Vtk exporter fix to save more than one product (#1439)

Co-authored-by: Sylwester Arabas <sylwester.arabas@agh.edu.pl>

Author: olastrz

PySDM/exporters/vtk_exporter.py

@@ -134,7 +134,7 @@ def export_products(self, particulator):
 
                 if isinstance(v, np.ndarray):
                     if v.shape == particulator.mesh.grid:
-                        payload[k] = v[:, :, np.newaxis]
+                        payload[k] = v[:, :, np.newaxis].copy()
                     else:
                         if self.verbose:
                             print(

tests/unit_tests/exporters/__init__.py

@@ -0,0 +1,57 @@
+""" checks for VTK exporter """
+
+from collections import namedtuple
+
+import numpy as np
+
+from PySDM.exporters import VTKExporter
+
+
+def test_vtk_exporter_copies_product_data(tmp_path):
+    """note: since VTK files contain unencoded binary data, we cannot use XML parsers;
+    not to introduce a new dependency to PySDM, we read the binary data with NumPy"""
+    # arrange
+    productc_filename = tmp_path / "prod"
+    sut = VTKExporter(products_filename=productc_filename)
+
+    grid = (1, 1)
+    arr = np.zeros(shape=grid, dtype=float)
+
+    incr = 666
+
+    def plusplus(arr):
+        arr += incr
+        return arr
+
+    prod = namedtuple(typename="MockProductA", field_names=("get",))(
+        get=lambda: plusplus(arr)
+    )
+
+    particulator = namedtuple(
+        typename="MockParticulator", field_names=("products", "n_steps", "dt", "mesh")
+    )(
+        n_steps=1,
+        products={
+            "a": prod,
+            "b": prod,
+        },
+        dt=0,
+        mesh=namedtuple(typename="MockMesh", field_names=("dimension", "grid", "size"))(
+            dimension=2,
+            grid=grid,
+            size=(1, 1),
+        ),
+    )
+
+    # act
+    sut.export_products(particulator)
+
+    # assert
+    offsets = (113, 129)
+    with open(str(productc_filename) + "_num0000000001.vts", mode="rb") as vtk:
+        binary_data = vtk.readlines()[14]
+        for i, off in enumerate(offsets):
+            assert (
+                np.frombuffer(binary_data[off : off + 8], dtype=np.float64)
+                == (i + 1) * incr
+            )