From a26109cb3006dd64e1bec38100397ad636813e98 Mon Sep 17 00:00:00 2001 From: Sam Wenham Date: Thu, 11 Jul 2019 12:29:38 +0100 Subject: Allow comments against names in logbooks in brackets Convert accent chars in names into simple chars as this is what people enter in the logbook --- parsers/logbooks.py | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'parsers/logbooks.py') diff --git a/parsers/logbooks.py b/parsers/logbooks.py index 12124ca..e5817a6 100644 --- a/parsers/logbooks.py +++ b/parsers/logbooks.py @@ -25,9 +25,10 @@ from utils import save_carefully # # the logbook loading section # -def GetTripPersons(trippeople, expedition, logtime_underground): +def GetTripPersons(trippeople, expedition, logtime_underground): res = [ ] author = None + round_bracket_regex = re.compile(r"[\(\[].*?[\)\]]") for tripperson in re.split(r",|\+|&|&(?!\w+;)| and ", trippeople): tripperson = tripperson.strip() mul = re.match(r"(.*?)$(?i)", tripperson) @@ -35,6 +36,7 @@ def GetTripPersons(trippeople, expedition, logtime_underground): tripperson = mul.group(1).strip() if tripperson and tripperson[0] != '*': #assert tripperson in personyearmap, "'%s' << %s\n\n %s" % (tripperson, trippeople, personyearmap) + tripperson = re.sub(round_bracket_regex, "", tripperson).strip() personyear = GetPersonExpeditionNameLookup(expedition).get(tripperson.lower()) if not personyear: print(" - No name match for: '%s'" % tripperson) @@ -172,8 +174,8 @@ def Parseloghtmltxt(year, expedition, txt): tripid, tripid1, tripdate, trippeople, triptitle, triptext, tu = s.groups() ldate = ParseDate(tripdate.strip(), year) #assert tripid[:-1] == "t" + tripdate, (tripid, tripdate) - #trippeople = re.sub(r"Ol(?!l)", "Olly", trippeople) - #trippeople = re.sub(r"Wook(?!e)", "Wookey", trippeople) + #trippeople = re.sub(r"Ol(?!l)", "Olly", trippeople) + #trippeople = re.sub(r"Wook(?!e)", "Wookey", trippeople) triptitles = triptitle.split(" - ") if len(triptitles) >= 2: tripcave = triptitles[0] -- cgit v1.2.3