feat(nwp-consumer): Add west-europe option for IFS (#225)

Author: devsjc

src/nwp_consumer/internal/entities/coordinates.py

@@ -510,6 +510,14 @@ def crop(
                 " and both must sit between 180 and -180 degrees.",
             ))
 
+        if (north > self.latitude[0] or south < self.latitude[-1]
+            or west < self.longitude[0] or east > self.longitude[-1]):
+            return Failure(ValueError(
+                "Cannot crop coordinates to a region outside the bounds of the map. "
+                f"Crop region '{north, west, south, east}' not in "
+                f"map bounds '{self.nwse()}'.",
+            ))
+
         # Determine the indices of the region in the latitude and longitude lists
         lat_indices = [
             i for i, lat in enumerate(self.latitude)

src/nwp_consumer/internal/entities/modelmetadata.py

@@ -98,6 +98,12 @@ def with_region(self, region: str) -> "ModelMetadata":
                 ).map(lambda coords: dataclasses.replace(
                     self, name=f"{self.name}_india", expected_coordinates=coords,
                 )).unwrap()
+            case "west-europe":
+                return self.expected_coordinates.crop(
+                    north=63, west=-12, south=35, east=26,
+                ).map(lambda coords: dataclasses.replace(
+                    self, name=f"{self.name}_west-europe", expected_coordinates=coords,
+                )).unwrap()
             case _:
                 log.warning(f"Unknown region '{region}', not cropping expected coordinates.")
                 return self
@@ -115,7 +121,7 @@ class Models:
     """Namespace containing known models."""
 
     ECMWF_HRES_IFS_0P1DEGREE: ModelMetadata = ModelMetadata(
-        name="HRES-IFS",
+        name="hres-ifs",
         resolution="0.1 degrees",
         expected_coordinates=NWPDimensionCoordinateMap(
             init_time=[],
@@ -147,7 +153,7 @@ class Models:
     """ECMWF's High Resolution Integrated Forecast System."""
 
     ECMWF_ENS_STAT_0P1DEGREE: ModelMetadata = ModelMetadata(
-        name="ENS-Stat",
+        name="ens-stat",
         resolution="0.1 degrees",
         expected_coordinates=NWPDimensionCoordinateMap(
             init_time=[],
@@ -166,7 +172,7 @@ class Models:
     """Summary statistics from ECMWF's Ensemble Forecast System."""
 
     ECMWF_ENS_0P1DEGREE: ModelMetadata = ModelMetadata(
-        name="ENS",
+        name="ens",
         resolution="0.1 degrees",
         expected_coordinates=NWPDimensionCoordinateMap(
             init_time=[],
@@ -193,7 +199,7 @@ class Models:
     """Full ensemble data from ECMWF's Ensemble Forecast System."""
 
     NCEP_GFS_1DEGREE: ModelMetadata = ModelMetadata(
-        name="NCEP-GFS",
+        name="ncep-gfs",
         resolution="1 degree",
         expected_coordinates=NWPDimensionCoordinateMap(
             init_time=[],
@@ -224,7 +230,7 @@ class Models:
     """NCEP's Global Forecast System."""
 
     MO_UM_GLOBAL_17KM: ModelMetadata = ModelMetadata(
-        name="UM-Global",
+        name="um-global",
         resolution="17km",
         expected_coordinates = NWPDimensionCoordinateMap(
             init_time=[],
@@ -259,7 +265,7 @@ class Models:
     """MetOffice's Unified Model, in the Global configuration, at a resolution of 17km."""
 
