from flask import Flask
from flask import request
from flask import render_template
from flask import redirect
from flask import jsonify
from flask import send_from_directory
import json
import numpy as np
import moduli
#from hugchat import hugchat
#from hugchat.login import Login
from numpy.linalg import norm
import datetime
import tensorflow as tf
import tensorflow_hub as hub
import tensorflow_text
from functools import wraps
import os
import pickle

model_url = "./hub_model/"
hub_layer = hub.load(model_url)

app = Flask(__name__)

def support_jsonp(func):
    """Wraps JSONified output for JSONP requests."""
    @wraps(func)
    def decorated_function(*args, **kwargs):
        callback = request.args.get('callback', False)
        if callback:
            resp = func(*args, **kwargs)
            resp.set_data('{}({})'.format(
                str(callback),
                resp.get_data(as_text=True)
            ))
            resp.mimetype = 'application/javascript'
            return resp
        else:
            return func(*args, **kwargs)
    return decorated_function

@app.route("/")
def flask_app():
    return '<br>' \
    '<h3><a href="http://18.102.72.47:5000" target="_blank">Hai acceso il server! Bene!&#128521 Adesso vai alla pagina La7 DEMO!!</a></h3>' \
   

@app.route('/broadcast_labels',methods=['POST','GET'])
@support_jsonp
def broadcast_labels():
    #if request.method == 'GET':
    s=request.args.get('S')
    dati=moduli.la7_labels(s)
    output=dati
    response = jsonify(output)
    return response

@app.route('/requestsSearch',methods=['POST','GET'])
@support_jsonp
def tasksSearch():
    domanda=request.args.get('frase').strip()
    s=request.args.get('S').strip()
    path=s
    #dati=moduli.la7_Search(s,domanda,hub_layer)
    print('domanda',domanda)
    dati=[]
    with open ('static/dati/'+str(path)+'/db_relatori_finale_emb', 'rb') as fp:
        word = pickle.load(fp) 
    word_orig=np.array(word, dtype="object")
    message_embeddings = hub_layer(domanda)[0].numpy()
    ris=[]
    for n,w in enumerate(word):
        contesto_embeddings=np.array(w[6])
        cosine = np.dot(message_embeddings,contesto_embeddings)/(norm(message_embeddings)*norm(contesto_embeddings))
        ris.append(cosine)
    ris_sort=np.argsort(ris)[::-1]
    l=0
    for n in ris_sort:
        if l==0:
          testo=str(word[n][5])
          testo_chi=str(word[n][2])+':'+str(word[n][1])
        else:
          testo=testo+' '+word[n][5]
          testo_chi=testo_chi+' --- '+str(word[n][2])+' '+str(word[n][1])
        l=len(testo)
        if l>800:
          break

    d={}
    d['text']=str(testo)
    d['chi']=str(testo_chi)
    dati.append(d)
    output=dati
    response = jsonify(output)
    return response



if __name__ == "__main__":
    app.run(host='0.0.0.0',  port=7860)