#!/usr/bin/python2.5
# coding=utf-8
# -*- encoding: utf-8 -*-

import sys, codecs, copy, Ft, os;
from Ft.Xml.Domlette import NonvalidatingReader;
from Ft.Xml.XPath import Evaluate;

sys.stdout = codecs.getwriter('utf-8')(sys.stdout);
sys.stderr = codecs.getwriter('utf-8')(sys.stderr);

if len(sys.argv) < 2: #{
    print 'python apertium2extract.py <dix file>';
    sys.exit(-1);
#}

dictionary = sys.argv[1];

if dictionary == os.path.basename(dictionary): #{
    dictionary = os.getcwd() + '/' + dictionary;
#}

doc = NonvalidatingReader.parseUri('file:///' + dictionary);
path = '/dictionary/pardefs/pardef';

paradigms = {};
#categories = ['__n', '__adj', '__vblex'];
categories = ['__n', '__adj'];

for node in Ft.Xml.XPath.Evaluate(path, contextNode=doc): #{
        pardef = node.getAttributeNS(None, 'n');

    if pardef not in paradigms: #{
        paradigms[pardef] = [];
    #}
    selected = 0;
    for category in categories: #{
        if pardef.count(category) > 0: #{
            selected = 1;
        #}
    #}

    if selected < 1: #{
        continue;
    #}

    for child in Ft.Xml.XPath.Evaluate('.//e', contextNode=node): #{
        for pair in Ft.Xml.XPath.Evaluate('.//p', contextNode=child): #{
            suffix = '';
            left = Ft.Xml.XPath.Evaluate('.//l', contextNode=pair)[0].firstChild;

            if type(left) != type(None): #{
                suffix = left.nodeValue;
            else: #{
                suffix = ''
            #}


            symbols = '';
            right =  Ft.Xml.XPath.Evaluate('.//r', contextNode=pair)[0];
            for sym in Ft.Xml.XPath.Evaluate('.//s', contextNode=right): #{
                symbol = '';
                if type(sym) != type(None): #{
                    symbol = sym.getAttributeNS(None, 'n');
                #}
                symbols = symbols + '.' + symbol;
            #}

            paradigms[pardef].append(suffix);
        #}
    #}
#}

universal_set = []

for paradigm in paradigms.keys(): #{
    for p in paradigms[paradigm]: #{
        universal_set.append(p);
    #}
#}

universal_set = set(universal_set);

for paradigm in paradigms.keys(): #{
    #paradigm rna__vblex =
    #        x
    #        { x+"ð" & x+"ði" & x+"ðu" & x+"ður" & x+"r" & ~(x+"dur" | x+"t")} ;

#   if paradigm.count('/') > 0: #{
        #continue;
    #}
    lset = len(set(paradigms[paradigm]));
    if lset < 2: #{
        continue;
    #}

    print '-- ' + paradigm;
#   print 'paradigm ' + paradigm.encode('ascii', 'ignore') + ' =';
    print 'paradigm ' + paradigm + ' =';
    print '\t' + 'x {';

    stems = '\t\t';
    count = 0;
    idx = 0;
    for pair in set(paradigms[paradigm]): #{
        if len(pair) >= 1: #{
            add = 'x+"' + pair + '" ';
            if idx != lset-1: #{
                add = 'x+"' + pair + '" & ';
            #}
            stems = stems + add;
        #}

        if count == 6 and count != lset -1: #{
            stems = stems + '\n\t\t';
            count = 0;
        #}
        count = count + 1;
        idx = idx + 1;
    #}
    complement = universal_set - set(paradigms[paradigm]);
    stems = stems + ' & \n\t\t~(';
    count = 0;
    for s in complement: #{
        stems = stems + 'x+"' + s + '" | ';
        if count == 6 and count != lset -1: #{
            stems = stems + '\n\t\t';
            count = 0;
        #}
        count = count + 1;
    #}
    stems = stems + ')\t' + '};';
    print stems.replace('| )', ')').replace('| \n\t\t)', ')');
    print '';
#}