Improve handling of the index columns (#296)

This PR removes generation of all know index columns; selective
generation is done instead. Columns without any values are dropped. This
allows cutting down on the processing time.
This PR also harmonises fi_t and uc_t processing transforms by
extracting common parts into separate functions.

Author: olejandro

xl2times/__main__.py

@@ -131,23 +131,23 @@ def convert_xl_to_times(
         transforms.process_tradelinks,
         transforms.include_tables_source,
         transforms.process_processes,
-        transforms.process_topology,
-        transforms.fill_in_missing_values,
+        transforms.create_model_topology,
+        transforms.fill_in_column_defaults,
         transforms.generate_uc_properties,
         transforms.expand_rows_parallel,  # slow
+        transforms.process_trade_links,
+        transforms.merge_tables,
         transforms.remove_invalid_values,
-        transforms.internalise_commodities,
-        transforms.generate_commodity_groups,
-        transforms.apply_fixups,
+        transforms.include_cgs_in_topology,
         transforms.fill_in_missing_pcgs,
-        transforms.generate_trade,
-        transforms.merge_tables,
         transforms.complete_processes,
-        transforms.process_units,
-        transforms.complete_commodity_groups,
+        transforms.create_model_units,
         transforms.process_wildcards,
         transforms.convert_aliases,
-        transforms.fix_topology,
+        transforms.enforce_availability,
+        transforms.complete_model_trade,
+        transforms.create_model_cgs,
+        transforms.apply_fixups,
         transforms.apply_transform_tables,
         transforms.generate_implied_topology,
         transforms.verify_uc_topology,