     MO_UM_GLOBAL_10KM: ModelMetadata = ModelMetadata(
-        name="UM-Global",
+        name="um-global",
         resolution="10km",
         expected_coordinates=NWPDimensionCoordinateMap(
             init_time=[],

src/nwp_consumer/internal/repositories/raw_repositories/ecmwf_mars.py

@@ -222,12 +222,14 @@ def repository() -> entities.RawRepositoryMetadata:
                 "default": entities.Models.ECMWF_HRES_IFS_0P1DEGREE.with_region("uk"),
                 "hres-ifs-uk": entities.Models.ECMWF_HRES_IFS_0P1DEGREE.with_region("uk"),
                 "hres-ifs-india": entities.Models.ECMWF_HRES_IFS_0P1DEGREE.with_region("india"),
+                "hres-ifs-west-europe": entities.Models.ECMWF_HRES_IFS_0P1DEGREE\
+                    .with_region("west-europe"),
                 "ens-stat-india": entities.Models.ECMWF_ENS_STAT_0P1DEGREE.with_region("india"),
                 "ens-stat-uk": entities.Models.ECMWF_ENS_STAT_0P1DEGREE.with_region("uk"),
                 "ens-uk": entities.Models.ECMWF_ENS_0P1DEGREE.with_region("uk")\
-                        .with_chunk_count_overrides(
-                            {"latitude": 2, "longitude": 2, "variable": 1, "ensemble_member": 5},
-                        ),
+                    .with_chunk_count_overrides(
+                        {"latitude": 2, "longitude": 2, "variable": 1, "ensemble_member": 5},
+                    ),
             },
         )
 
@@ -331,6 +333,7 @@ def _convert(path: pathlib.Path) -> ResultE[list[xr.DataArray]]:
                 f"Error context: {e}",
             ))
 
+        processed_das: list[xr.DataArray] = []
         try:
             # Merge the datasets back into one
             ds: xr.Dataset = xr.merge(
@@ -340,12 +343,11 @@ def _convert(path: pathlib.Path) -> ResultE[list[xr.DataArray]]:
             )
             del dss
 
-            if "ens" in ECMWFMARSRawRepository.model().name.lower():
-                # Add in missing coordinates for mean/std data
-                if "enfo-es" in path.name:
-                    ds = ds.expand_dims(dim={"ensemble_stat": ["std"]})
-                elif "enfo-em" in path.name:
-                    ds = ds.expand_dims(dim={"ensemble_stat": ["mean"]})
+            # Add in missing coordinates for mean/std data
+            if "enfo-es" in path.name:
+                ds = ds.expand_dims(dim={"ensemble_stat": ["std"]})
+            elif "enfo-em" in path.name:
+                ds = ds.expand_dims(dim={"ensemble_stat": ["mean"]})
             da: xr.DataArray = (
                 ds
                 .pipe(
@@ -356,7 +358,7 @@ def _convert(path: pathlib.Path) -> ResultE[list[xr.DataArray]]:
                 .expand_dims("init_time")
                 .to_dataarray(name=ECMWFMARSRawRepository.model().name)
             )
-            if "number" in da.coords:
+            if "ens" in path.as_posix():
                 da = da.rename({"number": "ensemble_member"})
             da = (
                 da.drop_vars(
@@ -370,11 +372,21 @@ def _convert(path: pathlib.Path) -> ResultE[list[xr.DataArray]]:
                 .sortby(variables=["step", "variable", "longitude"])
                 .sortby(variables="latitude", ascending=False)
             )
+            # Put each variable into its own DataArray:
+            # * Each raw file does not contain a full set of parameters
+            # * and so may not produce a contiguous subset of the expected coordinates.
+            processed_das.extend(
+                [
+                    da.where(cond=da.coords["variable"] == v, drop=True)
+                    for v in da.coords["variable"].values
+                ],
+            )
+
 
         except Exception as e:
             return Failure(ValueError(
                 f"Error processing DataArray for path '{path!s}'. Error context: {e}",
             ))
 
-        return Success([da])
+        return Success(processed_das)
 

src/nwp_consumer/internal/repositories/raw_repositories/noaa_s3.py

@@ -267,8 +267,8 @@ def _convert(path: pathlib.Path) -> ResultE[list[xr.DataArray]]:
             # * and so may not produce a contiguous subset of the expected coordinates.
             processed_das.extend(
                 [
-                    da.where(cond=da["variable"] == v, drop=True)
-                    for v in da["variable"].values
+                    da.where(cond=da.coords["variable"] == v, drop=True)
+                    for v in da.coords["variable"].values
                 ],
             )