change x to x_osgb, similar for y

nowcasting_dataset/data_sources/data_source.py

@@ -469,14 +469,14 @@ def get_example(
             x_meters_center=x_meters_center, y_meters_center=y_meters_center
         )
         selected_data = selected_data.sel(
-            x=slice(bounding_box.left, bounding_box.right),
-            y=slice(bounding_box.top, bounding_box.bottom),
+            x_osgb=slice(bounding_box.left, bounding_box.right),
+            y_osgb=slice(bounding_box.top, bounding_box.bottom),
         )
 
         # selected_sat_data is likely to have 1 too many pixels in x and y
         # because sel(x=slice(a, b)) is [a, b], not [a, b).  So trim:
         selected_data = selected_data.isel(
-            x=slice(0, self._square.size_pixels), y=slice(0, self._square.size_pixels)
+            x_osgb=slice(0, self._square.size_pixels), y_osgb=slice(0, self._square.size_pixels)
         )
 
         selected_data = self._post_process_example(selected_data, t0_dt)

nowcasting_dataset/data_sources/fake/batch.py

@@ -420,10 +420,10 @@ def topographic_fake(batch_size, image_size_pixels, metadata: Optional[Metadata]
                 image_size_pixels,
                 image_size_pixels,
             ),
-            dims=["x", "y"],
+            dims=["x_osgb", "y_osgb"],
             coords=dict(
-                x=x,
-                y=y,
+                x_osgb=x,
+                y_osgb=y,
             ),
             name="data",
         )
@@ -436,7 +436,7 @@ def topographic_fake(batch_size, image_size_pixels, metadata: Optional[Metadata]
 
 
 def create_image_array(
-    dims=("time", "x", "y", "channels"),
+    dims=("time", "x_osgb", "y_osgb", "channels"),
     seq_length=19,
     history_seq_length=5,
     image_size_pixels=64,
@@ -461,8 +461,8 @@ def create_image_array(
 
     ALL_COORDS = {
         "time": time,
-        "x": x,
-        "y": y,
+        "x_osgb": x,
+        "y_osgb": y,
         "channels": np.array(channels),
     }
     coords = [(dim, ALL_COORDS[dim]) for dim in dims]
@@ -571,8 +571,8 @@ def create_gsp_pv_dataset(
     )
 
     data["capacity_mwp"] = capacity
-    data["x_coords"] = x_coords
-    data["y_coords"] = y_coords
+    data["x_osgb"] = x_coords
+    data["y_osgb"] = y_coords
 
     # Add 1000 to the id numbers for the row numbers.
     # This is a quick way to make sure row number is different from id,

nowcasting_dataset/data_sources/gsp/gsp_data_source.py

@@ -317,8 +317,8 @@ def get_example(
             data=gsp_y_coords.values,
             dims=["id"],
         )
-        gsp["x_coords"] = gsp_x_coords
-        gsp["y_coords"] = gsp_y_coords
+        gsp["x_osgb"] = gsp_x_coords
+        gsp["y_osgb"] = gsp_y_coords
 
         # pad out so that there are always 32 gsp, fill with 0
         pad_n = self.n_gsp_per_example - len(gsp.id)

nowcasting_dataset/data_sources/gsp/gsp_model.py

@@ -14,8 +14,8 @@ class GSP(DataSourceOutput):
     _expected_data_vars = (
         "power_mw",
         "capacity_mwp",
-        "x_coords",
-        "y_coords",
+        "x_osgb",
+        "y_osgb",
     )
 
