import gensim
import os
import nltk
from Sastrawi.Stemmer.StemmerFactory import StemmerFactory
import swifter
from textblob import TextBlob
from nltk.stem import WordNetLemmatizer
from gensim.models import LdaModel
nltk.download('wordnet')


def lemmatize_stemming(text):
    factory = StemmerFactory()
    stemmer = factory.create_stemmer()

    return stemmer.stem(WordNetLemmatizer().lemmatize(text, pos='v'))


def preprocess(text):
    result=[]
    for token in gensim.utils.simple_preprocess(text) :
        if token not in gensim.parsing.preprocessing.STOPWORDS and len(token) > 3:
            result.append(lemmatize_stemming(token))
    return result


def get_topik(data):
    dictionary = gensim.corpora.Dictionary(data)
    dictionary.filter_extremes(no_below=5, no_above=0.5)
    bow_corpus = [dictionary.doc2bow(text) for text in data]

    num_topics = 1
    ldamodel = LdaModel(bow_corpus, num_topics=num_topics, id2word=dictionary, passes=20, alpha='auto', eta='auto')
    topics = ldamodel.show_topics(num_topics=num_topics, num_words=5, log=False, formatted=False)
    return [topic[0] for topic in topics[0][1]]


def load_lexicon(file):
    df=open(file,"r",encoding="utf-8", errors='replace')
    data=df.readlines();df.close()
    return [d.strip().lower() for d in data]


def prediksi_sentiment(text):
    negasi_path = os.path.join(os.path.dirname(__file__), '..', 'assets', 'negasi.txt')
    fpos_path = os.path.join(os.path.dirname(__file__), '..', 'assets', 's-pos.txt')
    fneg_path = os.path.join(os.path.dirname(__file__), '..', 'assets', 's-neg.txt')
    positif, negatif, negasi = load_lexicon(fpos_path), load_lexicon(fneg_path), load_lexicon(negasi_path)

    posWords = []
    negWords = [w for w in negatif if w in text]
    # percent_positif = 0
    # percent_negatif = 0
    for w in positif:
        if w in text:
            negated = False
            for n in negasi:
                if n+' '+w in text:
                    negWords.append(n+' '+w)
                    negated = True
                    break
            if not negated:
                posWords.append(w)
    nPos, nNeg = len(posWords), len(negWords)
    # count = nPos + nNeg
    
    # if count != 0:
    #     percent_positif = round((nPos / count) * 100)
    #     percent_negatif = round((nNeg / count) * 100)

    if nPos>nNeg:
        return 1
    if nPos<nNeg:
        return -1
    else:
        return 0
    

def prediksi_sentiment_text_blob(text):
    analysis = TextBlob(text)

    if analysis.sentiment.polarity > 0.0:
        return 1
    elif analysis.sentiment.polarity == 0.0:
        return 0
    else:
        return -1