Fix corpus building process for vrt files without entities

app/converters/vrt.py

@@ -97,7 +97,7 @@ def normalize_vrt_file(input_file, output_file):
             multi_line_tag_definition = False
             continue
         pos_attrs = line.rstrip('\n').split('\t')
-        if not has_ent_as_s_attr:
+        if not has_ent_as_s_attr and len(pos_attrs) > 4:
             if pos_attrs[4].lower() in ['null', 'none']:
                 if current_ent:
                     output_vrt += '</ent>\n'