Improve error message when loading data fails

Author: staadecker

switch_model/utilities/load_data.py

@@ -99,16 +99,16 @@ def __str__(self):
         return repr(self.value)
 
 
-def load_data(switch_data, optional, auto_select, optional_params, **kwds):
-    path = kwds["filename"]
+def load_data(switch_data, optional, auto_select, optional_params, **kwargs):
+    path = kwargs["filename"]
     # Skip if the file is missing
     if optional and not os.path.isfile(path):
         return
     # If this is a .dat file, then skip the rest of this fancy business; we'll
     # only check if the file is missing and optional for .csv files.
     filename, extension = os.path.splitext(path)
     if extension == ".dat":
-        switch_data.load(**kwds)
+        switch_data.load(**kwargs)
         return
 
     # copy the optional_params to avoid side-effects when the list is altered below
@@ -134,14 +134,14 @@ def load_data(switch_data, optional, auto_select, optional_params, **kwds):
     # Try to get a list of parameters. If param was given as a
     # singleton or a tuple, make it into a list that can be edited.
     params = []
-    if "param" in kwds:
+    if "param" in kwargs:
         # Tuple -> list
-        if isinstance(kwds["param"], tuple):
-            kwds["param"] = list(kwds["param"])
+        if isinstance(kwargs["param"], tuple):
+            kwargs["param"] = list(kwargs["param"])
         # Singleton -> list
-        elif not isinstance(kwds["param"], list):
-            kwds["param"] = [kwds["param"]]
-        params = kwds["param"]
+        elif not isinstance(kwargs["param"], list):
+            kwargs["param"] = [kwargs["param"]]
+        params = kwargs["param"]
     # optional_params may include Param objects instead of names. In
     # those cases, convert objects to names.
     for (i, p) in enumerate(optional_params):
@@ -157,11 +157,11 @@ def load_data(switch_data, optional, auto_select, optional_params, **kwds):
             optional_params.append(p.name)
     # How many index columns do we expect?
     # Grab the dimensionality of the index param if it was provided.
-    if "index" in kwds:
-        num_indexes = kwds["index"].dimen
+    if "index" in kwargs:
+        num_indexes = kwargs["index"].dimen
         if num_indexes == UnknownSetDimen:
             raise Exception(
-                f"Index {kwds['index'].name} has unknown dimension. Specify dimen= during its creation."
+                f"Index {kwargs['index'].name} has unknown dimension. Specify dimen= during its creation."
             )
     # Next try the first parameter's index.
     elif len(params) > 0:
@@ -185,7 +185,7 @@ def load_data(switch_data, optional, auto_select, optional_params, **kwds):
     # within the file (e.g., "cost" and "limit"). We could also require the data file
     # to be called "rfm_supply_tier.csv" for greater consistency/predictability.
     if auto_select:
-        if "select" in kwds:
+        if "select" in kwargs:
             raise InputError(
                 "You may not specify a select parameter if "
                 + "auto_select is set to True."
@@ -197,15 +197,15 @@ def get_column_name(p):
             else:
                 return p.name
 
-        kwds["select"] = headers[0:num_indexes] + [get_column_name(p) for p in params]
+        kwargs["select"] = headers[0:num_indexes] + [get_column_name(p) for p in params]
     # Check to see if expected column names are in the file. If a column
     # name is missing and its parameter is optional, then drop it from
     # the select & param lists.
-    if "select" in kwds:
-        if isinstance(kwds["select"], tuple):
-            kwds["select"] = list(kwds["select"])
+    if "select" in kwargs:
+        if isinstance(kwargs["select"], tuple):
+            kwargs["select"] = list(kwargs["select"])
         del_items = []
-        for (i, col) in enumerate(kwds["select"]):
+        for (i, col) in enumerate(kwargs["select"]):
             p_i = i - num_indexes
             if col not in headers:
                 if len(params) > p_i >= 0 and params[p_i].name in optional_params:
@@ -218,17 +218,20 @@ def get_column_name(p):
         # to first so that the indexes won't get messed up as we go.
         del_items.sort(reverse=True)
         for (i, p_i) in del_items:
-            del kwds["select"][i]
-            del kwds["param"][p_i]
+            del kwargs["select"][i]
+            del kwargs["param"][p_i]
 
     if optional and file_has_no_data_rows:
         # Skip the file.  Note that we are only doing this after having
         # validated the file's column headings.
         return
 
     # Use our custom DataManager to allow 'inf' in csvs.
-    if kwds["filename"][-4:] == ".csv":
-        kwds["using"] = "switch_csv"
+    if kwargs["filename"][-4:] == ".csv":
+        kwargs["using"] = "switch_csv"
     # All done with cleaning optional bits. Pass the updated arguments
     # into the DataPortal.load() function.
-    switch_data.load(**kwds)
+    try:
+        switch_data.load(**kwargs)
+    except:
+        raise Exception(f"Failed to load data from file {path}.")