Work around NotImplemented Quantity.tofile (#1436)

changes/1436.outlier_detection.rst

@@ -0,0 +1 @@
+Fix bug where on_disk=True could fail due to Quantities not implementing tofile.

romancal/outlier_detection/tests/test_outlier_detection.py

@@ -129,7 +129,8 @@ def test_outlier_do_detection_write_files_to_custom_location(tmp_path, base_imag
  assert all(x.exists() for x in outlier_files_path)
 
 
-def test_find_outliers(tmp_path, base_image):
+@pytest.mark.parametrize("on_disk", (True, False))
+def test_find_outliers(tmp_path, base_image, on_disk):
  """
  Test that OutlierDetection can find outliers.
  """
@@ -158,13 +159,31 @@ def test_find_outliers(tmp_path, base_image):
  imgs[0].data[img_0_input_coords[0], img_0_input_coords[1]] = cr_value
  imgs[1].data[img_1_input_coords[0], img_1_input_coords[1]] = cr_value
 
- input_models = ModelLibrary(imgs)
+ if on_disk:
+ # write out models and asn to disk
+ for img in imgs:
+ img.save(img.meta.filename)
+ asn_dict = {
+ "asn_id": "a3001",
+ "asn_pool": "none",
+ "products": [
+ {
+ "name": "test_asn",
+ "members": [
+ {"expname": m.meta.filename, "exptype": "science"} for m in imgs
+ ],
+ }
+ ],
+ }
+ input_models = ModelLibrary(asn_dict, on_disk=True)
+ else:
+ input_models = ModelLibrary(imgs)
 
  outlier_step = OutlierDetectionStep()
  # set output dir for all files created by the step
  outlier_step.output_dir = tmp_path.as_posix()
  # make sure files are written out to disk
- outlier_step.in_memory = False
+ outlier_step.in_memory = not on_disk
 
  result = outlier_step(input_models)
 

romancal/outlier_detection/utils.py

@@ -86,7 +86,7 @@ def _median_with_resampling(
 
  weight_threshold = compute_weight_threshold(drizzled_model.weight, maskpt)
  drizzled_model.data[drizzled_model.weight < weight_threshold] = np.nan
- computer.append(drizzled_model.data, i)
+ computer.append(drizzled_model.data.value, i)
  del drizzled_model
 
  # Perform median combination on set of drizzled mosaics
@@ -173,7 +173,7 @@ def _median_without_resampling(
 
  weight_threshold = compute_weight_threshold(wht, maskpt)
 
- data_copy = model.data.copy()
+ data_copy = model.data.value.copy()
  data_copy[wht < weight_threshold] = np.nan
  computer.append(data_copy, i)