Apply the fix to noadata issue at land cover level3

odc/stats/plugins/l34_utils/lc_level3.py

@@ -8,35 +8,31 @@
 def lc_level3(xx: xr.Dataset):
 
     # Cultivated pipeline applies a mask which feeds only terrestrial veg (110) to the model
-    # Just exclude no data (255) and apply the cultivated results
+    # Just exclude no data (255 or nan) and apply the cultivated results
+    # 255: load with product definition; nan: load without
+    # hence accormmodate both
+
     res = expr_eval(
-        "where(a<nodata, a, b)",
+        "where((a!=a)|(a>=nodata), b, a)",
         {"a": xx.cultivated_class.data, "b": xx.classes_l3_l4.data},
         name="mask_cultivated",
         dtype="float32",
-        **{"nodata": NODATA},
+        **{"nodata": xx.cultivated_class.attrs.get("nodata")},
     )
 
     # Mask urban results with bare sfc (210)
+
     res = expr_eval(
         "where(a==_u, b, a)",
         {
             "a": res,
             "b": xx.urban_classes.data,
         },
         name="mark_urban",
-        dtype="uint8",
+        dtype="float32",
         **{"_u": 210},
     )
 
-    # Add intertidal as water
-    res = expr_eval(
-        "where((a==223)|(a==221), 220, b)",
-        {"a": xx.classes_l3_l4.data, "b": res},
-        name="mark_urban",
-        dtype="uint8",
-    )
-
     # Mark nodata to 255 in case any nan
     res = expr_eval(
         "where(a==a, a, nodata)",
@@ -47,5 +43,12 @@ def lc_level3(xx: xr.Dataset):
         dtype="uint8",
         **{"nodata": NODATA},
     )
+    # Add intertidal as water
+    res = expr_eval(
+        "where((a==223)|(a==221), 220, b)",
+        {"a": xx.classes_l3_l4.data, "b": res},
+        name="mark_urban",
+        dtype="uint8",
+    )
 
     return res

odc/stats/plugins/lc_level34.py

@@ -60,9 +60,6 @@ def measurements(self) -> Tuple[str, ...]:
         _measurements = ["level3", "level4"]
         return _measurements
 
-    def native_transform(self, xx):
-        return xx
-
     def fuser(self, xx):
         return xx
 

tests/test_lc_l4_ctv.py

@@ -41,6 +41,7 @@ def image_groups(l34, urban, cultivated, woody, pv_pc_50):
         "cultivated_class": xr.DataArray(
             da.from_array(cultivated, chunks=(1, -1, -1)),
             dims=("spec", "y", "x"),
+            attrs={"nodata": 255},
         ),
         "woody_cover": xr.DataArray(
             da.from_array(woody, chunks=(1, -1, -1)),
@@ -129,12 +130,14 @@ def test_ctv_classes_woody():
 
     stats_l4 = StatsLccsLevel4()
     level3 = lc_level3.lc_level3(xx)
+
     lifeform = lc_lifeform.lifeform(xx)
     veg_cover = l4_veg_cover.canopyco_veg_con(xx, stats_l4.veg_threshold)
 
     l4_ctv = l4_cultivated.lc_l4_cultivated(
         xx.classes_l3_l4, level3, lifeform, veg_cover
     )
+
     assert (l4_ctv.compute() == expected_cultivated_classes).all()
 
 

tests/test_lc_level3.py

@@ -24,10 +24,10 @@ def image_groups():
                 [110, 110, 210],
                 [124, 110, 210],
                 [221, 210, 210],
-                [223, 255, 223],
+                [223, np.nan, 223],
             ]
         ],
-        dtype="uint8",
+        dtype="float32",
     )
 
     urban = np.array(
@@ -48,10 +48,10 @@ def image_groups():
                 [111, 112, 255],
                 [255, 112, 255],
                 [255, 255, 255],
-                [255, 255, 255],
+                [255, np.nan, np.nan],
             ]
         ],
-        dtype="uint8",
+        dtype="float32",
     )
 
     tuples = [