# -*- coding: utf-8 -*-
"""Copy of russian model testing.ipynb

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1c9k49wiWEvDa1zxIw65pUAsuzMlFn-tq
"""

#!pip install gradio
#!pip install translate

import nltk
from nltk.tokenize import word_tokenize
from nltk.corpus import stopwords

import pandas as pd

import numpy as np
import tensorflow as tf
from tensorflow.keras.preprocessing.text import Tokenizer
from tensorflow.keras.preprocessing.sequence import pad_sequences
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Embedding, LSTM, Dense, Bidirectional
from nltk.corpus import stopwords
from nltk.stem import WordNetLemmatizer
import nltk
from nltk.translate.bleu_score import sentence_bleu

nltk.download('stopwords')
nltk.download('wordnet')
nltk.download('punkt')

url = 'https://raw.githubusercontent.com/Obai33/NLP_PoemGenerationDatasets/main/russianpoems.csv'
text_data = pd.read_csv(url)

# removing duplicates and missing values
text_data.drop_duplicates(inplace = True)
text_data.dropna(inplace = True)
text_data

text_data = text_data['text']
text_data = text_data[500:700]

# Tokenization and lowercasing
tokenizer = Tokenizer()
tokenizer.fit_on_texts(text_data)

total_words = len(tokenizer.word_index) + 1

input_sequences = []
for line in text_data:
	token_list = tokenizer.texts_to_sequences([line])[0]
	for i in range(1, len(token_list)):
		n_gram_sequence = token_list[:i+1]
		input_sequences.append(n_gram_sequence)


# pad sequences
max_sequence_len = 100
input_sequences = np.array(pad_sequences(input_sequences, maxlen=max_sequence_len, padding='pre'))

# create predictors and label
xs, labels = input_sequences[:,:-1],input_sequences[:,-1]

ys = tf.keras.utils.to_categorical(labels, num_classes=total_words)

import requests
# URL of the model
url = 'https://github.com/Obai33/NLP_PoemGenerationDatasets/raw/main/modelrus1.h5'
# Local file path to save the model
local_filename = 'modelrus1.h5'

# Download the model file
response = requests.get(url)
with open(local_filename, 'wb') as f:
    f.write(response.content)

# Load the pre-trained model
model = tf.keras.models.load_model(local_filename)

# Import the necessary library for translation
import translate

# Function to translate text to English
def translate_to_english(text):
    translator = translate.Translator(from_lang="ru", to_lang="en")
    translated_text = translator.translate(text)
    return translated_text

def generate_russian_text(seed_text, next_words=50):
    generated_text = seed_text
    for _ in range(next_words):
      token_list = tokenizer.texts_to_sequences([generated_text])[0]
      token_list = pad_sequences([token_list], maxlen=max_sequence_len-1, 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
      generated_text += " " + output_word

      '''
        token_list = tokenizer.encode(generated_text, add_special_tokens=False)
        token_list = pad_sequences([token_list], maxlen=max_sequence_len-1, padding='pre')
        predicted = np.argmax(model.predict(token_list), axis=-1)
        output_word = tokenizer.decode(predicted[0])
        generated_text += " " + output_word
      '''
    #reconnected_text = generated_text.replace(" ##", "")
    t_text = translate_to_english(generated_text)
    return generated_text, t_text

import gradio as gr

# Update Gradio interface to include both Arabic and English outputs
iface = gr.Interface(
    fn=generate_russian_text,
    inputs="text",
    outputs=["text", "text"],
    title="Russian Poetry Generation",
    description="Enter Russian text to generate a small poem.",
    theme="compact"
)
# Run the interface
iface.launch()