scripts/opdsupdates.py
2014-09-13 21:23:13 +02:00

78 lines
2 KiB
Python

#!/usr/bin/env python
# -*- coding: utf-8 -*-
################################################################################
import time
import urllib2
import re
import smtplib
def findstring(url):
"""find if document's last update date is equal to current date"""
doc = urllib2.urlopen(url).read()
doctime = re.search(r'Catalogue g&eacute;n&eacute;r&eacute; le (.*)<',
doc)
if doctime != None:
print url.rstrip('\n') + " : "+doctime.group(1)
date = re.match(r'(.*) (.*) (.*)',doctime.group(1))
day = date.group(1)
month = date.group(2)
year = date.group(3)
#day = re.match(r'\d (.*)',doctime.group(1))
if day < 10:
day = '0' + day
months = { 'janv.' : '01',
'f&eacute;vr.' : '02',
'mars' : '03',
'avr.' : '04',
'mai' : '05',
'juin' : '06',
'juil.' : '07',
'aout' : '08',
'sept.' : '09',
'oct.' : '10',
'nov.' : '11',
'd&eacute;c' : '12'}
for i,j in months.iteritems():
month = month.replace(i,j)
doctime = time.strptime(day + ' ' + month + ' ' + year,
'%d %m %Y')[0:3]
curtime = time.localtime()[0:3]
if doctime == curtime:
rooturl = re.match('(.*)/_catalog/(.*)',url)
url = rooturl.group(1) + '/_catalog/index.html'
sendmail('OPDS Update','catalog ' + url + ' updated!')
return True
else:
return False
else:
sendmail('OPDS Error','string not found at catalog ' + url )
def sendmail(subject,corpus):
"""send an alert e-mail"""
fromaddr = "noreply@tourmentine.com"
mto = "n@tourmentine.com"
message = """From: No Reply <noreply@tourmentine.com>
To: nico <n@tourmentine.com>
Subject: """+subject+"""
"""+corpus+"""
"""
smtp = smtplib.SMTP()
smtp.connect()
smtp.sendmail(fromaddr,mto,message)
smtp.close()
print time.strftime('%c') + ' : Starting harvest...\n'
f = open('opdslist.txt')
for line in f:
if not re.match(r'^#',line):
findstring(line)
print '\n'