zemberek-python ile sözcük üretme (Word Generation)

zemberek_python_word_gen.py

import pprint

from zemberek.morphology import TurkishMorphology
from zemberek.morphology.lexicon import RootLexicon
from zemberek.morphology.morphotactics.turkish_morphotactics import get_morpheme_map

MORPHOLOGY = TurkishMorphology.create_with_defaults()

def generate_noun(word, morphology):
    morpheme_map = get_morpheme_map()
    def mm(id):
        return morpheme_map.get(id)

    # bk. https://docs.google.com/document/d/1bUW6i5KrjkeID8GD7rs8tbGWapZooLmt2gmdAUvPHW8
    # https://github.com/ahmetaa/zemberek-nlp/tree/master/morphology#word-generation
    # https://github.com/ahmetaa/zemberek-nlp/blob/master/examples/src/main/java/zemberek/examples/morphology/GenerateWords.java
    number = [mm("A1sg"), mm("A2sg"), mm("A3sg"), mm("A1pl"), mm("A2pl"), mm("A3pl")]
    possessives = [mm("P1sg"), mm("P2sg"), mm("P3sg"), mm("P1pl"), mm("P2pl"), mm("P3pl"), mm("Pnon")]
    cases = [mm("Dat"), mm("Loc"), mm("Abl"), mm("Gen"), mm("Acc"), mm("Ins"), mm("Nom")]
    derivational = [mm("Rel")] # masa-m-da-`ki`, dün-`kü`
    
    # bk. resources/lexicon.csv
    Noun_Lexicon_Suffix = ["Noun", "Noun_1", "Noun_2", "Noun_Time", "Noun_Time_1",
                           "Adj_Prop", "Adj", "Adv_Prop", "Adv_Time", "Pron", "Pron_Demons", 
                           "Pron_Demons_1", "Pron_Pers", "Pron_Quant", "Pron_Ques", "Pron_Reflex"]

    # morphology = TurkishMorphology.create_with_defaults()
    lex = RootLexicon.get_default()
    # simple_id = {k.split("_")[0]:k for k in lex.id_map.keys()}
    # word_item = lex.get_item_by_id(f"{word}_Noun_2")
    noun_res = {}
    words = []
    
    for nls in Noun_Lexicon_Suffix:
        word_nls = f"{word}_{nls}"
        if (w := lex.get_item_by_id(word_nls)):
            words.append((w, word_nls))
    
    for w in words:
        temp = []
        for number_m in number:
            for possessive_m in possessives:
                for case_m in cases:
                    temp += morphology.word_generator.generate(w[0], (number_m, possessive_m, case_m))
        for number_m in number:
            for possessive_m in possessives:
                for case_m in cases:
                    for deriv_m in derivational:
                        temp += morphology.word_generator.generate(w[0], (number_m, possessive_m, case_m, deriv_m))
        noun_res[w[1]] = [x.surface for x in temp]
    return noun_res

def generate_verb(verb, morphology):
    morpheme_map = get_morpheme_map()
    def mm(id):
        return morpheme_map.get(id)

    positiveNegatives = [mm("Verb"), mm("Neg")] # mm("Verb") -> positive yerine
    times = [mm("Imp"), mm("Aor"), mm("Past"), mm("Prog1"), mm("Prog2"), mm("Narr"), mm("Fut"), mm("Neces"), mm("Opt"), mm("Cond")] # cond bazen çalışmıyor
    persons = [mm("A1sg"), mm("A2sg"), mm("A3sg"), mm("A1pl"), mm("A2pl"), mm("A3pl")]
    Verb_Lexicon_Suffix = "Verb"

    # morphology = TurkishMorphology.create_with_defaults()
    lex = RootLexicon.get_default()
    # simple_id = {k.split("_")[0]:k for k in lex.id_map.keys()}
    word_item = lex.get_item_by_id(f"{verb}_{Verb_Lexicon_Suffix}")
    if not word_item:
        return []

    res = []
    for positiveNegative_m in positiveNegatives:
        for time_m in times:
            for person_m in persons:
                res += morphology.word_generator.generate(word_item, (positiveNegative_m, time_m, person_m))

    result = [x.surface for x in res]
    return result

def noun_verb_other(word, morphology):
    # morphology = TurkishMorphology.create_with_defaults()
    result = morphology.analyze(word)
    analysis = [(x.get_ending(), x.get_stem()) for x in result]
    return analysis


pp = pprint.PrettyPrinter(indent=2)
res = generate_noun("bilinç", MORPHOLOGY)
pp.pprint(res)

res_verb = generate_verb("savaşmak", MORPHOLOGY)
pp.pprint(res_verb)