Initial public release

backend/tests/test_export_file_states_completeness.py

@@ -0,0 +1,125 @@
+"""A downloaded export must carry its own provenance and its own limits.
+
+Counterpart to ``test_export_provenance_member``: this reads the file the
+service actually writes, so it proves the member survives serialisation rather
+than only that the helper builds it.
+"""
+
+from __future__ import annotations
+
+import json
+from datetime import datetime, timezone
+from pathlib import Path
+from uuid import uuid4
+
+from app.models import Dataset, Export
+from app.services.export_service import ExportService
+from app.services.storage_service import StorageService
+from app.services.vector_feature_service import VectorFeatureService
+
+from tests.test_sprint107_map_selection_export import FakeSession, _govern_fixture_dataset
+
+
+BBOX = {"min_x": 4.9, "min_y": 50.9, "max_x": 5.2, "max_y": 51.2, "crs": "EPSG:4326"}
+
+
+def _dataset(dataset_id):
+    dataset = _govern_fixture_dataset(
+        Dataset(
+            id=dataset_id,
+            project_id=uuid4(),
+            name="grb-buildings.geojson",
+            dataset_type="vector",
+            source="grb",
+            source_name="grb",
+            source_version="2024-06",
+            observed_at=datetime(2024, 6, 1, tzinfo=timezone.utc),
+            status="ready",
+        )
+    )
+    return dataset
+
+
+def _selection(*, feature_count: int, total: int, truncated: bool, warning: str | None = None):
+    return {
+        "selection_bbox": BBOX,
+        "feature_count": feature_count,
+        "total_feature_count": total,
+        "limit": 250,
+        "truncated": truncated,
+        "summary": {"selection_edge_warning": warning} if warning else {},
+        "geojson": {
+            "type": "FeatureCollection",
+            "features": [
+                {
+                    "type": "Feature",
+                    "geometry": {"type": "Point", "coordinates": [5.0, 51.0]},
+                    "properties": {"vector_feature_id": f"vf-{index}"},
+                }
+                for index in range(feature_count)
+            ],
+        },
+    }
+
+
+def _export(monkeypatch, tmp_path: Path, selection) -> dict:
+    dataset_id = uuid4()
+    dataset = _dataset(dataset_id)
+    db = FakeSession({(Dataset, dataset_id): dataset})
+    export_path = tmp_path / "selection.geojson"
+    monkeypatch.setattr(StorageService, "dataset_export_path", lambda *_args: str(export_path))
+    monkeypatch.setattr(VectorFeatureService, "select_features_by_bbox", lambda *_args, **_kwargs: selection)
+
+    ExportService.export_vector_selection_geojson(db, dataset_id, BBOX, limit=250, name="selection")
+
+    assert [item for item in db.added if isinstance(item, Export)]
+    return json.loads(export_path.read_text(encoding="utf-8"))
+
+
+def test_a_truncated_export_file_states_that_it_is_partial(monkeypatch, tmp_path: Path) -> None:
+    written = _export(monkeypatch, tmp_path, _selection(feature_count=2, total=1_400, truncated=True))
+
+    provenance = written["geointel_provenance"]
+    assert provenance["complete"] is False
+    assert "1400" in provenance["completeness_note"].replace(".", "")
+    # The features are still there; the file simply no longer implies it holds
+    # everything the selection contains.
+    assert len(written["features"]) == 2
+
+
+def test_a_complete_export_file_says_so(monkeypatch, tmp_path: Path) -> None:
+    written = _export(monkeypatch, tmp_path, _selection(feature_count=2, total=2, truncated=False))
+
+    provenance = written["geointel_provenance"]
+    assert provenance["complete"] is True
+    assert provenance["completeness_note"] is None
+
+
+def test_the_file_identifies_its_source_edition(monkeypatch, tmp_path: Path) -> None:
+    written = _export(monkeypatch, tmp_path, _selection(feature_count=1, total=1, truncated=False))
+
+    provenance = written["geointel_provenance"]
+    assert provenance["source_name"] == "grb"
+    assert provenance["source_version"] == "2024-06"
+    assert provenance["observed_at"].startswith("2024-06-01")
+    assert provenance["selection_bbox"] == BBOX
+
+
+def test_selection_caveats_travel_into_the_file(monkeypatch, tmp_path: Path) -> None:
+    written = _export(
+        monkeypatch,
+        tmp_path,
+        _selection(feature_count=1, total=1, truncated=False, warning="22 objecten liggen deels buiten de selectie."),
+    )
+
+    assert written["geointel_provenance"]["warnings"] == ["22 objecten liggen deels buiten de selectie."]
+
+
+def test_the_file_remains_a_valid_feature_collection(monkeypatch, tmp_path: Path) -> None:
+    """The provenance is a foreign member, not a change to the GeoJSON shape."""
+
+    written = _export(monkeypatch, tmp_path, _selection(feature_count=1, total=1, truncated=False))
+
+    assert written["type"] == "FeatureCollection"
+    assert isinstance(written["features"], list)
+    assert written["features"][0]["type"] == "Feature"