import numpy as np
import os
import gradio as gr
import xgboost as xgb
from sklearn.feature_extraction.text import TfidfVectorizer

os.environ["WANDB_DISABLED"] = "true"

model_file_name = "xgb_reg.pkl"
vectorizer_file_name = 'vectorizer.pk'

#load
xgb_model_loaded = pickle.load(open(model_file_name, "rb"))
vectorizer_loaded = pickle.load(open(vectorizer_file_name, "rb"))

# predict
def predict_sentiment(predict_texts):
    predictions_loaded = xgb_model_loaded.predict(
    vectorizer_loaded.transform(predict_texts))
    return vectorizer_loaded
    #le.inverse_transform(predictions_loaded)

interface = gr.Interface(
    fn=predict_sentiment,
    inputs='text',
    outputs=['text'],
    title='Croatian Movie reviews Sentiment Analysis',
    examples= ["Volim kavu","Ne volim kavu"],
    description='Get the positive/neutral/negative sentiment for the given input.'
)

interface.launch(inline = False)