mkz-programm: init

mkz-programm/scrape.rb

@@ -0,0 +1,87 @@
+#!/usr/bin/env ruby
+# coding: utf-8
+
+require 'uri'
+require 'open-uri'
+require 'nokogiri'
+require 'erb'
+
+def fetch_doc url
+  STDERR.puts "GET #{url}"
+  Nokogiri::HTML URI.open(url)
+end
+
+def fmt_time t
+  t.strftime "%Y%m%dT%H%M%S"
+end
+
+class Event
+  attr_accessor :title, :start, :end, :location, :url, :image
+end
+
+events = []
+
+url = "https://www.medienkulturzentrum.de/angebote/"
+while url
+  list_doc = fetch_doc url
+  list_doc.css("a[text()='weiterlesen']").each do |a|
+    ev = Event::new
+    ev.url = URI.join url, a.attr('href')
+    doc = fetch_doc ev.url
+
+    date = doc.css('section.page-document header strong')[0]
+    date = date ? date.parent.text : next
+    if date =~ /(\d+)\.(\d+)\.(\d+),\s*(\d+):(\d+)\s*-\s*(\d+)\.(\d+)\.(\d+),\s*(\d+):(\d+)/
+      ev.start = Time::local $3.to_i, $2.to_i, $1.to_i, $4.to_i, $5.to_i, 0
+      ev.end = Time::local $8.to_i, $7.to_i, $6.to_i, $9.to_i, $10.to_i, 0
+    elsif date =~ /(\d+)\.(\d+)\.(\d+),\s*(\d+):(\d+)\s*-\s*(\d+):(\d+)/
+      ev.start = Time::local $3.to_i, $2.to_i, $1.to_i, $4.to_i, $5.to_i, 0
+      ev.end = Time::local $3.to_i, $2.to_i, $1.to_i, $6.to_i, $7.to_i, 0
+    elsif date =~ /(\d+)\.(\d+)\.(\d+)\s*-\s*(\d+)\.(\d+)\.(\d+)/
+      ev.start = Time::local $3.to_i, $2.to_i, $1.to_i, 0, 0, 0
+      ev.end = Time::local $6.to_i, $5.to_i, $4.to_i, 23, 59, 59
+    else
+      puts "Unrecognized date: #{date}"
+      next
+    end
+    ev.title = doc.css('section.page-document header h1').text
+    ev.location = doc.css('section.page-document address').text
+                    .lines
+                    .collect { |s| s.strip }
+                    .filter { |s| not s.empty? }
+                    .join(", ")
+    ev.image = doc.css('section.page-document article img[1]').attr('src')
+
+    events << ev
+  end
+
+  next_page = list_doc.css(".wp-pagenavi .current + .page")[0]
+  url = if next_page
+          URI.join url, next_page.attr('href')
+        else
+          nil
+        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
+  METHOD:PUBLISH
+  CLASS:PUBLIC
+  UID:<%= ev.url %>
+  DTSTART:<%= fmt_time(ev.start) %>
+  DTEND:<%= fmt_time(ev.end) %>
+  SUMMARY:<%= ev.title %>
+  LOCATION:<%= ev.location %>
+  URL:<%= ev.url %>
+  ATTACH;FMTTYPE=image/jpeg:<%= ev.image %>
+  END:VEVENT
+  <% end  %>
+  END:VCALENDAR
+EOF
+
+puts ical.result