#!/usr/bin/python3

# Copyright: (2013-2015) Michael Till Beck <Debianguru@gmx.de>
# License: GPL-2.0+

import urllib.request, urllib.error, urllib.parse
import urllib.parse
from lxml import etree
from cssselect import GenericTranslator
import re
import io

import smtplib
from email.mime.text import MIMEText
from email.header import Header
from urllib.parse import urljoin

import os
import sys
import getopt
import traceback

import subprocess

import time
from time import strftime
import random

import importlib
config = None

defaultEncoding = 'utf-8'
maxTitleLength = 150

# this is how an empty RSS feed looks like
emptyfeed = """<?xml version="1.0"?>
<rss version="2.0">
 <channel>
  <title>MailWebsiteChanges Feed</title>
  <link>https://github.com/Debianguru/MailWebsiteChanges</link>
  <description>MailWebsiteChanges Feed</description>
 </channel>
</rss>"""

# Attributes in HTML files storing URI values. These values are automatically translated to absolute URIs.
uriAttributes = [['//img[@src]', 'src'], ['//a[@href]', 'href']]
cmdscheme = 'cmd://'

mailsession = None


# translates all relative URIs found in trees to absolute URIs
def toAbsoluteURIs(trees, baseuri):
        for tree in trees:
                if isinstance(tree, str):
                        continue
                for uriAttribute in uriAttributes:
                        tags = tree.xpath(uriAttribute[0])
                        for tag in tags:
                                if tag.attrib.get(uriAttribute[1]) != None:
                                        if urllib.parse.urlparse(tag.attrib[uriAttribute[1]]).scheme == '':
                                                tag.attrib[uriAttribute[1]] = urllib.parse.urljoin(baseuri, tag.attrib[uriAttribute[1]])


def parseSite(site):
        file, content, titles, warning = None, None, None, None

        uri = site['uri']
        contenttype = site.get('type', 'html')
        contentregex = site.get('contentregex', '')
        titleregex = site.get('titleregex', '')
        enc = site.get('encoding', defaultEncoding)

        contentxpath = site.get('contentxpath', '')
        if contentxpath == '' and site.get('contentcss', '') != '':
                # CSS
                contentxpath = GenericTranslator().css_to_xpath(site.get('contentcss'))
        titlexpath = site.get('titlexpath', '')
        if titlexpath == '' and site.get('titlecss', '') != '':
                titlexpath = GenericTranslator().css_to_xpath(site.get('titlecss'))

        try:

                if uri.startswith(cmdscheme):
                        # run command and retrieve output
                        process = subprocess.Popen(uri[len(cmdscheme):], stdout=subprocess.PIPE, shell=True, close_fds=True)
                        file = process.stdout
                else:
                        # open website
                        req = urllib.request.Request(uri)
                        if 'user-agent' in site:
                            req.add_header('User-Agent', site['user-agent'])
                        file = urllib.request.urlopen(req)


                if contenttype == 'text' or (contentxpath == '' and titlexpath == ''):
                        contents = [file.read().decode(enc)]
                        titles = []
                else:
                        baseuri = uri
                        if contenttype == 'html':
                                parser = etree.HTMLParser(encoding=enc)
                        else:
                                parser = etree.XMLParser(recover=True, encoding=enc)

                        tree = etree.parse(file, parser)

                        # xpath
                        contentresult = tree.xpath(contentxpath) if contentxpath else []
                        titleresult = tree.xpath(titlexpath) if titlexpath else []

                        # translate relative URIs to absolute URIs
                        if contenttype == 'html':
                                basetaglist = tree.xpath('/html/head/base')
                                if len(basetaglist) != 0:
                                        baseuri = basetaglist[0].attrib['href']
                                if len(contentresult) != 0:
                                        toAbsoluteURIs(contentresult, baseuri)
                                if len(titleresult) != 0:
                                        toAbsoluteURIs(titleresult, baseuri)

