# $Id: markow.py 750 2005-08-13 20:28:39Z md $

"""Class for handling markow chains.

A little amusement, find more at http://c0re.jp/

  --drt@un.bewaff.net

"""

from __future__ import generators
import random
import pprint

_NOWORD = ''

class markow:
    """Class for using a simple text based markow chain.

    e.g.
    # create a new chain
    m = markow()
    # initialize the cahin with some data
    m.build('laber rababer and about 1000 additional words go here ...')
    # output some text based on thet chain
    print ' '.join(m.putput)
    """
    
    def __init__(self, prefixlen=3):
        self.startstate = [_NOWORD] * prefixlen
        # how to copy by calue in python?
        self.statetab = {}

    def build(self, s):
        state = self.startstate * 1
        for x in s.split():
            if tuple(state) not in self.statetab:
                self.statetab[tuple(state)] = []
            self.statetab[tuple(state)].append(x)
            state.pop(0)
            state.append(x)

        # add tail
        if tuple(state) not in self.statetab:
            self.statetab[tuple(state)] = []
        self.statetab[tuple(state)].append(_NOWORD) 

    def buildfromfile(self, fd):
        state = self.startstate * 1
        for line in fd.readlines():
            for x in line.split():
                if tuple(state) not in self.statetab:
                    self.statetab[tuple(state)] = []
                self.statetab[tuple(state)].append(x)
                state.pop(0)
                state.append(x)

        # add tail
        if tuple(state) not in self.statetab:
            self.statetab[tuple(state)] = []
        self.statetab[tuple(state)].append(_NOWORD) 


    def output(self, maxlen=50000):
        state = self.startstate * 1
        l = []
        for i in range(maxlen):
            suffix  = random.choice(self.statetab[tuple(state)])
            if suffix == _NOWORD:
                return l
            l.append(suffix)
            state.pop(0)
            state.append(suffix)
    

if __name__ == '__main__':
    # run an example
    # produce an comp.lang.python lookalike message
    import nntplib

    # change this to your server name. Append username and password at the end if needed
    # eg s = nntplib.NNTP('news.cis.dfn.de', 119, 'user', 'pass')
    s = nntplib.NNTP('news.t-online.de', 119)
    resp, count, first, last, name = s.group('comp.lang.python')
    resp, subs = s.xhdr('subject', first + '-' + last)
    m = markow()
    for id, sub in subs[-10:]:
        m.build('\n'.join(s.body(id)[3]))

    column = 0
    for x in m.output():
        if x == '>' or column > 70:
            print
            column = 0
        print x,
        column += len(x) + 1
    s.quit()