Rough fix for .museum e-mail address detection

master
Sven Slootweg 11 years ago
parent 2911e7e949
commit 4b19e9025e

@ -75,8 +75,8 @@ grammar = {
'(?P<val>[a-z0-9-]+\.d?ns[0-9]*\.([a-z0-9-]+\.)+[a-z0-9]+)',
'(?P<val>([a-z0-9-]+\.)+[a-z0-9]+)(\s+([0-9]{1,3}\.){3}[0-9]{1,3})',
'[^a-z0-9.-](?P<val>d?ns\.([a-z0-9-]+\.)+[a-z0-9]+)'],
'emails': ['(?P<val>[\w.-]+@[\w.-]+\.[\w]{2,4})',
'(?P<val>[\w.-]+\sAT\s[\w.-]+\sDOT\s[\w]{2,4})']
'emails': ['(?P<val>[\w.-]+@[\w.-]+\.[\w]{2,6})', # Really need to fix this, much longer TLDs now exist...
'(?P<val>[\w.-]+\sAT\s[\w.-]+\sDOT\s[\w]{2,6})']
},
"_dateformats": (
'(?P<day>[0-9]{1,2})[./ -](?P<month>Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)[./ -](?P<year>[0-9]{4}|[0-9]{2})'

Loading…
Cancel
Save