#!/usr/bin/env python

from subprocess import Popen, PIPE
import re

class SVN:

    def __init__(self,address):
        stdin = ''
        command = Popen(['svn','log',address], 
                        shell=False, stdin=PIPE, 
                        stdout=PIPE) #,stderr=PIPE)
        (stdout, _) = command.communicate(stdin)
        revs = re.findall(r'(r\d+)|(\d\d\d\d-\d\d-\d\d)|(\d\d:\d\d)',stdout.decode('UTF-8'))
        self.list_of_revisions = []
        index = 0
        while index < len(revs):
            self.list_of_revisions.append((revs[index][0],
                                           revs[index+1][1],
                                           revs[index+2][2]))
            index += 3
    
    def revisions(self):
        return self.list_of_revisions
    
    def _sort_neg_pos(self,lst):
        positive = set()
        negative = set()
        for change in lst:
            (pos,neg) = (set(),set())
            for line in change:
                if line[0] == '+':
                    pos.add(line[1:].split('\t')[0])
                else:
                    neg.add(line[1:].split('\t')[0])
            if len(neg) == 0:
                positive.update(pos)
            elif len(pos) == 0:
                negative.update(neg)
        return (positive,negative)

    def difference(self,revision1,revision2):
        stdin = ''
        command = Popen(['svn', 'diff','-r',
                     '%s:%s' % (revision1,revision2),address], 
                        shell=False, stdin=PIPE, 
                        stdout=PIPE) #,stderr=PIPE)
        (stdout, _) = command.communicate(stdin)
        return self._sort_neg_pos(
            [[c for c in re.split(r'\n',s) 
              if c != '' and c[0] in '+-'] 
             for s in re.split(r'@@.+@@\n',stdout.decode('UTF-8'))])


def ref(address,item):
    if address == '':
        return item
    else:
        return '<a href="%s%s">%s</a>' % (address,item,item)

def to_word(lex):
    return re.split(r'\.\.',lex)[0].replace('_',' ')

def produce_rss(address,number=20,ref_address=''):
    svn = SVN(address)
    revs =  svn.revisions()
    result = '''<?xml version="1.0" encoding="utf-8" ?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <atom:link href="http://spraakbanken.gu.se/rss/saldo_dev.xml" rel="self" type="application/rss+xml" />
    <title>SALDO dev</title>
    <link>http://spraakbanken.gu.se/saldo</link>
    <description>SALDO dev</description>
    <image>
        <title>SALDO dev</title>
        <url>http://spraakbanken.gu.se/sites/spraakbanken.gu.se/files/img/saldo/saldo.gif</url>
        <link>http://spraakbanken.gu.se/saldo</link>
    </image>\n'''
    count = 0
    for i in range(len(revs)-1):
        (rev,d,t) = revs[i]
        (rev2,_,_) = revs[i+1]
        (pos,neg) = svn.difference(rev2,rev)
        if len(pos) > 0:
            count += 1
            if len(pos) > 1:
                dots = ' ...'
            else:
                dots = ''
            result += '  <item>\n'
            result += '   <title>%s%s (%s %s)</title>\n' % (list(pos)[0],dots,d,t) 
            result += '   <guid isPermaLink="false">%s</guid>\n' % (rev)
            result += '   <link>http://spraakbanken.gu.se/eng/research/saldo/history#%s</link>\n' % (rev) 
            result += '   <description>%s</description>\n' %(", ".join([p for p in sorted(pos)]))
            result += '  </item>\n'
            if count >= number:
                break
    result += '</channel>\n'
    result += '</rss>'
    return result

if __name__ == '__main__':
    address = 'https://svn.spraakdata.gu.se/repos/sblex/sal/trunk/saldo30.txt'
    print produce_rss(address, number=20, ref_address='http://spraakbanken.gu.se/ws/saldo-ws/lid/html/').encode('UTF-8')