                        if contentxpath != '' and titlexpath != '' and len(contentresult) != len(titleresult):
                                warning = 'WARNING: number of title blocks (' + str(len(titleresult)) + ') does not match number of content blocks (' + str(len(contentresult)) + ')'
                        elif contentxpath and len(contentresult) == 0:
                                warning = 'WARNING: content selector became invalid!'
                        elif titlexpath and len(titleresult) == 0:
                                warning = 'WARNING: title selector became invalid!'
                        else:
                                if len(contentresult) == 0:
                                        contentresult = titleresult
                                if len(titleresult) == 0:
                                        titleresult = contentresult

                        if isinstance(contentresult, str):
                                contents = [contentresult]
                        else:
                                contents = [etree.tostring(s, encoding=defaultEncoding, pretty_print=True).decode(defaultEncoding) for s in contentresult]
                        if isinstance(titleresult, str):
                                titles = [getSubject(titleresult)]
                        else:
                                titles = [getSubject(' '.join(s.xpath('.//text()'))) for s in titleresult]

        except IOError as e:
                warning = 'WARNING: could not open URL; maybe content was moved?\n\n' + str(e)

        if file is not None:
                file.close()

        if uri.startswith(cmdscheme) and process.wait() != 0:
                warning = 'WARNING: process terminated with an error'

        if warning:
                return {'content': content, 'titles': titles, 'warning': warning}

        # parse regex
        if contentregex:
                contents = [x for y in [re.findall(r'' + contentregex, c, re.S) for c in contents] for x in y]
        if titleregex:
                titles = [x for y in [re.findall(r'' + titleregex, c, re.S) for c in titles] for x in y]

        if contentregex and titleregex and len(contents) != len(titles):
                warning = 'WARNING: number of title blocks (' + str(len(titles)) + ') does not match number of content blocks (' + str(len(contents)) + ') after regex'
        elif contentregex and len(contents) == 0:
                warning = 'WARNING: content regex became invalid!'
        elif titleregex and len(titles) == 0:
                warning = 'WARNING: title regex became invalid!'
        else:
                if len(contents) == 0:
                        contents = titles
                if len(titles) == 0:
                        titles = [getSubject(c) for c in contents]

        return {'contents': contents, 'titles': titles, 'warning': warning}


# returns a short subject line
def getSubject(textContent):
        if textContent == None or textContent == '':
                return config.subjectPostfix
        textContent = re.sub(' +', ' ', re.sub('\s', ' ', textContent)).strip()
        return (textContent[:maxTitleLength] + ' [..]') if len(textContent) > maxTitleLength else textContent


# generates a new RSS feed item
def genFeedItem(subject, content, link, change):
        feeditem = etree.Element('item')
        titleitem = etree.Element('title')
        titleitem.text = subject + ' #' + str(change)
        feeditem.append(titleitem)
        linkitem = etree.Element('link')
        linkitem.text = link
        feeditem.append(linkitem)
        descriptionitem = etree.Element('description')
        descriptionitem.text = content
        feeditem.append(descriptionitem)
        guiditem = etree.Element('guid')
        guiditem.text = str(random.getrandbits(32))
        feeditem.append(guiditem)
        dateitem = etree.Element('pubDate')
        dateitem.text = strftime("%a, %d %b %Y %H:%M:%S %Z", time.localtime())
        feeditem.append(dateitem)

        return feeditem


# sends mail notification
def sendmail(receiver, subject, content, sendAsHtml, link):
        global mailsession

        if sendAsHtml:
                baseurl = None
                if link != None:
                        content = '<p><a href="' + link + '">' + subject + '</a></p>\n' + content
                        baseurl = urljoin(link, '/')
                mail = MIMEText('<html><head><title>' + subject + '</title>' + ('<base href="' + baseurl + '">' if baseurl else '') + '</head><body>' + content + '</body></html>', 'html', defaultEncoding)
        else:
                if link != None:
                        content = link + '\n\n' + content
                mail = MIMEText(content, 'text', defaultEncoding)

        mail['From'] = config.sender
        mail['To'] = receiver
        mail['Subject'] = Header(subject, defaultEncoding)

        # initialize session once, not each time this method gets called
        if mailsession is None:
                mailsession = smtplib.SMTP(config.smtphost, config.smtpport)
                if config.useTLS:
                        mailsession.ehlo()
                        mailsession.starttls()
                mailsession.login(config.smtpusername, config.smtppwd)

        mailsession.sendmail(config.sender, receiver.split(','), mail.as_string())


