From b931680c737d166f466165683b6e71161942d617 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andreas=20Sch=C3=A4rtl?= <andreas.schaertl@fau.de>
Date: Tue, 28 Apr 2020 20:06:21 +0200
Subject: [PATCH] add fix-rdf-file.py
passed an rdf file w/ bad iris on stdin, it returns
a fixed version
---
ulo/fix-rdf-file.py | 36 ++++++++++++++++++++++++++++++++++++
1 file changed, 36 insertions(+)
create mode 100755 ulo/fix-rdf-file.py
diff --git a/ulo/fix-rdf-file.py b/ulo/fix-rdf-file.py
new file mode 100755
index 0000000..a671246
--- /dev/null
+++ b/ulo/fix-rdf-file.py
@@ -0,0 +1,36 @@
+#! /usr/bin/env python3
+
+
+from urllib import parse
+import re
+import sys
+
+
+def fix_quoted(s: str) -> str:
+ payload = s.strip('"')
+ fixed = parse.quote(payload)
+ return '"%s"' % fixed
+
+
+def print_fixed_line(line: str):
+ # https://stackoverflow.com/questions/249791
+ regex = r"(\"(?:[^\"]|\\.)*\")"
+ matches = re.findall(regex, line)
+
+ for match in matches:
+ fixed = fix_quoted(match)
+ line = line.replace(match, fixed)
+
+ print(line, end='')
+
+
+def main():
+ for line in sys.stdin:
+ print_fixed_line(line)
+
+
+if __name__ == '__main__':
+ try:
+ main()
+ except (KeyboardInterrupt, SystemExit, BrokenPipeError):
+ pass
--
GitLab