glamr/tables: fix the uniref100 column in exported readabundance data

Take into account that the FK field can be referred to by the field name
or the attname.

mibios/glamr/tables.py

@@ -371,7 +371,7 @@ def get_export_fields(self, columns):
         """
         field_names = []
         for i in super().get_export_fields():
-            if i == 'ref_id':
+            if i == 'ref' or i == 'ref_id':
                 # adds a join, but almost each row will have a distinct uniref,
                 # so probably better than prefetch/caching
                 field_names.append('ref__accession')