adding GlacioBasis folder to inputs and appending

src/pypromice/process/aws.py

@@ -55,7 +55,8 @@ def __init__(
         """
         assert os.path.isfile(config_file), "cannot find " + config_file
         assert os.path.isdir(inpath), "cannot find " + inpath
-        logger.info(
+
+        logger.debug(
             "AWS("
             f"config_file={config_file},"
             f" inpath={inpath},"

src/pypromice/process/join_l3.py

@@ -49,6 +49,13 @@ def parse_arguments_joinl3(debug_args=None):
         required=False,
         help="Path to GC-Net historical L1 folder",
     )
+    parser.add_argument(
+        "-gb",
+        "--folder_glaciobasis",
+        type=str,
+        required=False,
+        help="Path to GlacioBasis historical data folder",
+    )
 
     parser.add_argument(
         "-o",
@@ -147,7 +154,7 @@ def readNead(infile):
         ds = df.to_xarray()
         ds.attrs = meta
 
-        # renaming variables
+        # renaming variables if GC-Net
         file_path = pypromice.resources.DEFAULT_VARIABLES_ALIASES_GCNET_PATH
         var_name = pd.read_csv(file_path)
         var_name = var_name.set_index("old_name").GEUS_name
@@ -160,7 +167,7 @@ def readNead(infile):
 
         # renaming variables to the GEUS names
         ds = ds.rename(var_name)
-
+                
         # variables always dropped from the historical GC-Net files
         # could be move to the config files at some point
         standard_vars_to_drop = [
@@ -201,6 +208,28 @@ def loadArr(infile, isNead):
             df = pd.read_csv(infile)
             df["time"] = pd.to_datetime(df["time"]).dt.tz_localize(None)
             df = df.set_index("time")
+
+            # renaming variables if GlacioBasis
+            file_path = pypromice.resources.DEFAULT_VARIABLES_ALIASES_GLACIOBASIS_PATH
+            var_name = pd.read_csv(file_path)
+            var_name = var_name.set_index("old_name").GEUS_name
+            msk = [v for v in var_name.index if v in df.columns]
+            var_name = var_name.loc[msk].to_dict()
+
+            # postprocessing for glaciobasis
+            if 'ice_ablation' in df.columns:
+                df['ice_ablation'] = (-df['ice_ablation'].diff()).cumsum()
+                df['z_surf_combined'] = df['ice_ablation'].values
+
+            # renaming variables to the GEUS names
+            df = df.rename(columns=var_name)
+
+            # variables always dropped from the historical GC-Net files
+            # could be move to the config files at some point
+            standard_vars_to_drop = ["I"]
+
+            # Drop the variables if they are present in the dataset
+            df = df.drop(columns=[var for var in standard_vars_to_drop if var in df.columns])
             ds = xr.Dataset.from_dataframe(df)
 
     elif infile.split(".")[-1].lower() in "nc":
@@ -366,7 +395,7 @@ def build_station_list(config_folder: str, target_station_site: str) -> list:
     return station_info_list
 
 
-def join_l3(config_folder, site, folder_l3, folder_gcnet, outpath, variables, metadata):
+def join_l3(config_folder, site, folder_l3, folder_gcnet, folder_glaciobasis, outpath, variables, metadata):
     # Get the list of station information dictionaries associated with the given site
     list_station_info = build_station_list(config_folder, site)
 
@@ -377,9 +406,17 @@ def join_l3(config_folder, site, folder_l3, folder_gcnet, outpath, variables, me
 
         filepath = os.path.join(folder_l3, stid, stid + "_hour.nc")
         isNead = False
-        if station_info["project"].lower() in ["historical gc-net"]:
-            filepath = os.path.join(folder_gcnet, stid + ".csv")
-            isNead = True
+
+        if not os.path.isfile(filepath):
+            if station_info["project"].lower() in ["historical gc-net"]:
+                filepath = os.path.join(folder_gcnet, stid + ".csv")
+                isNead = True
+
+            if station_info["project"].lower() in ["glaciobasis"]:
+                filepath = os.path.join(folder_glaciobasis, stid + ".csv")
+                isNead = False
+
+        # import pdb; pdb.set_trace()
         if not os.path.isfile(filepath):
             logger.error(
                 "\n***\n"
@@ -551,6 +588,7 @@ def main():
         args.site,
         args.folder_l3,
         args.folder_gcnet,
+        args.folder_glaciobasis,
         args.outpath,
         args.variables,
         args.metadata,

src/pypromice/resources/__init__.py

@@ -7,6 +7,7 @@
 DEFAULT_METADATA_PATH = (Path(__file__).parent / "file_attributes.csv").absolute()
 DEFAULT_VARIABLES_PATH = (Path(__file__).parent / "variables.csv").absolute()
 DEFAULT_VARIABLES_ALIASES_GCNET_PATH = (Path(__file__).parent / "variable_aliases_GC-Net.csv").absolute()
+DEFAULT_VARIABLES_ALIASES_GLACIOBASIS_PATH = (Path(__file__).parent / "variable_aliases_GlacioBasis.csv").absolute()
 
 def load_metadata(path: Union[None, str, Path] = None) -> Dict[str, str]:
     """

src/pypromice/resources/variable_aliases_GlacioBasis.csv

@@ -0,0 +1,8 @@
+GEUS_name,old_name
+wspd_u,wspd
+rh_u_cor,rh_u_corr
+dsr_cor,dsr_corr
+usr_cor,usr_corr
+cc,cloud_cover
+z_boom_u,z_boom
+z_ice_surf,ice_ablation