fix(platform): govern geospatial analysis and raster handoffs

backend/tests/test_raster_operations_service.py

@@ -1,14 +1,16 @@
 from __future__ import annotations
 
-from types import ModuleType, SimpleNamespace
+from types import SimpleNamespace
 from uuid import uuid4
 from pathlib import Path
 import importlib
+from hashlib import sha256
 
 from geoalchemy2.shape import from_shape
 from app.core.errors import AppError
 from app.models import Area, Dataset, DatasetVersion
 from app.services.raster_operations_service import RasterOperationsService
+from app.services.storage_service import StorageService
 from app.api.routes.datasets import _run_job_sync
 from shapely.geometry import box
 import pytest
@@ -688,6 +690,104 @@ def test_raster_tile_returns_manifest_payload(monkeypatch, tmp_path) -> None:
     assert payload["manifest"]["tile_server"] is None
 
 
+@pytest.mark.parametrize(
+    ("dimension", "expected"),
+    [
+        (512, [0]),
+        (513, [0, 1]),
+        (960, [0, 448]),
+        (961, [0, 448, 449]),
+    ],
+)
+def test_raster_tile_offsets_use_full_tiles_and_one_unique_edge_start(dimension, expected) -> None:
+    assert RasterOperationsService._tile_offsets(dimension, tile_size=512, step=448) == expected
+
+
+def test_raster_tile_rejects_limit_before_creating_output(monkeypatch, tmp_path) -> None:
+    project_id = uuid4()
+    dataset_id = uuid4()
+    source = tmp_path / "large-raster.tif"
+    source.write_bytes(b"source")
+    dataset = Dataset(
+        id=dataset_id,
+        project_id=project_id,
+        name="large-raster.tif",
+        dataset_type="raster",
+        source="user_upload",
+        storage_path=str(source),
+        original_filename="large-raster.tif",
+        stored_filename="large-raster.tif",
+        content_type="image/tiff",
+        size_bytes=6,
+    )
+    db = FakeSession([dataset])
+
+    class FakeSource:
+        width = 2048
+        height = 2048
+        count = 1
+        crs = None
+
+        def __enter__(self):
+            return self
+
+        def __exit__(self, exc_type, exc, tb):
+            return None
+
+    fake_rasterio = SimpleNamespace(open=lambda _path: FakeSource())
+    monkeypatch.setattr(
+        "app.services.raster_operations_service._import_rasterio",
+        lambda: (fake_rasterio, SimpleNamespace()),
+    )
+    tile_root = tmp_path / "tiles-that-must-not-exist"
+    monkeypatch.setattr(
+        StorageService,
+        "raster_tiles_root",
+        staticmethod(lambda *_args: tile_root),
+    )
+
+    with pytest.raises(AppError) as error:
+        RasterOperationsService.tile(db, dataset_id, tile_size=512, overlap=64, max_tiles=1)
+
+    assert error.value.code == "RASTER_TILE_LIMIT_EXCEEDED"
+    assert error.value.details == {"expected_tile_count": 25, "max_tiles": 1}
+    assert not tile_root.exists()
+
+
+def test_raster_tile_rejects_changed_source_bytes_before_creating_output(monkeypatch, tmp_path) -> None:
+    project_id = uuid4()
+    dataset_id = uuid4()
+    source = tmp_path / "changed-raster.tif"
+    source.write_bytes(b"changed")
+    dataset = Dataset(
+        id=dataset_id,
+        project_id=project_id,
+        name="changed-raster.tif",
+        dataset_type="raster",
+        source="user_upload",
+        storage_path=str(source),
+        original_filename="changed-raster.tif",
+        stored_filename="changed-raster.tif",
+        content_type="image/tiff",
+        size_bytes=7,
+        checksum_sha256=sha256(b"original").hexdigest(),
+        data_contract_key="raster.generic",
+    )
+    db = FakeSession([dataset])
+    tile_root = tmp_path / "tiles-that-must-not-exist"
+    monkeypatch.setattr(
+        StorageService,
+        "raster_tiles_root",
+        staticmethod(lambda *_args: tile_root),
+    )
+
+    with pytest.raises(AppError) as error:
+        RasterOperationsService.tile(db, dataset_id)
+
+    assert error.value.code == "DATASET_STORAGE_CHECKSUM_MISMATCH"
+    assert not tile_root.exists()
+
+
 def test_raster_clip_persists_derived_dataset(monkeypatch, tmp_path) -> None:
     project_id = uuid4()
     dataset_id = uuid4()
@@ -1393,4 +1493,3 @@ def test_run_job_sync_serializes_index_job_output_dataset_id(monkeypatch) -> Non
     assert result["job_type"] == "raster.ndvi"
     assert result["output_dataset_id"] == str(output_dataset_id)
     assert result["result_json"]["output_dataset_id"] == str(output_dataset_id)
-