scrapers/azconni/scrape.rb

92 lines
2.2 KiB
Ruby

#!/usr/bin/env ruby
# coding: utf-8
require 'open-uri'
require 'nokogiri'
require 'erb'
MONTHS = [
"Januar", "Februar", "März", "April",
"Mai", "Juni", "Juli", "August",
"September", "Oktober", "November", "Dezember",
]
def fmt_time t
t.strftime "%Y%m%dT%H%M%S"
end
class Event
attr_accessor :title, :url
attr_reader :date, :location, :dtstart, :dtend
def initialize
@location = "AZ Conni, Rudolf-Leonhard-Straße 39"
end
def date=(s)
if s =~ /(\d{1,2})\.\s*(.+?)\s*(\d{4})[^\d]*(\d{1,2}):(\d{2})[^\d]*(\d{1,2}):(\d{2})/
year = $3.to_i
month = MONTHS.index($2) + 1
day = $1.to_i
@dtstart = Time::local year, month, day, $4.to_i, $5.to_i, 0
@dtend = Time::local year, month, day, $6.to_i, $7.to_i, 0
elsif s =~ /(\d{1,2})\.\s*(.+?)\s*(\d{4})[^\d]*(\d{1,2}):(\d{2})/
year = $3.to_i
month = MONTHS.index($2) + 1
day = $1.to_i
@dtstart = Time::local year, month, day, $4.to_i, $5.to_i, 0
@dtend = Time::local year, month, day, $4.to_i + 1, $5.to_i, 0
else
raise "Invalid date: #{s.inspect}"
end
while @dtend < @dtstart
# @dtend is in new day
@dtend += 86400
end
end
end
events = []
url = "https://www.azconni.de/termine/"
doc = Nokogiri::HTML URI.open(url)
doc.css(".termin").each do |termin|
event_url = URI.join url, termin.css("header a").attr("href")
event_doc = Nokogiri::HTML URI.open(event_url)
termin = event_doc.css(".termin.full")[0]
title = termin.css("header").text
category = termin.css(".categories").text
begin
ev = Event::new
ev.title = (category.empty? ? "" : "#{category}: ") + title
ev.url = event_url
ev.date = termin.css(".time").text
events << ev
rescue
STDERR.puts "Omitting: #{$!}"
# STDERR.puts $!.backtrace
end
end
ical = ERB::new <<~EOF
BEGIN:VCALENDAR
VERSION:2.0
METHOD:PUBLISH
X-WR-TIMEZONE;VALUE=TEXT:Europe/Berlin
<% events.each do |ev| %>
BEGIN:VEVENT
SUMMARY:<%= ev.title %>
DTSTART:<%= fmt_time ev.dtstart %>
DTEND:<%= fmt_time ev.dtend %>
UID:<%= ev.url %>
URL:<%= ev.url %>
LOCATION:<%= ev.location %>
END:VEVENT
<% end %>
END:VCALENDAR
EOF
puts ical.result