X-Git-Url: https://git.mdrn.pl/wolnelektury.git/blobdiff_plain/5105e426102784140df92c6a6ce1ffdf70866a4e..6c49f9cfb80c746034d7113883c7c355f547187e:/src/wolnelektury/utils.py diff --git a/src/wolnelektury/utils.py b/src/wolnelektury/utils.py index d20039cd2..2657a5542 100644 --- a/src/wolnelektury/utils.py +++ b/src/wolnelektury/utils.py @@ -155,3 +155,14 @@ class UnicodeCSVWriter(object): # the original re.escape messes with unicode def re_escape(s): return re.sub(r"[(){}\[\].*?|^$\\+-]", r"\\\g<0>", s) + + +BOT_BITS = ['bot', 'slurp', 'spider', 'facebook', 'crawler', 'parser', 'http'] + + +def is_crawler(request): + user_agent = request.META.get('HTTP_USER_AGENT') + if not user_agent: + return True + user_agent = user_agent.lower() + return any(bot_bit in user_agent for bot_bit in BOT_BITS)