     @classmethod
@@ -28,8 +28,8 @@ def model_validation(cls, v):
         v.check_data_var_dim(v.power_mw, ("example", "time_index", "id_index"))
         v.check_data_var_dim(v.capacity_mwp, ("example", "time_index", "id_index"))
         v.check_data_var_dim(v.time, ("example", "time_index"))
-        v.check_data_var_dim(v.x_coords, ("example", "id_index"))
-        v.check_data_var_dim(v.y_coords, ("example", "id_index"))
+        v.check_data_var_dim(v.x_osgb, ("example", "id_index"))
+        v.check_data_var_dim(v.y_osgb, ("example", "id_index"))
 
         return v
 

nowcasting_dataset/data_sources/nwp/nwp_data_source.py

@@ -87,6 +87,7 @@ def open(self) -> None:
         """
         self._data = self._open_data()
         np.testing.assert_array_equal(self._data["variable"].values, self.channels)
+        self._data = self._data.rename({"x": "x_osgb", "y": "y_osgb"})
 
     def _open_data(self) -> xr.DataArray:
         data = open_nwp(self.zarr_path, consolidated=self.consolidated)

nowcasting_dataset/data_sources/nwp/nwp_model.py

@@ -14,7 +14,7 @@ class NWP(DataSourceOutput):
     # Use to store xr.Dataset data
 
     __slots__ = ()
-    _expected_dimensions = ("time", "x", "y", "channels")
+    _expected_dimensions = ("time", "x_osgb", "y_osgb", "channels")
     _expected_data_vars = ("data",)
 
     @classmethod
@@ -24,7 +24,7 @@ def model_validation(cls, v):
         v.check_nan_and_inf(data=v.data)
 
         v.check_data_var_dim(
-            v.data, ("example", "time_index", "x_index", "y_index", "channels_index")
+            v.data, ("example", "time_index", "x_osgb_index", "y_osgb_index", "channels_index")
         )
 
         return v

nowcasting_dataset/data_sources/optical_flow/optical_flow_data_source.py

@@ -154,8 +154,8 @@ def _put_predictions_into_data_array(
             data=predictions,
             coords=(
                 ("time", datetime_index_of_predictions),
-                ("x", satellite_data_cropped.coords["x"].values),
-                ("y", satellite_data_cropped.coords["y"].values),
+                ("x_osgb", satellite_data_cropped.coords["x_osgb"].values),
+                ("y_osgb", satellite_data_cropped.coords["y_osgb"].values),
                 ("channels", satellite_data.coords["channels"].values),
             ),
             name="data",

nowcasting_dataset/data_sources/optical_flow/optical_flow_model.py

@@ -10,7 +10,7 @@ class OpticalFlow(DataSourceOutput):
     """Class to store optical flow data as a xr.Dataset with some validation"""
 
     __slots__ = ()
-    _expected_dimensions = ("time", "x", "y", "channels")
+    _expected_dimensions = ("time", "x_osgb", "y_osgb", "channels")
     _expected_data_vars = ("data",)
 
     @classmethod

nowcasting_dataset/data_sources/pv/pv_data_source.py

@@ -293,8 +293,8 @@ def get_example(
             data=pv_system_row_number,
             dims=["id"],
         )
-        pv["x_coords"] = x_coords
-        pv["y_coords"] = y_coords
+        pv["x_osgb"] = x_coords
+        pv["y_osgb"] = y_coords
         pv["pv_system_row_number"] = pv_system_row_number
 
         # pad out so that there are always n_pv_systems_per_example, pad with zeros

nowcasting_dataset/data_sources/pv/pv_model.py

@@ -16,8 +16,8 @@ class PV(DataSourceOutput):
         "power_mw",
         "capacity_mwp",
         "pv_system_row_number",
-        "x_coords",
-        "y_coords",
+        "x_osgb",
+        "y_osgb",
     )
 
