Subversion Repositories navi

#!/usr/bin/python

# -*- coding: utf-8 -*-

#    This file is part of Tsim Apiak.

#

#    Tsim Apiak is free software: you can redistribute it and/or modify

#    it under the terms of the GNU General Public Licence as published by

#    the Free Software Foundation, either version 3 of the Licence, or

#    (at your option) any later version. 

# 

#    In addition to this, you must also comply with clause 4 of the

#    Apache Licence, version 2.0, concerning attribution. Where there

#    is a contradiction between the two licences, the GPL

#    takes preference.

#

#    Tsim Apiak is distributed in the hope that it will be useful,

#    but WITHOUT ANY WARRANTY; without even the implied warranty of

#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

#    GNU General Public License for more details.

#

#    You should have received a copy of the GNU General Public License

#    along with Tsim Apiak.  If not, see <http://www.gnu.org/licenses/>.

import dbconnector

import parsenum

import re

#wordlist = [{"id": 0, "navi": u"tawtute", "infix": u"tawtute", "type": u"n."}] + dbconnector.getnavilist() + [{"id": 0, "navi": u"na'vi", "infix": u"na'vi", "type": u"n."}] # XXX HACK - extra proper nouns

wordlist = dbconnector.getnavilist()

BROKENWORDS = ((u"sami", u"si", u"", u"am", u"", (()), (()), False), (u"to", u"to", u"", u"", u"", (()), (()), False), (u"frato", u"to", u"", u"", u"", [[u"fra"]], (()), False), (u"mengenga", u"ngenga", u"", u"", u"", [[u"me"]], (()), False), (u"pxengenga", u"ngenga", u"", u"", u"", [[u"pxe"]], (()), False), (u"kìmä", u"kä", u"", u"ìm", u"", (()), (()), False))

# XXX HACK - These are words that are either not in Eana Eltu, or that get interpreted wrongly for whatever reason. The latter should be removed from this list when the parser gets more sophisticated. The former should also have an entry in the equivalent array in the translator! If it can take infixes, consider adding it to the main wordlist above (see the examples). The order is - original, Na'vi root, 0-pos infix, 1-pos infix, 2-pos infix, PREFIXES, suffixes. Things that can take affixes should go in the above list instead.

INFIXES1 = (u"awn", u"eyk", u"us", u"äp", u"")

INFIXES2 = (u"ìyev", u"iyev", u"ìmìy", u"arm", u"asy", u"ilv", u"ìmv", u"imv", u"ìrm", u"irv", u"ìsy", u"aly", u"ary", u"ìly", u"ìry", u"ìlm", u"alm", u"am", u"ay", u"er", u"ìm", u"iv", u"ìy", u"ol", u"")

INFIXES3 = (u"äng", u"ats", u"eiy", u"ei", u"uy", u"")

PREFIXES = (u"tsay", u"fray", u"say", u"fay", u"fra", u"pxe", u"fne", u"tsa", u"kel", u"lek", u"sa", u"pe", u"fe", u"le", u"nì", u"sä", u"tì", u"sì", u"ay", u"me", u"fì", u"ke", u"he", u"px", u"a", u"m", u"k")

ADPOSITIONS = (u"mungwrr", u"kxamlä", u"pximaw", u"pxisre", u"tafkip", u"nemfa", u"takip", u"mìkam", u"teri", u"fkip", u"luke", u"pxel", u"pxaw", u"rofa", u"ìlä", u"fpi", u"ftu", u"kip", u"lok", u"maw", u"sre", u"sìn", u"vay", u"eo", u"fa", u"hu", u"io", u"ka", u"mì", u"na", u"ne", u"ro", u"ta", u"uo", u"wä", u"äo", u"to", u"sì")

POSTFIXES = ADPOSITIONS + (u"tsyìp", u"eyä", u"ìri", u"aru", u"ati", u"ayä", u"ari", u"ay", u"ìl", u"it", u"lo", u"ri", u"ru", u"ti", u"ur", u"ve", u"yä", u"ya", u"tu", u"vi", u"yu", u"an", u"ng", u"ke", u"al", u"at", u"ar", u"ey", u"e", u"o", u"l", u"t", u"y", u"a", u"ä", u"r")

#prefixesn = ur"(?P<npr>(?:(?:fì|tsa)?(?:me|pxe|ay|fra)?|(?:fay)?|(?:tsay)?)(?:fne)?(?:tì|sä)?"

#prefixesv = ur"(?P<vpr>(?:nì|sä|tì|rä'ä |ke )?)"

