read xarray in cache dir for caravan

hydrodataset/caravan.py

@@ -1,4 +1,5 @@
 import collections
+import glob
 import re
 import warnings
 from tqdm import tqdm
@@ -774,7 +775,7 @@ def _check_data(self, regions):
     def read_attr_xrdataset(self, gage_id_lst=None, var_lst=None, **kwargs):
         # Define the path to the attributes file
         file_path = os.path.join(
-            self.data_source_description["ATTR_DIR"],
+            CACHE_DIR,
             "caravan_attributes.nc",
         )
 
@@ -791,22 +792,17 @@ def read_attr_xrdataset(self, gage_id_lst=None, var_lst=None, **kwargs):
         return ds
 
     def read_ts_xrdataset(self, gage_id_lst, t_range, var_lst, **kwargs):
-        # TS_dir is same as flow_dir/forcing_dir
-        base_path = self.data_source_description["FLOW_DIR"]
-        file_paths = []
-
-        # Generate list of file paths based on sites_id
-        for site in gage_id_lst:
-            # Split the site string to get the region and site_code
-            region, site_code = site.split("_")
-
-            file_path = os.path.join(base_path, region, f"{region}_{site_code}.nc")
-            file_paths.append(file_path)
+        file_paths = sorted(
+            glob.glob(os.path.join(CACHE_DIR, "*caravan*timeseries*.nc"))
+        )
 
         # Open the dataset in a lazy manner using dask
         parallel = kwargs.get("parallel", False)
         combined_ds = xr.open_mfdataset(
-            file_paths, combine="nested", concat_dim="basin", parallel=parallel
+            file_paths,
+            combine="nested",
+            concat_dim="gauge_id",
+            parallel=parallel,
         )
 
         def extract_unit(variable_name, units_string):
@@ -829,6 +825,8 @@ def extract_unit(variable_name, units_string):
             combined_ds = combined_ds[var_lst]
         if t_range:
             combined_ds = combined_ds.sel(date=slice(*t_range))
+        if gage_id_lst:
+            combined_ds = combined_ds.sel(gauge_id=gage_id_lst)
 
         # some units are not recognized by pint_xarray, hence we manually set them
         unit_mapping = {"W/m2": "watt / meter ** 2", "m3/m3": "meter^3/meter^3"}
@@ -840,9 +838,8 @@ def extract_unit(variable_name, units_string):
                 unit = unit_mapping.get(unit, unit)
                 combined_ds[var].attrs["units"] = unit
 
-        # Assign basin names as coordinate
-        combined_ds = combined_ds.assign_coords(basin=gage_id_lst)
         combined_ds = combined_ds.rename({"date": "time"})
+        combined_ds = combined_ds.rename({"gauge_id": "basin"})
         return combined_ds
 
     @property