setop/agenda_culturel
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix problème import (tiret manquant parfois)

Browse Source
This commit is contained in:
Jean-Marie Favreau 2025-03-28 16:04:49 +01:00
parent 9f7204a35f
commit 83bcc43e35
1 changed files with 12 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
src/agenda_culturel/import_tasks/custom_extractors/lefotomat.py
View File

@ -1,4 +1,5 @@
from bs4 import BeautifulSoup
import re
from ..extractor import Extractor
from ..twosteps_extractor import TwoStepsExtractor
@ -51,10 +52,17 @@ class CExtractor(TwoStepsExtractor):
else:
image = None
desc = soup.select("head meta[name=description]")[0]["content"]
start_day = Extractor.parse_french_date(desc.split("-")[0])
start_time = Extractor.parse_french_time(desc.split("-")[1])
end_time = Extractor.parse_french_time(desc.split("-")[2])
end_day = Extractor.guess_end_day(start_day, start_time, end_time)
res = re.match("(.*202[5-9])[ -]*([0-9h:]+)[ -]*([0-9h:]*)", desc)
if res:
start_day = Extractor.parse_french_date(res[1])
start_time = Extractor.parse_french_time(res[2])
end_time = Extractor.parse_french_time(res[3])
end_day = Extractor.guess_end_day(start_day, start_time, end_time)
else:
start_day = None
start_time = None
end_time = None
end_day = None
location = self.nom_lieu
descriptions = soup.select("div.vce-col-content")