From 8e4133d45e6ab4c539ffb5e8def34a2193f40ee6 Mon Sep 17 00:00:00 2001
From: Cyprian Laskowski <cyp@cjvt.si>
Date: Thu, 19 Nov 2020 17:00:55 +0100
Subject: [PATCH] IssueID #1487: added helper script for fixing xml ids

---
 scripts/fix_xml_ids.py | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)
 create mode 100644 scripts/fix_xml_ids.py

diff --git a/scripts/fix_xml_ids.py b/scripts/fix_xml_ids.py
new file mode 100644
index 0000000..7b0ceea
--- /dev/null
+++ b/scripts/fix_xml_ids.py
@@ -0,0 +1,21 @@
+import argparse
+import codecs
+import re
+
+arg_parser = argparse.ArgumentParser(description='Fix invalid XML ids.')
+arg_parser.add_argument('-infile', type=str, help='Input file')
+arg_parser.add_argument('-outfile', type=str, help='Output file')
+arguments = arg_parser.parse_args()
+input_file_name = arguments.infile
+output_file_name = arguments.outfile
+
+output_file = codecs.open(output_file_name, 'w')
+input_file = codecs.open(input_file_name, 'r')
+
+for line in input_file:
+    line = re.sub('xml:id="(?=\d)','xml:id="s', line)
+    line = line.replace('#', '#s')
+    output_file.write(line)
+    
+input_file.close()
+output_file.close()