#!/usr/bin/env python
import cgi, urllib, os, sys, BeautifulSoup
from xml.sax.saxutils import escape
form = cgi.FieldStorage()
if os.environ.get('REQUEST_METHOD','').lower() == 'get':
if form.has_key('url'):
print "Content-type: application/xml; charset=utf-8\r\n\r\n",
sys.argv = ['', form.getvalue('url')]
else:
print 'Content-type: text/html\r\n\r\n',
print 'Scrape SYO'
print ''
sys.exit(0)
print '''
'''
outline = ' '
for url in sys.argv[1:]:
soup = BeautifulSoup.BeautifulStoneSoup(urllib.urlopen(url).read())
for row in soup.findAll('tr'):
cols=row.findAll('td')
xmlUrl=cols[2].find('a')['href']
url = cols[0].find('a')
if url:
text=url.string
url='url="' + url['href'] + '" '
else:
text=cols[0].string
url=''
print outline % (escape(url), escape(text), escape(xmlUrl))
print '''
'''