stock_exchange.py

# Start evaluation traders:
# Compare their performance over the testing period (2012-2017) against a buy-and-hold trader.
from evaluating.evaluator_utils import initialize_portfolios
from utils import read_stock_market_data
from evaluating.portfolio_evaluator import PortfolioEvaluator
from model.CompanyEnum import CompanyEnum
from dependency_injection_containers import Traders
import datetime
from definitions import PERIOD_1, PERIOD_2

if __name__ == "__main__":
    # Load stock market data for training and testing period
    stock_market_data = read_stock_market_data([CompanyEnum.COMPANY_A, CompanyEnum.COMPANY_B], [PERIOD_1, PERIOD_2])

    # Define portfolio-name/trader mappings
    portfolio_name_trader_mappings = [
        # Benchmark trader
        ('Buy-and-hold Trader', Traders.BuyAndHoldTrader(), 'chartreuse'),

        # Simple traders
        ('Simple Trader (perfect prediction)', Traders.SimpleTrader_with_perfect_prediction(), 'pink'),
        ('Simple Trader (NN binary perfect prediction)', Traders.SimpleTrader_with_nn_binary_perfect_prediction(),
         'yellow'),
        ('Simple Trader (NN binary prediction)', Traders.SimpleTrader_with_nn_binary_prediction(), 'brown'),

        # Deep Q-Learning traders
        ('DQL Trader (perfect prediction)', Traders.DqlTrader_with_perfect_prediction(), 'gray'),
        ('DQL Trader (NN binary perfect prediction)', Traders.DqlTrader_with_nn_binary_perfect_prediction(), 'magenta'),
        ('DQL Trader (NN binary prediction)', Traders.DqlTrader_with_nn_binary_prediction(), 'burlywood'),

        # Code-Camp Task 0 traders
        (
        'Team Blue Simple Trader (perfect prediction)', Traders.TeamBlueSimpleTrader_with_perfect_prediction(), 'blue'),
        ('Team Green Simple Trader (perfect prediction)', Traders.TeamGreenSimpleTrader_with_perfect_prediction(),
         'green'),
        ('Team Black Simple Trader (perfect prediction)', Traders.TeamBlackSimpleTrader_with_perfect_prediction(),
         'black'),
        ('Team Red Simple Trader (perfect prediction)', Traders.TeamRedSimpleTrader_with_perfect_prediction(), 'red'),

        # Code-Camp Task 1 traders
        ('Simple Trader (Team Blue prediction)', Traders.SimpleTrader_with_team_blue_prediction(), 'blue'),
        ('Simple Trader (Team Green prediction)', Traders.SimpleTrader_with_team_green_prediction(), 'green'),
        ('Simple Trader (Team Black prediction)', Traders.SimpleTrader_with_team_black_prediction(), 'black'),
        ('Simple Trader (Team Red prediction)', Traders.SimpleTrader_with_team_red_prediction(), 'red'),

        # Code-Camp Task 2 traders    
        ('Team Blue DQL Trader (perfect prediction)', Traders.TeamBlueDqlTrader_with_perfect_prediction(), 'blue'),
        ('Team Green DQL Trader (perfect prediction)', Traders.TeamGreenDqlTrader_with_perfect_prediction(), 'green'),
        ('Team Black DQL Trader (perfect prediction)', Traders.TeamBlackDqlTrader_with_perfect_prediction(), 'black'),
        ('Team Red DQL Trader (perfect prediction)', Traders.TeamRedDqlTrader_with_perfect_prediction(), 'red')

    ]

    # Define portfolios for the traders and create a portfolio/trader mapping
    portfolio_trader_mappings = initialize_portfolios(10000.0, portfolio_name_trader_mappings)

    # Evaluate their performance over the testing period
    evaluator = PortfolioEvaluator([], True)
    evaluator.inspect_over_time_with_mapping(stock_market_data, portfolio_trader_mappings,
                                             date_offset=datetime.date(2012, 1, 3))