LENIT = ((u"px", u"p"), (u"tx", u"t"), (u"kx", u"k"), (u"ts", u"s"), (u"t", u"s"), (u"p", u"f"), (u"k", u"h"), (u"'", u""))

def parseword(wordin):

    tempid = 0

    temptype = u""

    for brokenword in BROKENWORDS: # XXX HACK - this is all code to work around bugs that shouldn't exist

        if wordin[0] == brokenword[0]:

            for word in wordlist:

                if brokenword[1] == word["navi"]:

                    tempid = word["id"]

                    temptype = word["type"]

            return {"word": {"id": tempid, "navi": brokenword[1], "infix": u"", "type": temptype}, "pref": brokenword[5], "post": brokenword[6], "len": brokenword[7], "inf": (brokenword[2], brokenword[3], brokenword[4]) }

    ret = {"word": {"id": 0, "navi": u"[" + wordin[0] + u"]", "infix": u"", "type": u""}}

    for word in wordlist:

        word["navi"] = word["navi"].lower()

        foundit = True

        foundprefs = []

        foundposts = []

        lenited = False

        splitword = word["infix"].split(u" ")

        foundins = [u"", u"", u""]

        if len(wordin) < len(splitword):

            foundit = False

            continue

        for wor in range(len(splitword)):

            if not foundit:

                break

            foundprefs.append([])

            foundposts.append([])

            center = u""

            if u"<1>" in splitword[wor]:

                tempin1 = []

                tempin2 = []

                tempin3 = []

                for in1 in INFIXES1:

                    if in1 in wordin[wor]:

                        tempin1.append(in1)

                for in2 in INFIXES2:

                    if in2 in wordin[wor]:

                        tempin2.append(in2)

                for in3 in INFIXES3:

                    if in3 in wordin[wor]:

                        tempin3.append(in3)

                for in1 in tempin1:

                    for in2 in tempin2:

                        for in3 in tempin3:

                            if splitword[wor].replace(u"<1><2>", in1 + in2).replace(u"<3>", in3).replace(u"lll", u"l").replace(u"rrr", u"r") in wordin[wor]:

                                center = splitword[wor].replace(u"<1><2>", in1 + in2).replace(u"<3>", in3).replace(u"lll", u"l").replace(u"rrr", u"r")

                                foundins = [in1, in2, in3]

                                break

                        if center != u"":

                            break

                    if center != u"":

                        break

            else:

                if splitword[wor] in wordin[wor]:

                    center = splitword[wor]

                if center == u"" and (wordin[wor] == u"paya" or splitword[wor] != u"pxay"): # XXX HACK - workaround to fix pay being lenited pxay. Maybe fixable without hardcoding?

                    for i in LENIT:

                        temp = u""

                        if splitword[wor].startswith(i[0]):

                            temp = i[1] + splitword[wor][len(i[0]):]

                            if temp in wordin[wor]:

                                lenited = True

                                center = temp

                if center == u"":

                    if splitword[wor].endswith(u"nga"):

                        temp = splitword[wor][:-3] + u"ng"

                        if temp in wordin[wor]:

                            center = temp

                    if splitword[wor].endswith(u"po"):

                        temp = splitword[wor][:-2] + u"p"

                        if temp in wordin[wor]:

                            center = temp

                    if splitword[wor].endswith(u"tsa"):

                        temp = splitword[wor][:-3] + u"ts"

                        if temp in wordin[wor]:

                            center = temp

            if center == u"":

                foundit = False

                break

            temp = wordin[wor].split(center)

            if len(temp) != 2:

                foundit = False

                break

            pref, posf = temp

            last = u""

            while last != pref:

                last = pref

                for pre in PREFIXES:

                    if pref != u"":

                        if pref.endswith(pre):

                            if pre in foundprefs[wor]:

                                break

                            foundprefs[wor].append(pre)

                            pref = pref[:-len(pre)]

                            break

            if pref != u"":

                foundit = False

                foundprefs = []

                break

            last = u""

            while last != posf:

                last = posf

                for pos in POSTFIXES:

                    if posf != u"":

                        if posf.startswith(pos):

                            if pos in foundposts[wor]:

                                break

                            if pos != u"ä" or word["navi"] != u"pey": # XXX HACK - fix for peyä. THIS SHOULD NOT BE HERE!

                                foundposts[wor].append(pos)

                                posf = posf[len(pos):]

                                break

                            else:

                                break

            if posf != u"":

                foundit = False

                foundposts = []

                break

        if foundit == True:

            foundword = word

            break

    ret["pref"] = foundprefs

    ret["post"] = foundposts

    ret["inf"] = foundins

    ret["len"] = lenited

    if foundit == True:

        ret["word"] = foundword

    return ret

def parsesent(sent):

    sent = sent.strip().lower().replace(u"’", u"'")

    sent = re.sub(ur"[^\wìä' ]", u"", sent)

    sent = re.sub(ur"\ +", u" ", sent)

    sent = sent.split(u" ")

    ret = []

    left = len(sent)

    while left:

        word = parsenum.parse(sent[len(sent) - left])

        if word == None:

            word = parseword(sent[-left:])

        left -= len(word["word"]["navi"].split(" "))

        ret.append(word)

    return ret