Update for xarray backend

This commit adds more strict checks to coord/dim attrs to make sure
the Dataset and/or DataArray dimension sizes matches the what is
in the input files.

src/grib2io/xarray_backend.py

@@ -424,12 +424,15 @@ def parse_grib_index(index, filters):
     # productDefinitionTemplateNumber
     non_geo_dims = list()
 
+    total_dim_count = 1
+
     #TODO Eventually re-work this section making 'non_geo_dims'
 
     # refDate always added for now (could add only based on
     # typOfGeneratingProcess)
     if 'refDate' not in index.columns:
         index = index.assign(refDate=index.msg.apply(lambda msg: msg.refDate))
+        total_dim_count *= index['refDate'].nunique()
     @dataclass(init=False)
     class RefDateDim:
         refDate: pd.Index = PdIndex()
@@ -439,65 +442,91 @@ class RefDateDim:
     # typOfGeneratingProcess)
     if 'leadTime' not in index.columns:
         index = index.assign(leadTime=index.msg.apply(lambda msg: msg.leadTime))
+        total_dim_count *= index['leadTime'].nunique()
     @dataclass(init=False)
     class LeadTimeDim:
         leadTime: pd.Index = PdIndex()
     non_geo_dims.append(LeadTimeDim)
 
     if 'valueOfFirstFixedSurface' not in index.columns:
         index = index.assign(valueOfFirstFixedSurface=index.msg.apply(lambda msg: msg.valueOfFirstFixedSurface))
+        total_dim_count *= index['valueOfFirstFixedSurface'].nunique()
     @dataclass(init=False)
     class ValueOfFirstFixedSurfaceDim:
         valueOfFirstFixedSurface: pd.Index = PdIndex()
     non_geo_dims.append(ValueOfFirstFixedSurfaceDim)
 
-    # logic for parsing possible dims from specific product definition section
+    # Logic for parsing possible dims from specific product definition section
 
     if pdtn in {5,9}:
         # Probability forecasts at a horizontal level or in a horizontal layer
-        # in a continuous or non-continuous time interval.  (see Template
-        # 4.9)
+        # in a continuous or non-continuous time interval.  (see Template 4.9)
         index = index.assign(thresholdLowerLimit = index.msg.apply(lambda msg: msg.thresholdLowerLimit))
         index = index.assign(thresholdUpperLimit = index.msg.apply(lambda msg: msg.thresholdUpperLimit))
+        total_dim_count *= index['thresholdLowerLimit'].nunique() # Only need to do this with one attr
 
-        if index['thresholdLowerLimit'].nunique() > 1:
+        if index['thresholdLowerLimit'].nunique() > 1 and \
+           index['thresholdLowerLimit'].nunique() == total_dim_count:
             @dataclass(init=False)
             class ThresholdLowerLimitDim:
                 thresholdLowerLimit: pd.Index = PdIndex()
             non_geo_dims.append(ThresholdLowerLimitDim)
-        if index['thresholdUpperLimit'].nunique() > 1:
+
+        index = index.assign(thresholdUpperLimit = index.msg.apply(lambda msg: msg.thresholdUpperLimit))
+        total_dim_count *= index['thresholdUpperLimit'].nunique()
+        if index['thresholdUpperLimit'].nunique() > 1 and \
+           index['thresholdUpperLimit'].nunique() == total_dim_count:
             @dataclass(init=False)
             class ThresholdUpperLimitDim:
                 thresholdUpperLimit: pd.Index = PdIndex()
             non_geo_dims.append(ThresholdUpperLimitDim)
 
     if pdtn in {6,10}:
         # Percentile forecasts at a horizontal level or in a horizontal layer
-        # in a continuous or non-continuous time interval.  (see Template
-        # 4.10)
-        index = index.assign(percentileValue = index.msg.apply(lambda msg: msg.percentileValue))
-
-        @dataclass(init=False)
-        class PercentileValueDim:
-            percentileValue: pd.Index = PdIndex()
-        non_geo_dims.append(PercentileValueDim)
+        # in a continuous or non-continuous time interval.  (see Template 4.10)
+        if 'percentileValue' not in index.columns:
+            index = index.assign(percentileValue = index.msg.apply(lambda msg: msg.percentileValue))
+            total_dim_count *= index['percentileValue'].nunique()
+
+        if index['percentileValue'].nunique() >= 1:
+            if index['percentileValue'].nunique() in {1,total_dim_count}:
+                @dataclass(init=False)
+                class PercentileValueDim:
+                    percentileValue: pd.Index = PdIndex()
+                non_geo_dims.append(PercentileValueDim)
+            else:
+                unique_percentile = index.percentileValue.unique()
+                raise ValueError(f'filter to a single percentileValue; found: {[str(i) for i in unique_percentile]}')
 
     if pdtn in {8,9,10,11,12,13,14,42,43,45,46,47,61,62,63,67,68,72,73,78,79,82,83,84,85,87,91}:
         if 'duration' not in index.columns:
             index = index.assign(duration=index.msg.apply(lambda msg: msg.duration))
-
-        @dataclass(init=False)
-        class Duration:
-            duration: pd.Index = PdIndex()
-        non_geo_dims.append(Duration)
+            total_dim_count *= index['duration'].nunique()
+
+        if index['duration'].nunique() >= 1:
+            if index['duration'].nunique() in {1,total_dim_count}:
+                @dataclass(init=False)
+                class Duration:
+                    duration: pd.Index = PdIndex()
+                non_geo_dims.append(Duration)
+            else:
+                unique_duration = index.duration.unique()
+                raise ValueError(f'filter to a single duration; found: {[str(i) for i in unique_duration]}')
 
     if pdtn in {1,11,33,34,41,43,45,47,49,54,56,58,59,63,68,77,79,81,83,84,85,92}:
         if 'perturbationNumber' not in index.columns:
             index = index.assign(perturbationNumber = index.msg.apply(lambda msg: msg.perturbationNumber))
-        @dataclass(init=False)
-        class perturbationNumber:
-            perturbationNumber: pd.Index = PdIndex()
-        non_geo_dims.append(perturbationNumber)
+            total_dim_count *= index['perturbationNumber'].nunique()
+
+        if index['perturbationNumber'].nunique() >= 1:
+            if index['perturbationNumber'].nunique() in {1,total_dim_count}:
+                @dataclass(init=False)
+                class perturbationNumber:
+                    perturbationNumber: pd.Index = PdIndex()
+                non_geo_dims.append(perturbationNumber)
+            else:
+                unique_pertnum = index.duration.unique()
+                raise ValueError(f'filter to a single duration; found: {[str(i) for i in unique_pertnum]}')
 
     return index, non_geo_dims
 
@@ -520,7 +549,6 @@ def build_da_without_coords(index, cube, filename) -> xr.DataArray:
     DataArray
         DataArray without coordinates
     """
-
     dim_names = [k for k in cube.__dataclass_fields__.keys() if cube[k] is not None and len(cube[k]) > 1]
     constant_meta_names = [k for k in cube.__dataclass_fields__.keys() if cube[k] is None]
     dims = {k: len(cube[k]) for k in dim_names}