     @classmethod
@@ -29,8 +29,8 @@ def model_validation(cls, v):
         v.check_data_var_dim(v.power_mw, ("example", "time_index", "id_index"))
         v.check_data_var_dim(v.capacity_mwp, ("example", "id_index"))
         v.check_data_var_dim(v.time, ("example", "time_index"))
-        v.check_data_var_dim(v.x_coords, ("example", "id_index"))
-        v.check_data_var_dim(v.y_coords, ("example", "id_index"))
+        v.check_data_var_dim(v.x_osgb, ("example", "id_index"))
+        v.check_data_var_dim(v.y_osgb, ("example", "id_index"))
         v.check_data_var_dim(v.pv_system_row_number, ("example", "id_index"))
 
         return v

nowcasting_dataset/data_sources/satellite/satellite_data_source.py

@@ -54,6 +54,7 @@ def open(self) -> None:
         self._data = self._open_data()
         if "variable" in self._data.dims:
             self._data = self._data.rename({"variable": "channels"})
+        self._data = self._data.rename({"x": "x_osgb", "y": "y_osgb"})
         if not set(self.channels).issubset(self._data.channels.values):
             raise RuntimeError(
                 f"One or more requested channels are not available in {self.zarr_path}!"
@@ -101,11 +102,13 @@ def get_spatial_region_of_interest(
             The selected data around the center
         """
         # Get the index into x and y nearest to x_center_osgb and y_center_osgb:
-        x_index_at_center = np.searchsorted(data_array.x.values, x_center_osgb) - 1
-        y_index_at_center = np.searchsorted(data_array.y.values, y_center_osgb) - 1
+        x_index_at_center = np.searchsorted(data_array.x_osgb.values, x_center_osgb) - 1
+        y_index_at_center = np.searchsorted(data_array.y_osgb.values, y_center_osgb) - 1
         # Put x_index_at_center and y_index_at_center into a pd.Series so we can operate
         # on them both in a single line of code.
-        x_and_y_index_at_center = pd.Series({"x": x_index_at_center, "y": y_index_at_center})
+        x_and_y_index_at_center = pd.Series(
+            {"x_osgb": x_index_at_center, "y_osgb": y_index_at_center}
+        )
         half_image_size_pixels = self._square.size_pixels // 2
         min_x_and_y_index = x_and_y_index_at_center - half_image_size_pixels
         max_x_and_y_index = x_and_y_index_at_center + half_image_size_pixels
@@ -114,8 +117,8 @@ def get_spatial_region_of_interest(
         suggested_reduction_of_image_size_pixels = (
             max(
                 (-min_x_and_y_index.min() if (min_x_and_y_index < 0).any() else 0),
-                (max_x_and_y_index.x - len(data_array.x)),
-                (max_x_and_y_index.y - len(data_array.y)),
+                (max_x_and_y_index.x_osgb - len(data_array.x_osgb)),
+                (max_x_and_y_index.y_osgb - len(data_array.y_osgb)),
             )
             * 2
         )
@@ -129,9 +132,9 @@ def get_spatial_region_of_interest(
                 "Requested region of interest of satellite data steps outside of the available"
                 " geographical extent of the Zarr data.  The requested region of interest extends"
                 f" from pixel indicies"
-                f" x={min_x_and_y_index.x} to x={max_x_and_y_index.x},"
-                f" y={min_x_and_y_index.y} to y={max_x_and_y_index.y}.  In the Zarr data,"
-                f" len(x)={len(data_array.x)}, len(y)={len(data_array.y)}. Try reducing"
+                f" x={min_x_and_y_index.x_osgb} to x={max_x_and_y_index.x_osgb},"
+                f" y={min_x_and_y_index.y_osgb} to y={max_x_and_y_index.y_osgb}.  In the Zarr data,"
+                f" len(x)={len(data_array.x_osgb)}, len(y)={len(data_array.y_osgb)}. Try reducing"
                 f" image_size_pixels from {self._square.size_pixels} to"
                 f" {new_suggested_image_size_pixels} pixels."
             )
@@ -140,8 +143,8 @@ def get_spatial_region_of_interest(
         # Note that isel is *exclusive* of the end of the slice.
         # e.g. isel(x=slice(0, 3)) will return the first, second, and third values.
         data_array = data_array.isel(
-            x=slice(min_x_and_y_index.x, max_x_and_y_index.x),
-            y=slice(min_x_and_y_index.y, max_x_and_y_index.y),
+            x_osgb=slice(min_x_and_y_index.x_osgb, max_x_and_y_index.x_osgb),
+            y_osgb=slice(min_x_and_y_index.y_osgb, max_x_and_y_index.y_osgb),
         )
         return data_array
 

nowcasting_dataset/data_sources/satellite/satellite_model.py

@@ -12,7 +12,7 @@ class Satellite(DataSourceOutput):
     """Class to store satellite data as a xr.Dataset with some validation"""
 
