From 5876793d01ab2e392e93eec9323dd6a7edb53170 Mon Sep 17 00:00:00 2001 From: Kévin Le Gouguec Date: Mon, 21 Feb 2022 19:57:49 +0100 Subject: Extract more stuff out of build-concerts.py --- helpers.py | 14 ++++++++++++++ 1 file changed, 14 insertions(+) (limited to 'helpers.py') diff --git a/helpers.py b/helpers.py index 87ed9c1..a746ea2 100644 --- a/helpers.py +++ b/helpers.py @@ -131,3 +131,17 @@ def read_concerts(filename): for match in re.finditer(_CONCERT_RE, f.read()) ) return tuple(sorted(concerts, key=attrgetter('time'))) + + +_TOUCHUPS = ( + (re.compile('([0-9])(st|nd|rd|th|er|ère|nde|ème)'), r'\1\2'), + (re.compile('(https://[^ ]+)'), r'\1'), + (re.compile('([^ ]+@[^ ]+)'), r'\1'), +) + + +def touchup_plaintext(plaintext): + text = plaintext + for regexp, repl in _TOUCHUPS: + text = regexp.sub(repl, text) + return text -- cgit v1.2.3