# returns a list of all content that is stored locally for a specific site
def getFileContents(shortname):
        result = []
        for f in os.listdir('.'):
                if f.startswith(shortname + '.') and f.endswith('.txt'):
                        file = open(f, 'rb')
                        result.append(file.read().decode('utf-8'))
                        file.close()
        return result


# updates list of content that is stored locally for a specific site
def storeFileContents(shortname, parseResult):
        for f in os.listdir('.'):
                if f.startswith(shortname + '.') and f.endswith('.txt'):
                        os.remove(f)

        i = 0
        for c in parseResult['contents']:
                file = open(shortname + '.' + str(i) + '.txt', 'wb')
                file.write(c.encode('utf-8'))
                file.close()
                i += 1


def pollWebsites():

        # parse existing feed or create a new one
        if config.enableRSSFeed:
                if os.path.isfile(config.rssfile):
                        feedXML = etree.parse(config.rssfile)
                else:
                        feedXML = etree.parse(io.StringIO(emptyfeed))

        # start polling sites
        for site in config.sites:

                print('polling site [' + site['shortname'] + '] ...')
                parseResult = parseSite(site)
                receiver = site.get('receiver', config.receiver)

                # if something went wrong, notify the user
                if parseResult['warning']:
                        subject = '[' + site['shortname'] + '] WARNING'
                        print('WARNING: ' + parseResult['warning'])
                        if config.enableMailNotifications:
                                sendmail(receiver, subject, parseResult['warning'], False, None)
                        if config.enableRSSFeed:
                                feedXML.xpath('//channel')[0].append(genFeedItem(subject, parseResult['warning'], site['uri'], 0))
                else:
                        # otherwise, check which parts of the site were updated
                        changes = 0
                        fileContents = getFileContents(site['shortname'])
                        i = 0
                        for content in parseResult['contents']:
                                if content not in fileContents:
                                        changes += 1

                                        subject = '[' + site['shortname'] + '] ' + parseResult['titles'][i]
                                        print('    ' + subject)
                                        if config.enableMailNotifications and len(fileContents) > 0:
                                                sendmail(receiver, subject, content, (site.get('type', 'html') == 'html'), site['uri'])

                                        if config.enableRSSFeed:
                                                feedXML.xpath('//channel')[0].append(genFeedItem(subject, content, site['uri'], changes))
                                i += 1


                        if changes > 0:
                                storeFileContents(site['shortname'], parseResult)
                                print('        ' + str(changes) + ' updates')
 
        # store feed
        if config.enableRSSFeed:
                for o in feedXML.xpath('//channel/item[position()<last()-' + str(config.maxFeeds - 1) + ']'):
                        o.getparent().remove(o)
                file = open(config.rssfile, 'w')
                file.write(etree.tostring(feedXML, pretty_print=True, xml_declaration=True, encoding=defaultEncoding).decode(defaultEncoding))
                file.close()


if __name__ == "__main__":

        configMod = 'config'
        dryrun = None

        try:
                opts, args = getopt.getopt(sys.argv[1:], 'hc:d:', ['help', 'config=', 'dry-run='])
        except getopt.GetoptError:
                print('Usage: mwc.py --config=config --dry-run=shortname')
                sys.exit(1)
        for opt, arg in opts:
                if opt == '-h':
                        print('Usage: mwc.py --config=config')
                        exit()
                elif opt in ('-c', '--config'):
                        configMod = arg
                elif opt in ('-d', '--dry-run'):
                        dryrun = arg

        config = importlib.import_module(configMod)

        if dryrun:
                for site in config.sites:
                        if site['shortname'] == dryrun:
                                parseResult = parseSite(site)
                                print(parseResult)
                                break
        else:
                try:
                        pollWebsites()
                except:
                        msg = str(sys.exc_info()[0]) + '\n\n' + traceback.format_exc()
                        print(msg)
                        if config.receiver != '':
                                sendmail(config.receiver, '[mwc] Something went wrong ...', msg, False, None)

                if mailsession:
                        mailsession.quit()
                        mailsession = None