#!/usr/bin/env python
# -*- coding: utf8 -*-

import sys
import codecs
import annotate.compound as compound
import re

def read_csv(num_of_fields, tolerates=-1):
    if tolerates == -1:
        tolerates = num_of_fields
    for line in sys.stdin:
        e = line[:-1].split('\t')
        if len(e) == num_of_fields:
            yield e
        elif len(e) >= tolerates and len(e) < num_of_fields:
            f = [""] * (num_of_fields - len(e))
            yield e + f
    return

def read_csv_file(filepath, num_of_fields, tolerates=-1):
    if tolerates == -1:
        tolerates = num_of_fields
    with codecs.open('saldoexempel.txt', encoding='utf-8') as f:
        for line in f:
            line = line.encode("utf-8")
            e = line[:-1].split('\t')
            if len(e) == num_of_fields:
                yield e
            elif len(e) >= tolerates and len(e) < num_of_fields:
                f = [""] * (num_of_fields - len(e))
                yield e + f
    return

def jsitem(key, value):
    if type(value) == list:
        realvalue = '[' + ", ".join(map(lambda x: '"' + x + '"', value)) + "]"
    else:
        if value[0] == '"':
            value = value[1:-1]
        realvalue = '"' + value + '"'
    return '"' + key + '" : ' + realvalue

if __name__ == '__main__':
    # First open SweFN
    lus = {}
    for (name, _, _, examples_raw, _, _, _, _, _, proplus, _, _, _, _) in read_csv(num_of_fields=14):
        #examples_raw = examples_raw.decode("utf-8")
        lus_block = proplus#.decode("utf-8")
        lu_lines = [line.strip() for line in lus_block.split(";;")]
        for lu_line in lu_lines:
            if len(lu_line) != 0 and " " in lu_line:
                (pos, lus_raw) = tuple(lu_line.strip().split(" ", 1))
                newlus = [lu.strip() for lu in lus_raw.split(",")]
                for nlu in newlus:
                    nlu = nlu.strip()
                    if nlu:
                        if nlu not in lus:
                            lus[nlu] = {}
                            lus[nlu]["frame"] = []
                        if name not in lus[nlu]:
                            lus[nlu]["frame"].append(name)
    # Then open saldoe.txt and get the remaining information

    for (sense, prdesc, scdesc, pos, info, example) in read_csv_file("saldoexempel.txt", num_of_fields=6):
        sense = sense.strip()
        example = example.strip()
        if "+" in info:
            # It is a New LU
            if sense not in lus:
                lus[sense] = {}
            lus[sense]["pos"] = pos
            lus[sense]["pr"] = prdesc
            lus[sense]["sc"] = scdesc
            lus[sense]["info"] = info.replace('"', "&quot;")
            lus[sense]["example"] = example.replace('"', "&quot;")


    lines = []
    for lu in lus.keys():
        obje = lus[lu]
        lines.append('"' + lu + '" : {' + ", ".join([jsitem(x, obje[x]) for x in obje.keys()]) + '}')
    #for lu in lus.keys():
    #    lines.append('"' + lu + '" : [' + ", ".join(map(lambda x: '"' + x + '"', lus[lu])) + "]")
    output = 'generated_swefn_newlus = {' + ",\n".join(lines) + '};'
    print output#.decode("utf-8")