netflow, make sure get_timeserie_data() returns string type objects in stead of bytes, to prevent 17e4e9c0fc (commitcomment-33587137)

we might find some other small type interpretation differences, but the bumpiest seems to be gone now.

src/opnsense/scripts/netflow/get_timeseries.py

@@ -61,7 +61,7 @@ if __name__ == '__main__':
                     record_key = []
                     for key_field in cmd_args.key_fields.split(','):
                         if key_field in record and record[key_field] is not None:
-                            record_key.append(str(record[key_field]))
+                            record_key.append(record[key_field])
                         else:
                             record_key.append('')
                     record_key = ','.join(record_key)

src/opnsense/scripts/netflow/lib/aggregates/__init__.py

@@ -267,7 +267,10 @@ class BaseFlowAggregator(object):
                 result_record = dict()
                 for field_indx in range(len(field_names)):
                     if len(record) > field_indx:
-                        result_record[field_names[field_indx]] = record[field_indx]
+                        if type(record[field_indx]) == bytes:
+                            result_record[field_names[field_indx]] = record[field_indx].decode()
+                        else:
+                            result_record[field_names[field_indx]] = record[field_indx]
                 if 'start_time' in result_record:
                     result_record['end_time'] = result_record['start_time'] \
                                                 + datetime.timedelta(seconds=self.resolution)