genwordmodel.py

import sys
import os
import textinput
import gensim
import string
from pathlib import Path


BASE_DIR = Path(__file__).resolve().parent
MODEL_PATH = BASE_DIR / "MODELS"
MODEL_WORDS = 'model.words'

MAX_WORD2VEC_WINDOW = 10
WORD2VEC_SG = 1
WORD2VEC_SIZE = 300
WORD2VEC_MINWORD_COUNT = 1

def build_word_model(word_list):

    print('building word models')

    model = gensim.models.Word2Vec(word_list, 
        min_count = WORD2VEC_MINWORD_COUNT, 
        vector_size = WORD2VEC_SIZE, 
        window = MAX_WORD2VEC_WINDOW,
        sg = WORD2VEC_SG)
    print(MODEL_PATH)

    MODEL_PATH.mkdir(parents=True, exist_ok=True)
    path = MODEL_PATH / f"{MODEL_WORDS}.{MAX_WORD2VEC_WINDOW}"
    print(f'Saving word model: {path}')
    model.save(str(path))


if __name__ == '__main__':

    textinput.load_bible()
    build_word_model(textinput.AllStoppedWords)

    print('done')