omb: accept dataframe

OpenDrift · Dec 4, 2024 · cd69e6d · cd69e6d
1 parent ed5e14a
commit cd69e6d
Show file tree

Hide file tree

Showing 2 changed files with 26 additions and 23 deletions.
diff --git a/trajan/readers/omb.py b/trajan/readers/omb.py
@@ -60,11 +60,15 @@ def append_dict_with_entry(dict_in: dict, parsed_entry: ParsedIridiumMessage):
         dict_in[parsed_entry.device_from]["T"] = []
     dict_in[parsed_entry.device_from][parsed_entry.kind].append(parsed_entry)
 
+EXPECTED_COLUMNS = [
+    'Date Time (UTC)', 'Device', 'Direction', 'Payload', 'Approx Lat/Lng',
+    'Payload (Text)', 'Length (Bytes)', 'Credits'
+]
 
-def read_omb_csv(path_in: Path,
+def read_omb_csv(path_in: Path|pd.DataFrame,
                  dict_instruments_params: dict = None,
                  modified_wave_packet_properties: dict = None) -> xr.Dataset:
-    logger.debug("read path to pandas")
+    logger.debug(f"reading {path_in}..")
 
     if modified_wave_packet_properties is None:
         nbr_bins_waves = _BD_YWAVE_NBR_BINS
@@ -76,21 +80,19 @@ def read_omb_csv(path_in: Path,
     ########################################
     # generic pandas read to be able to open from a variety of files
 
-    omb_dataframe = pd.read_csv(path_in)
-
-    ########################################
-    # check this is actually a Rock7 data file
+    if isinstance(path_in, pd.DataFrame):
+        omb_dataframe = path_in
+    else:
+        omb_dataframe = pd.read_csv(path_in)
+        ########################################
+        # check this is actually a Rock7 data file
 
-    columns = omb_dataframe.columns.to_list()
-    expected_columns = [
-        'Date Time (UTC)', 'Device', 'Direction', 'Payload', 'Approx Lat/Lng',
-        'Payload (Text)', 'Length (Bytes)', 'Credits'
-    ]
+        columns = omb_dataframe.columns.to_list()
 
-    if not set(expected_columns).issubset(set(columns)):
-        raise RuntimeError(
-            f"does not look like a Rock7 file; got colmns {columns}, expected {expected_columns}, missing: {set(expected_columns) - set(columns)}"
-        )
+        if not set(EXPECTED_COLUMNS).issubset(set(columns)):
+            raise RuntimeError(
+                f"does not look like a Rock7 file; got colmns {columns}, expected {EXPECTED_COLUMNS}, missing: {set(EXPECTED_COLUMNS) - set(columns)}"
+            )
 
     ########################################
     # decode

diff --git a/trajan/traj.py b/trajan/traj.py
@@ -361,27 +361,28 @@ def assign_cf_attrs(self,
             'long_name': 'trajectory name'
         }
 
+
         ds = ds.assign_attrs({
             'Conventions':
             'CF-1.10',
             'featureType':
             'trajectory',
             'geospatial_lat_min':
-            np.nanmin(self.tlat),
+            np.nanmin(self.tlat) if self.ds.sizes[self.obs_dim] > 0 else np.nan,
             'geospatial_lat_max':
-            np.nanmax(self.tlat),
+            np.nanmax(self.tlat) if self.ds.sizes[self.obs_dim] > 0 else np.nan,
             'geospatial_lon_min':
-            np.nanmin(self.tlon),
+            np.nanmin(self.tlon) if self.ds.sizes[self.obs_dim] > 0 else np.nan,
             'geospatial_lon_max':
-            np.nanmax(self.tlon),
+            np.nanmax(self.tlon) if self.ds.sizes[self.obs_dim] > 0 else np.nan,
             'time_coverage_start':
             pd.to_datetime(
-                np.nanmin(ds['time'].values[ds['time'].values != np.datetime64(
-                    'NaT')])).isoformat(),
+                np.nanmin(ds[self.time_varname].values[ds[self.time_varname].values != np.datetime64(
+                    'NaT')])).isoformat() if self.ds.sizes[self.obs_dim] > 0 else np.nan,
             'time_coverage_end':
             pd.to_datetime(
-                np.nanmax(ds['time'].values[ds['time'].values != np.datetime64(
-                    'NaT')])).isoformat(),
+                np.nanmax(ds[self.time_varname].values[ds[self.time_varname].values != np.datetime64(
+                    'NaT')])).isoformat() if self.ds.sizes[self.obs_dim] > 0 else np.nan,
         })
 
         if creator_name: