LSI_prueba

from BusquedasSem import *
import pandas as pd
import seaborn as sns


def main():

    df = pd.read_csv('./client0-sort.csv')
    abstracts = df['Abstract'].values

    abstracts_aux = []
    for abstract in abstracts:
        text = minimizar(abstract)
        text = deletePunt(text=text)
        text = deleteStop(text=text, leng='english')
        #text = nltk.tokenize.word_tokenize(text)
        text = deleteWord('CD', text)
        text = stemmingLemmatizer(text)
        abstracts_aux.append(text)
    #print(abstracts_aux)

    words = getWordsText('explosive emulsion; plastic explosive; oil with water; robust')
    #print(words)
    #dictionary = gensim.corpora.Dictionary(abstracts_aux)

    lsi_score = LSIscore(words, abstracts_aux)
    #print(lsi_score)
    #print(type(lsi_score))
    #print(lsi_score[0])

    scores = []
    for i in lsi_score:
        scores.append(i)


    df_lsi_score = pd.DataFrame(scores, columns=['LSI Score'])
    df_abstracts = pd.DataFrame(abstracts, columns=['Abstract'])
    df_pca_abstracts = pd.concat([df_lsi_score, df_abstracts], axis=1)

    df = df_pca_abstracts.sort_values(['LSI Score'], ascending=False)
    df.to_csv('LSI_Sorted_Abstracts_prueba.csv')


if __name__ == "__main__":
    main()