stable/freqtrade/tests/test_backtesting.py

# pragma pylint: disable=missing-docstring

import json
import logging
import os
from typing import Tuple, Dict

import arrow
import pytest
from arrow import Arrow
from pandas import DataFrame
from tabulate import tabulate

from freqtrade import exchange
from freqtrade.analyze import parse_ticker_dataframe, populate_indicators, \
    populate_buy_trend, populate_sell_trend
from freqtrade.exchange import Bittrex
from freqtrade.main import min_roi_reached
from freqtrade.persistence import Trade
from freqtrade.tests import load_backtesting_data

logger = logging.getLogger(__name__)


def format_results(results: DataFrame):
    return 'Made {} buys. Average profit {:.2f}%. ' \
           'Total profit was {:.3f}. Average duration {:.1f} mins.'.format(
               len(results.index),
               results.profit.mean() * 100.0,
               results.profit.sum(),
               results.duration.mean() * 5,
           )


def print_pair_results(pair: str, results: DataFrame):
    print('For currency {}:'.format(pair))
    print(format_results(results[results.currency == pair]))


def preprocess(backdata) -> Dict[str, DataFrame]:
    processed = {}
    for pair, pair_data in backdata.items():
        processed[pair] = populate_indicators(parse_ticker_dataframe(pair_data))
    return processed


def get_timeframe(data: Dict[str, Dict]) -> Tuple[Arrow, Arrow]:
    """
    Get the maximum timeframe for the given backtest data
    :param data: dictionary with backtesting data
    :return: tuple containing min_date, max_date
    """
    min_date, max_date = None, None
    for values in data.values():
        values = sorted(values, key=lambda d: arrow.get(d['T']))
        if not min_date or values[0]['T'] < min_date:
            min_date = values[0]['T']
        if not max_date or values[-1]['T'] > max_date:
            max_date = values[-1]['T']
    return arrow.get(min_date), arrow.get(max_date)


def generate_text_table(data: Dict[str, Dict], results: DataFrame, stake_currency) -> str:
    """
    Generates and returns a text table for the given backtest data and the results dataframe
    :return: pretty printed table with tabulate as str
    """
    tabular_data = []
    headers = ['pair', 'buy count', 'avg profit', 'total profit', 'avg duration']
    for pair in data:
        result = results[results.currency == pair]
        tabular_data.append([
            pair,
            len(result.index),
            '{:.2f}%'.format(result.profit.mean() * 100.0),
            '{:.08f} {}'.format(result.profit.sum(), stake_currency),
            '{:.2f}'.format(result.duration.mean() * 5),
        ])

    # Append Total
    tabular_data.append([
        'TOTAL',
        len(results.index),
        '{:.2f}%'.format(results.profit.mean() * 100.0),
        '{:.08f} {}'.format(results.profit.sum(), stake_currency),
        '{:.2f}'.format(results.duration.mean() * 5),
    ])
    return tabulate(tabular_data, headers=headers)


def backtest(backtest_conf, processed, mocker):
    trades = []
    exchange._API = Bittrex({'key': '', 'secret': ''})
    mocker.patch.dict('freqtrade.main._CONF', backtest_conf)
    for pair, pair_data in processed.items():
        pair_data['buy'] = 0
        pair_data['sell'] = 0
        ticker = populate_sell_trend(populate_buy_trend(pair_data))
        # for each buy point
        for row in ticker[ticker.buy == 1].itertuples(index=True):
            trade = Trade(
                open_rate=row.close,
                open_date=row.date,
                amount=backtest_conf['stake_amount'],
                fee=exchange.get_fee() * 2
            )
            # calculate win/lose forwards from buy point
            for row2 in ticker[row.Index:].itertuples(index=True):
                if min_roi_reached(trade, row2.close, row2.date) or row2.sell == 1:
                    current_profit = trade.calc_profit(row2.close)

                    trades.append((pair, current_profit, row2.Index - row.Index))
                    break
    labels = ['currency', 'profit', 'duration']
    return DataFrame.from_records(trades, columns=labels)


@pytest.mark.skipif(not os.environ.get('BACKTEST'), reason="BACKTEST not set")
def test_backtest(backtest_conf, mocker):
    print('')

    config = None
    conf_path = os.environ.get('BACKTEST_CONFIG')
    if conf_path:
        print('Using config: {} ...'.format(conf_path))
        with open(conf_path, 'r') as conf_file:
            config = json.load(conf_file)

    ticker_interval = int(os.environ.get('BACKTEST_TICKER_INTERVAL') or 5)
    print('Using ticker_interval: {} ...'.format(ticker_interval))

    data = {}
    if os.environ.get('BACKTEST_LIVE'):
        print('Downloading data for all pairs in whitelist ...')
        exchange._API = Bittrex({'key': '', 'secret': ''})
        for pair in config['exchange']['pair_whitelist']:
            data[pair] = exchange.get_ticker_history(pair, ticker_interval)
    else:
        print('Using local backtesting data (ignoring whitelist in given config)...')
        data = load_backtesting_data(ticker_interval)

    config = config or backtest_conf

    print('Using stake_currency: {} ...\nUsing stake_amount: {} ...'.format(
        config['stake_currency'], config['stake_amount']
    ))

    min_date, max_date = get_timeframe(data)
    print('Measuring data from {} up to {} ...'.format(
        min_date.isoformat(), max_date.isoformat()
    ))

    results = backtest(config, preprocess(data), mocker)
    print('====================== BACKTESTING REPORT ======================================\n\n'
          'NOTE: This Report doesn\'t respect the limits of max_open_trades, \n'
          '      so the projected values should be taken with a grain of salt.\n')
    print(generate_text_table(data, results, config['stake_currency']))