Merge branch 'testing'

Author: jmp75

src/efts_io/_ncdf_stf2.py

@@ -93,6 +93,7 @@ def _create_cf_time_axis(data: xr.DataArray, timestep_str: str) -> tuple[np.ndar
     units = f"{timestep_str} since {formatted_string_with_tz}"
     return axis, units, calendar
 
+
 def _validate_station_id_for_int32(station_id: np.ndarray, intdata_type: str) -> None:
     """Validate that station_id values can be safely stored as int32.
 
@@ -114,6 +115,7 @@ def _validate_station_id_for_int32(station_id: np.ndarray, intdata_type: str) ->
                 f"station_id values must be in the int32 range [{np.iinfo(np.int32).min}, {np.iinfo(np.int32).max}] to be stored in STF2.0 format.",
             )
 
+
 def write_nc_stf2(
     out_nc_file: str,
     dataset: xr.Dataset,
@@ -481,7 +483,8 @@ def add_optional_variables(data: xr.DataArray, ncfile: Dataset, var_id: str) ->
         # This prevents double-close in the exception handler
         ncfile.close()
 
-def _get_stationid_data_types(stf_nc_vers: Any, d_type:np.ndarray, d_type_long:np.ndarray) -> None:
+
+def _get_stationid_data_types(stf_nc_vers: Any, d_type: np.ndarray, d_type_long: np.ndarray) -> None:
     if int(stf_nc_vers) == 1:
         d_type[1] = "fcast"
         d_type_long[1] = "forecast"

src/efts_io/wrapper.py

@@ -87,8 +87,8 @@ def load_from_stf2_file(file_path: str, time_zone_timestamps: bool) -> xr.Datase
 
     # work around https://jira.csiro.au/browse/WIRADA-635
     # lead_time can be a problem with xarray, so do not decode "times"
-    x = xr.open_dataset(file_path, decode_times=False)
-
+    # we also use mask_and_scale=False because of https://github.com/csiro-hydroinformatics/efts-io/issues/24
+    x = xr.open_dataset(file_path, decode_times=False, mask_and_scale=False)
     # replace the time and station names coordinates values
     # TODO This is probably not a long term solution for round-tripping a read/write or vice and versa
     decod = times.CFDatetimeCoder(use_cftime=True)
@@ -181,11 +181,19 @@ def __init__(self, data: Union[str, xr.Dataset]) -> None:
         self.ENS_MEMBER_DIMNAME = ENS_MEMBER_DIMNAME
         # self.identifiers_dimensions: list = []
         self.data: xr.Dataset
+        from pathlib import Path
+
+        if data is None:
+            raise ValueError("input cannot be None")
+        if isinstance(data, Path):
+            data = str(data)
         if isinstance(data, str):
             new_dataset = load_from_stf2_file(data, self.time_zone_timestamps)
             self.data = new_dataset
-        else:
+        elif isinstance(data, xr.Dataset):
             self.data = data
+        else:
+            raise TypeError(f"Unsupported type {type(data)}")
 
         self.stf2_int_datatype = "i4"  # default integer type for STF2 saving
 
@@ -275,8 +283,10 @@ def append_history(self, message: str, timestamp: Optional[datetime] = None) ->
         message: The message to append.
         timestamp: If not provided, the current UTC time is used.
         """
+        from datetime import UTC
+
         if timestamp is None:
-            timestamp = datetime.now(datetime.timezone.utc).isoformat()
+            timestamp = datetime.now(UTC).isoformat()
 
         current_history = self.data.attrs.get(HISTORY_ATTR_KEY, "")
         if current_history:

tests/test_create.py

@@ -126,8 +126,8 @@ def _create_test_ds():
             },
         }
     )
-    # 2025-11 It has been decided to support transparent conversion of string station 
-    # IDs to integers on save for STF2.0. 
+    # 2025-11 It has been decided to support transparent conversion of string station
+    # IDs to integers on save for STF2.0.
     # It was otherwise confusing for users.
     # besides it helps to promote the use of string station IDs in memory datasets.
     assert w.writeable_to_stf2()
@@ -192,7 +192,7 @@ def _saving_to_stf2(station_ids, intdata_type="i4", delete=True):
     # create a temporary file
     import tempfile
 
-    # save to STF2.0 will clean up the file if write fails, 
+    # save to STF2.0 will clean up the file if write fails,
     # so in that case we should allow for deletion to be True or False
     with tempfile.NamedTemporaryFile(suffix=".nc", delete=delete) as tmp:
         filename = tmp.name