import streamlit as st
import pandas as pd
import numpy as np
import re
import pickle
from tensorflow.keras.preprocessing.sequence import pad_sequences
from tensorflow.keras.models import load_model

# Modeli yükle
model = load_model('sherlock_model.h5')

tokenizer = pickle.load(open( "tokenizer.pkl", "rb"))
word_index_df = pd.read_csv("word_index.csv", header=None, index_col=0)
word_index = {}
for word, index in word_index_df.to_dict()[1].items():
    word_index[word] = int(index)
tokenizer.word_index = word_index

st.title('Next Word Generator :writing_hand:')
st.write("This app predicts the next word using a model trained on words from the book 'THE ADVENTURES OF SHERLOCK HOLMES'.") 
st.write("Write a few words and indicate how many words you want them to guess.")

def next_words(seed_text, n):
    for _ in range(n):
        token_list = tokenizer.texts_to_sequences([seed_text])[0]
        token_list = pad_sequences([token_list], maxlen=17, padding='pre')
        predicted = np.argmax(model.predict(token_list), axis=-1)
        output_word = ""
        for word, index in tokenizer.word_index.items():
            if index == predicted:
                output_word = word
                break
        seed_text += " " + output_word
    return seed_text

# Giriş metnini al
text = st.text_area("Enter text", height=80)
n=st.number_input("Word number", 1,100)

if st.button("Predict"):
    sonuc=next_words(text, n)
    st.info(f'Prediction : {sonuc}')