app.py

from flask import Flask, request
import pandas as pd
import numpy as np
import json
import pickle
import os
import requests
from os import getenv
from dotenv import load_dotenv
from sklearn.preprocessing import StandardScaler
from .preprocess import Magic
from .sentiment import TwitterSentiment

load_dotenv()

def create_app():
    '''
        Create Flask App
    '''
    APP = Flask(__name__)

    model = pickle.load(open("model.p", "rb"))

    def generate_df(ticker):

        macd = 'https://www.alphavantage.co/query?function=MACD&symbol=' + ticker + '&interval=daily&series_type=open&apikey=SXG08DL4S2EW8SKC'
        response1 = requests.get(macd)
        if "Note" in response1.json().keys():
            return (response1.json()["Note"])
        df_macd = pd.DataFrame.from_dict(response1.json()['Technical Analysis: MACD']).T

        stoch = 'https://www.alphavantage.co/query?function=STOCH&symbol=' + ticker + '&interval=daily&apikey=SXG08DL4S2EW8SKC'
        response2 = requests.get(stoch)
        if "Note" in response2.json().keys():
            return (response2.json()["Note"])
        df_stoch = pd.DataFrame.from_dict(response2.json()['Technical Analysis: STOCH']).T

        aroon = 'https://www.alphavantage.co/query?function=AROONOSC&symbol=' + ticker + '&interval=daily&time_period=10&apikey=SXG08DL4S2EW8SKC'
        response4 = requests.get(aroon)
        if "Note" in response4.json().keys():
            return (response4.json()["Note"])
        df_aroon = pd.DataFrame.from_dict(response4.json()['Technical Analysis: AROONOSC']).T

        dx = 'https://www.alphavantage.co/query?function=DX&symbol=' + ticker + '&interval=daily&time_period=10&apikey=SXG08DL4S2EW8SKC'
        response5 = requests.get(dx)
        if "Note" in response5.json().keys():
            return (response5.json()["Note"])
        df_dx = pd.DataFrame.from_dict(response5.json()['Technical Analysis: DX']).T

        url = 'https://www.alphavantage.co/query?function=TIME_SERIES_DAILY&symbol=' + ticker + '&interval=5min&outputsize=full&apikey=SXG08DL4S2EW8SKC'
        response6 = requests.get(url)
        if "Note" in response6.json().keys():
            return (response6.json()["Note"])
        df = pd.DataFrame.from_dict(response6.json()['Time Series (Daily)']).T

        df = df.join(df_macd)
        df = df.join(df_stoch)
        df = df.join(df_aroon)
        df = df.join(df_dx)

        return df

    @APP.route('/')
    @APP.route('/api',methods=['POST'])
    def hello_world(ticker):
        inp = ticker

        if request.method == 'POST':
            inp = request.values['ticker']
        market_df = generate_df(inp)
        if type(market_df)==str:
            return market_df

        market_df = market_df.dropna()

        X = market_df[['5. volume', 'MACD', 'AROONOSC','MACD_Hist', 'MACD_Signal', 'DX', 'SlowD', 'SlowK']]
        sc = StandardScaler()
        X = sc.fit_transform(X)
        y_prebro = model.predict_proba(X[0].reshape(1, -1))

        magic = Magic(ticker)
        historical = magic.output_historical()
        future = magic.output_future()

        twitter = TwitterSentiment(ticker)
        sentiment = twitter.output_twitter()

        json_obj = {'TA': {'sell':y_prebro[0][0], 'hold':y_prebro[0][1], 'buy':y_prebro[0][2]},
                    'Sentiment': {'sell':sentiment['Sell'], 'hold':sentiment['Hold'], 'buy':sentiment['Buy']},
                    'Historical': {'sell':historical['Sell'], 'hold':historical['Hold'], 'buy':historical['Buy']},
                    'Future': {'sell':future['Sell'], 'hold':future['Hold'], 'buy':future['Buy']}
                    }

        response = json.dumps(json_obj)

        return response