     __slots__ = ()
-    _expected_dimensions = ("time", "x", "y", "channels")
+    _expected_dimensions = ("time", "x_osgb", "y_osgb", "channels")
     _expected_data_vars = ("data",)
 
     @classmethod
@@ -22,7 +22,7 @@ def model_validation(cls, v):
         # previously nans were filled with -1s, so lets make sure there are none
         v.check_dataset_not_equal(data=v.data, value=-1)
         v.check_data_var_dim(
-            v.data, ("example", "time_index", "x_index", "y_index", "channels_index")
+            v.data, ("example", "time_index", "x_osgb_index", "y_osgb_index", "channels_index")
         )
 
         return v
@@ -32,5 +32,5 @@ class HRVSatellite(Satellite):
     """Class to store HRV satellite data as a xr.Dataset with some validation"""
 
     __slots__ = ()
-    _expected_dimensions = ("time", "x", "y", "channels")
+    _expected_dimensions = ("time", "x_osgb", "y_osgb", "channels")
     _expected_data_vars = ("data",)

nowcasting_dataset/data_sources/topographic/topographic_data_source.py

@@ -39,10 +39,13 @@ def __post_init__(self, image_size_pixels: int, meters_per_pixel: int):
             )
 
         self._data = self._data.fillna(0)  # Set nodata values to 0 (mostly should be ocean)
+        self._data = self._data.rename({"x": "x_osgb", "y": "y_osgb"})
         # Add CRS for later, topo maps are assumed to be in OSGB
         self._data.attrs["crs"] = OSGB
         # Distance between pixels, giving their spatial extant, in meters
-        self._stored_pixel_size_meters = abs(self._data.coords["x"][1] - self._data.coords["x"][0])
+        self._stored_pixel_size_meters = abs(
+            self._data.coords["x_osgb"][1] - self._data.coords["x_osgb"][0]
+        )
         self._meters_per_pixel = meters_per_pixel
 
     @staticmethod
@@ -72,8 +75,8 @@ def get_example(
             x_meters_center=x_meters_center, y_meters_center=y_meters_center
         )
         selected_data = self._data.sel(
-            x=slice(bounding_box.left, bounding_box.right),
-            y=slice(bounding_box.top, bounding_box.bottom),
+            x_osgb=slice(bounding_box.left, bounding_box.right),
+            y_osgb=slice(bounding_box.top, bounding_box.bottom),
         )
         if self._stored_pixel_size_meters != self._meters_per_pixel:
             # Rescale here to the exact size, assumes that the above is good slice
@@ -87,7 +90,7 @@ def get_example(
         # selected_data is likely to have 1 too many pixels in x and y
         # because sel(x=slice(a, b)) is [a, b], not [a, b).  So trim:
         selected_data = selected_data.isel(
-            x=slice(0, self._square.size_pixels), y=slice(0, self._square.size_pixels)
+            x_osgb=slice(0, self._square.size_pixels), y_osgb=slice(0, self._square.size_pixels)
         )
 
         selected_data = self._post_process_example(selected_data, t0_dt)

nowcasting_dataset/data_sources/topographic/topographic_model.py

@@ -7,7 +7,7 @@ class Topographic(DataSourceOutput):
     """Class to store topographic data as a xr.Dataset with some validation"""
 
     __slots__ = ()
-    _expected_dimensions = ("x", "y")
+    _expected_dimensions = ("x_osgb", "y_osgb")
     _expected_data_vars = ("data",)
 
     @classmethod
@@ -16,6 +16,6 @@ def model_validation(cls, v):
 
         v.check_nan_and_inf(data=v.data)
 
