Merge branch 'test-install'

[ypp-sc-tools.db-live.git] / yarrg / yppedia-ocean-scraper

diff --git a/yarrg/yppedia-ocean-scraper b/yarrg/yppedia-ocean-scraper
index ad35c4d708a432566941542156e10339d5ed421f..476c1cd9b381bb0d5d5ec1979cc4928bff9fa9f4 100755 (executable)
--- a/yarrg/yppedia-ocean-scraper
+++ b/yarrg/yppedia-ocean-scraper
@@ -42,6 +42,23 @@ import re as regexp
 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()
+
+
 ocean = None
 soup = None
 opts = None
@@ -95,8 +112,7 @@ def parse_chart():
        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})
@@ -105,7 +121,7 @@ def parse_ocean():
                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 = []