market.py

import pandas as pd
import numpy as np


class Market:

    def __init__(self, file, number_of_assets=1):
        self.rawData = pd.read_csv(file)

        # Compute number of signals. Remove 1 to account for the date column (first column)
        self.number_of_signals = len(self.rawData.values[0, :]) - number_of_assets - 1
        self.duration = len(self.rawData.values[:, 1])

        self.signals = self.rawData.iloc[:, 2:].values
        self.signals = self.preprocess_data(self.signals)

        # temporarily force coded the cash asset
        # self.asset_prices = self.rawData.IntelPriceUSD.values
        asset_risk_free = np.ones([self.duration, 1])
        asset_others = self.rawData.iloc[:, 1:number_of_assets+1].values.reshape(self.duration, number_of_assets)
        self.asset_prices = np.append(asset_risk_free, asset_others, axis=1)
        # temporarily force add the cash asset in total number of asset
        self.number_of_assets = number_of_assets + 1

    def preprocess_data(self, signals):
        # input := np.array
        # thats where we should modify the signals, e.g. standardization
        signals -= np.mean(signals, 0)
        signals /= np.std(signals, 0)

        return signals