-        v.check_data_var_dim(v.data, ("example", "x_index", "y_index"))
+        v.check_data_var_dim(v.data, ("example", "x_osgb_index", "y_osgb_index"))
 
         return v

nowcasting_dataset/filesystem/utils.py

@@ -154,9 +154,11 @@ def download_to_local(remote_filename: str, local_filename: str):
     filesystem = get_filesystem(remote_filename)
     try:
         filesystem.get(remote_filename, local_filename)
-    except FileNotFoundError:
-        _LOG.error(f"Could not find {remote_filename}")
-        raise FileNotFoundError(f"Could not find {remote_filename}")
+    except FileNotFoundError as e:
+        _LOG.error(e)
+        message = f"Could not copy {remote_filename} to {local_filename}"
+        _LOG.error(message)
+        raise FileNotFoundError(message)
 
 
 def upload_one_file(remote_filename: str, local_filename: str, overwrite: bool = True):

tests/data_sources/fake/test_fake.py

@@ -32,11 +32,11 @@ def test_model(configuration):  # noqa: D103
 
     for i in range(configuration.process.batch_size):
         for data_source_name in ["satellite", "hrvsatellite", "opticalflow", "topographic", "nwp"]:
-            assert x_center_osgb[i] <= batch.__getattribute__(data_source_name).x.max()
-            assert x_center_osgb[i] >= batch.__getattribute__(data_source_name).x.min()
-            assert y_center_osgb[i] <= batch.__getattribute__(data_source_name).y.max()
-            assert y_center_osgb[i] >= batch.__getattribute__(data_source_name).y.min()
+            assert x_center_osgb[i] <= batch.__getattribute__(data_source_name).x_osgb.max()
+            assert x_center_osgb[i] >= batch.__getattribute__(data_source_name).x_osgb.min()
+            assert y_center_osgb[i] <= batch.__getattribute__(data_source_name).y_osgb.max()
+            assert y_center_osgb[i] >= batch.__getattribute__(data_source_name).y_osgb.min()
         # check first system is the center coordinates
         for data_source_name in ["gsp", "pv"]:
-            assert x_center_osgb[i] == batch.__getattribute__(data_source_name).x_coords[i, 0]
-            assert y_center_osgb[i] == batch.__getattribute__(data_source_name).y_coords[i, 0]
+            assert x_center_osgb[i] == batch.__getattribute__(data_source_name).x_osgb[i, 0]
+            assert y_center_osgb[i] == batch.__getattribute__(data_source_name).y_osgb[i, 0]

tests/data_sources/fake/test_fake_datasource_output.py

@@ -56,7 +56,7 @@ def test_satellite(configuration):
 
     s = satellite_fake(configuration=configuration)
 
-    assert s.x is not None
+    assert s.x_osgb is not None
 
 
 def test_sun():

tests/data_sources/gsp/test_gsp_data_source.py

@@ -121,8 +121,8 @@ def test_gsp_pv_data_source_get_example():
     )
 
     assert len(example.id) == len(example.power_mw[0])
-    assert len(example.x_coords) == len(example.y_coords)
-    assert len(example.x_coords) > 0
+    assert len(example.x_osgb) == len(example.y_osgb)
+    assert len(example.x_osgb) > 0
     assert pd.Timestamp(example.time[0].values) <= end_dt
     assert pd.Timestamp(example.time[0].values) >= start_dt
 
@@ -152,9 +152,9 @@ def test_gsp_pv_data_source_get_batch():
     )
 
     assert len(batch.power_mw[0]) == 4
-    assert len(batch.id[0]) == len(batch.x_coords[0])
-    assert len(batch.x_coords[1]) == len(batch.y_coords[1])
-    assert len(batch.x_coords[2]) > 0
+    assert len(batch.id[0]) == len(batch.x_osgb[0])
+    assert len(batch.x_osgb[1]) == len(batch.y_osgb[1])
+    assert len(batch.x_osgb[2]) > 0
     # assert T0_DT in batch[3].keys()