Add script to parse user agents with historical data set.

(imported from commit e529c0b914ed3d3d06e9581a6239676f68c97b3f)

tools/test_user_agent_parsing.py

@@ -0,0 +1,33 @@
+#!/usr/bin/python
+import re
+from collections import defaultdict
+import os
+import sys
+
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..'))
+
+from zerver.lib.user_agent import parse_user_agent
+
+user_agents_parsed = defaultdict(int)
+user_agents_path = os.path.join(os.path.dirname(__file__), "user_agents_unique")
+parse_errors = 0
+for line in file(user_agents_path).readlines():
+    line = line.strip()
+    match = re.match('^(?P<count>[0-9]+) "(?P<user_agent>.*)"$', line)
+    if match is None:
+        print line
+        continue
+    groupdict = match.groupdict()
+    count = groupdict["count"]
+    user_agent = groupdict["user_agent"]
+    ret = parse_user_agent(user_agent)
+    if ret is None:
+        print "parse error", line
+        parse_errors += 1
+        continue
+    user_agents_parsed[ret["name"]] += int(count)
+
+for key in user_agents_parsed:
+    print "    ", key, user_agents_parsed[key]
+
+print "%s parse errors!" % (parse_errors,)

zerver/lib/user_agent.py

@@ -0,0 +1,10 @@
+import re
+
+# Warning: If you change this parsing, please test using
+#   tools/test_user_agent_parsing.py
+# And extend tools/user_agents_unique with any new test cases
+def parse_user_agent(user_agent):
+    match = re.match("^(?P<name>[^/ ]*[^0-9/(]*)(/(?P<version>[^/ ]*))?([ /].*)?$", user_agent)
+    if match is None:
+        return None
+    return match.groupdict()