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

# Fonction summarize
def summarizer_FS(Text):
  # Tokeniser article_text en phrase
  sentences_list = nltk.sent_tokenize(Text)
  # Stopwords
  stopwords = nltk.corpus.stopwords.words('english')
  # Dictionnaire de fréquences des mots
  word_frequencies = {}
  # calculer les fréquences des mots
  for word in nltk.word_tokenize(Text):
    if word not in stopwords:
      if word not in word_frequencies.keys():
        word_frequencies[word] = 1
      else:
        word_frequencies[word] += 1
  # Fréquence maximale de word_frequencies
  freq_maximum =max(word_frequencies.values())
  # Calculer la fréquence pondérée de chaque mot
  for word in word_frequencies.keys():
    word_frequencies[word] =  word_frequencies[word]/freq_maximum
  # Liste des scores de chaque phrase
  sentence_scores = {}
  # Calculer le score de chaque phrase
  for sent in sentences_list:
    for word in nltk.word_tokenize(sent.lower()):
      if word in word_frequencies.keys():
        # if len(sent.split(' ')) < 30:
          if sent not in sentence_scores.keys():
            sentence_scores[sent]= word_frequencies[word]
          else:
            sentence_scores[sent] +=word_frequencies[word]
  # Ordonner les phrases par pondération et recupérer les 10 premières phrases
  summary_sentences = sorted(sentence_scores, key = sentence_scores.get, reverse = True)[:5]
  # regrouper ensemble les phrases qui ont les poids les plus élévés
  summary = ' '.join(summary_sentences)

  return summary

# Importer pipeline
from transformers import pipeline
# Modèle de summarization
summarizer = pipeline(task="summarization",
                      model="facebook/bart-large-cnn")

# Définir une fonction summarize_t
def summarizer_t(input):
    output = summarizer(input)
    return output[0]['summary_text']


# Déployer
import gradio as gr
import os
# Création de blocks
demo = gr.Blocks()

summarizer1 = gr.Interface(fn=summarizer_FS,
                    inputs=[gr.Textbox(label="Text to summarize", lines=6)],
                    outputs=[gr.Textbox(label="Result", lines=3)],
                    title="Text summarization with NLTK",
                    description="Summarize any text from scracth under the hood!"
                   )
summarizer2 = gr.Interface(fn=summarizer_t,
                    inputs=[gr.Textbox(label="Text to summarize", lines=6)],
                    outputs=[gr.Textbox(label="Result", lines=3)],
                    title="Text summarization with bart-large-cnn ",
                    description="Summarize any text using facebook/bart-large-cnn model under the hood!"
                   )

with demo:
    gr.TabbedInterface(
        [summarizer1,
         summarizer2],
        ["Summarize from scrath",
         "Summarize using bart"],
    )

demo.launch()