app.py

import pandas as pd
import streamlit as st
from vnstock import *
import seaborn as sns
import matplotlib.pyplot as plt
import plotly.express as px
import plotly.graph_objects as go
import statsmodels.api as sm
from statsmodels.tsa.arima.model import ARIMA
from prophet import Prophet
from datetime import datetime, timedelta
import pytz



start_date = str((datetime.now(pytz.timezone('Asia/Ho_Chi_Minh')) - timedelta(days=365)).strftime("%Y-%m-%d"))
end_date = str((datetime.now(pytz.timezone('Asia/Ho_Chi_Minh')) - timedelta(days=0)).strftime("%Y-%m-%d"))

def prophet_ts(symbol, periods = 10):
    data =  stock_historical_data(symbol=symbol, 
                            start_date=start_date, 
                            end_date=end_date, resolution='1D', type='stock')
    m = Prophet()
    pdf = pd.DataFrame()
    pdf['ds'] = data.index 
    pdf['y'] = data.close.values
    m.fit(pdf)
    future = m.make_future_dataframe(periods=periods)
    forecast = m.predict(future)
    fig = go.Figure()
    fig.add_trace(go.Scatter(x= pdf.ds, 
                            y=pdf.y,
                            name = f"{symbol}_true"
                            ))
    fig.add_trace(go.Scatter(x= forecast.ds, 
                                 y=forecast.yhat,
                                 name = f"{symbol}_pred"
                                ))
    return fig
class TS: 
    def __init__(self, symbol):
        self.symbol = symbol
    def get_data(self):
        data =  stock_historical_data(symbol=self.symbol, 
                            start_date=start_date, 
                            end_date=end_date, resolution='1D', type='stock')
        pdf = pd.DataFrame()
        pdf['ds'] = data.index 
        pdf['y'] = data.close.values
        return pdf
    def prophet(self, period = 28):
        df = self.get_data()
        model = Prophet()
        model.fit(df)
        future = model.make_future_dataframe(periods=period)
        forecast = model.predict(future)
        return self.viz(df, forecast)
    def viz(self, data, future):
        fig = go.Figure()
        fig.add_trace(go.Scatter(x= data.ds, 
                                y=data.y,
                                name = f"{self.symbol}"
                                ))
        fig.add_trace(go.Scatter(x= future.ds, 
                                     y=future.yhat,
                                     name = f"{self.symbol}_pred"
                                    ))
        return fig
    def arima(self, period = 3):
        df = self.get_data()
        df.index = df.ds
        model = ARIMA(df.y, order = (1, 0, 7))
        model = model.fit()
        forecast = pd.DataFrame()
        predictions = model.forecast(period)
        forecast["yhat"] = predictions
        forecast["ds"] = pd.date_range(start = df.index.max() + timedelta(days = 1), end = df.index.max() + timedelta(days = period))
        
        return self.viz(df, forecast)



# ts = TS(symbol = "FPT")
# data = ts.arima()
# print(data.index.min(), data.index.max())


st.title("Vietnam Trading by Time Series")
model = st.selectbox("model", ["Prophet", "ARIMA"])
sb = st.text_input('Symbol', 'FPT')
period = st.slider('Period', 1, 365, 28)
# fig = prophet_ts(symbol=sb, periods = periods)
ts = TS(symbol = sb)
fig = None
with st.spinner('Wait for it...'):
    if model == "Prophet":
        fig = ts.prophet(period = period)
    elif model == "ARIMA":
        fig = ts.arima(period = period)
st.success('Done!')
st.plotly_chart(fig, use_container_width=True)