part3.py

import os
import json
import time
import pandas as pd
from datetime import date
import matplotlib.pyplot as plt
from prophet import Prophet
from statsmodels.tsa.holtwinters import SimpleExpSmoothing, Holt

from part1 import fetch_ts, convert_to_ts

def get_forecast(df, month, country="ALL"): 

    df_ts = df.copy() if country=="ALL" else df[df['country']==country].copy()
    if len(df_ts) <= 180: 
        return None
    
    actual = df_ts[df_ts['inv_month'] == month.isoformat()]['value'].sum()

    df_ts = df_ts[['inv_month', 'inv_date', 'value']].groupby(['inv_month', 'inv_date']).sum().reset_index()
    
    df_train = df_ts[df_ts['inv_month'] < month.isoformat()]
    df_train['inv_date'] = pd.to_datetime(df_train['inv_date'])
    df_train.rename(columns={'inv_date':'ds', 'value':'y'}, inplace=True)

    m = Prophet(yearly_seasonality=20)
    m.fit(df_train)
    future = m.make_future_dataframe(periods=60, freq='D')
    df_forecast = m.predict(future)
    df_forecast['inv_month'] = df_forecast['ds'].apply(lambda v: date(v.year, v.month, 1).isoformat())
    forecast = df_forecast[df_forecast['inv_month'] == month.isoformat()]['yhat'].sum()

    exp_model = SimpleExpSmoothing(df_ts[['inv_month', 'inv_date', 'value']].set_index(['inv_month', 'inv_date'])).fit(smoothing_level=0.2, optimized=False)
    exp_forecast = forecast_df = sum(exp_model.forecast(30))
    
    return actual, forecast, exp_forecast

def test_forecast(df_ts, df_hist):

    df_ts = pd.concat([df_hist, df_ts])
    df_ts.sort_values('inv_date', inplace=True)
    months = [date(2019, 8, 1), date(2019, 9, 1), date(2019, 10, 1), date(2019, 11, 1), date(2019, 12, 1)]

    forecasts = []
    for m in months:
        actual, forecast, exp_forecast = get_forecast(df_ts, m)
        forecasts.append((m, actual, forecast, exp_forecast))

    print(forecasts)
    pd.DataFrame(forecasts, columns=["Month", "Actual", "Forecast", "EXP_Forecast"]).to_csv('prophet_forecast.log')

    cumu_actual = 0
    cumu_abs_error = 0
    cumu_exp_abs_error = 0
    for (m, actual, forecast, exp_forecast) in forecasts: 
        print(f"Month {m.strftime('%Y-%m')}: Actual {actual:.2}, Forecast {forecast:.2}, ERROR: {abs(forecast-actual)/actual:.1%}, EXP_ERROR: {abs(exp_forecast-actual)/actual:.1%}")
        if m != date(2019, 12, 1):
            cumu_actual += actual
            cumu_abs_error += abs(forecast-actual)
            cumu_exp_abs_error += abs(exp_forecast-actual)

    print(f"MAE: {cumu_abs_error/cumu_actual:.1%} vs. EXP MAE: {cumu_exp_abs_error/cumu_actual:.1%}")

if __name__ == "__main__":
    clean = False
    data_dir = 'cs-production'
    df_all = fetch_ts(data_dir, clean=clean)
    df_ts = convert_to_ts(data_dir, df_all, clean=clean)

    df_hist = pd.read_csv(r'cs-train\ts_data.csv')
    test_forecast(df_ts, df_hist)