic_raster",
     )
     db = FakeSession({(Dataset, dataset_id): dataset})
     export_path = tmp_path / "space-occupation-analysis.json"
@@ -209,7 +271,26 @@ def test_evolution_map_result_is_recomputed_and_persisted(tmp_path, monkeypatch)
     project_id = uuid4()
     earlier_id = uuid4()
     later_id = uuid4()
-    db = FakeSession()
+    earlier_dataset = governed_dataset(
+        project_id=project_id,
+        dataset_id=earlier_id,
+        name="forest-earlier.geojson",
+        dataset_type="vector",
+        source_key="inbo_bwk_natura2000",
+    )
+    later_dataset = governed_dataset(
+        project_id=project_id,
+        dataset_id=later_id,
+        name="forest-later.geojson",
+        dataset_type="vector",
+        source_key="inbo_bwk_natura2000",
+    )
+    db = FakeSession(
+        {
+            (Dataset, earlier_id): earlier_dataset,
+            (Dataset, later_id): later_dataset,
+        }
+    )
     export_path = tmp_path / "forest-evolution.json"
     captured: dict = {}