import sys
import os
import urllib
-import urllib2
import re as regexp
+import subprocess
from optparse import OptionParser
from BeautifulSoup import BeautifulSoup
+
+# For fuck's sake!
+import codecs
+import locale
+def fix_stdout():
+ sys.stdout = codecs.EncodedFile(sys.stdout, locale.getpreferredencoding())
+ def null_decode(input, errors='strict'):
+ return input, len(input)
+ sys.stdout.decode = null_decode
+# From
+# http://ewx.livejournal.com/457086.html?thread=3016574
+# http://ewx.livejournal.com/457086.html?thread=3016574
+# lightly modified.
+# See also Debian #415968.
+fix_stdout()
+
+
+# User agent:
+class YarrgURLopener(urllib.FancyURLopener):
+ base_version= urllib.URLopener().version
+ proc= subprocess.Popen(
+ ["./database-info-fetch", "useragentstringmap",
+ base_version, "manual islands/topology fetch"],
+ shell=False,
+ stderr=None,
+ stdout=subprocess.PIPE,
+ )
+ version = proc.communicate()[0].rstrip('\n');
+ assert(proc.returncode is not None and proc.returncode == 0)
+urllib._urlopener = YarrgURLopener()
+
ocean = None
soup = None
opts = None
url = ('http://yppedia.puzzlepirates.com/' +
(url_base % urllib.quote(ocean,'')))
debug('fetching',url)
- dataf = urllib2.urlopen(url)
+ dataf = urllib.urlopen(url)
debug('fetched',dataf)
- soup = BeautifulSoup(dataf,
- convertEntities=BeautifulSoup.HTML_ENTITIES)
+ soup = BeautifulSoup(dataf)
title_arch_re = regexp.compile('(\\S.*\\S) Archipelago \\((\\S+)\\)$')
def parse_chart():
ta = soup.find('textarea')
debug('ta',ta)
- rc = ta.renderContents()
- debug('rc',rc)
- return rc
+ s = ta.string
+ debug('s',s)
+ s = regexp.sub(r'\<\;', '<', s)
+ s = regexp.sub(r'\>\;', '>', s)
+ s = regexp.sub(r'\"\;', '"', s)
+ s = regexp.sub(r'\&\;', '&', s)
+ debug('s',s)
+ return s
def parse_ocean():
- firstarch = soup.find('a', attrs = {'title': title_arch_ok})
- debug('fa',firstarch)
+ content = soup.find('div', attrs = {'id': 'content'})
def findall_title_arch_ok(t):
return t.findAll('a', attrs = {'title': title_arch_ok})
if u.name != 'table': return False
return len(findall_title_arch_ok(u)) > 1
- archestable = firstarch.findParent('table', attrs={'border':'1'})
+ archestable = content.findChild('table', attrs={'border':'1'})
debug('at',archestable)
archsoups = []