stable/scripts/plot_profit.py

231 lines
6.9 KiB
Python
Raw Normal View History

2018-01-12 09:55:49 +00:00
#!/usr/bin/env python3
"""
Script to display profits
Mandatory Cli parameters:
-p / --pair: pair to examine
Optional Cli parameters
-c / --config: specify configuration file
-s / --strategy: strategy to use
-d / --datadir: path to pair backtest data
--timerange: specify what timerange of data to use
--export-filename: Specify where the backtest export is located.
"""
2019-03-23 18:18:10 +00:00
import json
2018-03-25 19:37:14 +00:00
import logging
2018-01-12 09:55:49 +00:00
import sys
from argparse import Namespace
2018-12-15 12:54:35 +00:00
from pathlib import Path
from typing import List, Optional
2018-01-12 09:55:49 +00:00
2019-03-23 18:18:10 +00:00
import numpy as np
import plotly.graph_objs as go
from plotly import tools
from plotly.offline import plot
from freqtrade.arguments import Arguments
from freqtrade.configuration import Configuration
2018-12-15 12:54:35 +00:00
from freqtrade.data import history
from freqtrade.exchange import timeframe_to_seconds
from freqtrade.misc import common_datearray
from freqtrade.resolvers import StrategyResolver
from freqtrade.state import RunMode
2018-03-25 19:37:14 +00:00
logger = logging.getLogger(__name__)
2018-01-12 09:55:49 +00:00
2018-01-12 21:15:50 +00:00
# data:: [ pair, profit-%, enter, exit, time, duration]
# data:: ["ETH/BTC", 0.0023975, "1515598200", "1515602100", "2018-01-10 07:30:00+00:00", 65]
def make_profit_array(data: List, px: int, min_date: int,
2019-03-23 18:28:06 +00:00
interval: str,
filter_pairs: Optional[List] = None) -> np.ndarray:
pg = np.zeros(px)
filter_pairs = filter_pairs or []
2018-01-12 09:55:49 +00:00
# Go through the trades
# and make an total profit
# array
for trade in data:
pair = trade[0]
if filter_pairs and pair not in filter_pairs:
2018-01-12 09:55:49 +00:00
continue
profit = trade[1]
trade_sell_time = int(trade[3])
ix = define_index(min_date, trade_sell_time, interval)
2018-01-20 18:49:04 +00:00
if ix < px:
logger.debug('[%s]: Add profit %s on %s', pair, profit, trade[4])
2018-01-20 18:49:04 +00:00
pg[ix] += profit
2018-01-12 09:55:49 +00:00
# rewrite the pg array to go from
# total profits at each timeframe
# to accumulated profits
pa = 0
2018-01-23 05:20:17 +00:00
for x in range(0, len(pg)):
2018-01-12 09:55:49 +00:00
p = pg[x] # Get current total percent
pa += p # Add to the accumulated percent
pg[x] = pa # write back to save memory
return pg
def plot_profit(args: Namespace) -> None:
2018-01-12 09:55:49 +00:00
"""
Plots the total profit for all pairs.
Note, the profit calculation isn't realistic.
But should be somewhat proportional, and therefor useful
in helping out to find a good algorithm.
"""
# We need to use the same pairs, same ticker_interval
2018-01-12 09:55:49 +00:00
# and same timeperiod as used in backtesting
# to match the tickerdata against the profits-results
timerange = Arguments.parse_timerange(args.timerange)
2018-01-12 09:55:49 +00:00
config = Configuration(args, RunMode.OTHER).get_config()
# Init strategy
try:
strategy = StrategyResolver({'strategy': config.get('strategy')}).strategy
except AttributeError:
logger.critical(
'Impossible to load the strategy. Please check the file "user_data/strategies/%s.py"',
config.get('strategy')
)
2018-06-08 00:01:46 +00:00
exit(1)
# Load the profits results
try:
filename = args.exportfilename
with open(filename) as file:
data = json.load(file)
except FileNotFoundError:
logger.critical(
'File "backtest-result.json" not found. This script require backtesting '
'results to run.\nPlease run a backtesting with the parameter --export.')
2018-06-08 00:01:46 +00:00
exit(1)
# Take pairs from the cli otherwise switch to the pair in the config file
if args.pairs:
filter_pairs = args.pairs
filter_pairs = filter_pairs.split(',')
else:
filter_pairs = config['exchange']['pair_whitelist']
ticker_interval = strategy.ticker_interval
2018-01-12 09:55:49 +00:00
pairs = config['exchange']['pair_whitelist']
if filter_pairs:
pairs = list(set(pairs) & set(filter_pairs))
logger.info('Filter, keep pairs %s' % pairs)
2018-12-13 05:34:37 +00:00
tickers = history.load_data(
2019-03-23 18:28:06 +00:00
datadir=Path(str(config.get('datadir'))),
pairs=pairs,
ticker_interval=ticker_interval,
refresh_pairs=False,
timerange=timerange
)
dataframes = strategy.tickerdata_to_dataframe(tickers)
2018-01-12 09:55:49 +00:00
# NOTE: the dataframes are of unequal length,
# 'dates' is an merged date array of them all.
2018-01-12 09:55:49 +00:00
dates = common_datearray(dataframes)
min_date = int(min(dates).timestamp())
max_date = int(max(dates).timestamp())
num_iterations = define_index(min_date, max_date, ticker_interval) + 1
2018-01-12 09:55:49 +00:00
# Make an average close price of all the pairs that was involved.
# this could be useful to gauge the overall market trend
2018-01-12 09:55:49 +00:00
# We are essentially saying:
# array <- sum dataframes[*]['close'] / num_items dataframes
# FIX: there should be some onliner numpy/panda for this
avgclose = np.zeros(num_iterations)
2018-01-12 09:55:49 +00:00
num = 0
for pair, pair_data in dataframes.items():
close = pair_data['close']
maxprice = max(close) # Normalize price to [0,1]
logger.info('Pair %s has length %s' % (pair, len(close)))
for x in range(0, len(close)):
avgclose[x] += close[x] / maxprice
# avgclose += close
num += 1
2018-01-12 09:55:49 +00:00
avgclose /= num
# make an profits-growth array
pg = make_profit_array(data, num_iterations, min_date, ticker_interval, filter_pairs)
2018-01-12 09:55:49 +00:00
#
# Plot the pairs average close prices, and total profit growth
#
avgclose = go.Scattergl(
x=dates,
y=avgclose,
name='Avg close price',
)
profit = go.Scattergl(
x=dates,
y=pg,
name='Profit',
)
fig = tools.make_subplots(rows=3, cols=1, shared_xaxes=True, row_width=[1, 1, 1])
fig.append_trace(avgclose, 1, 1)
fig.append_trace(profit, 2, 1)
2018-01-12 09:55:49 +00:00
for pair in pairs:
pg = make_profit_array(data, num_iterations, min_date, ticker_interval, [pair])
pair_profit = go.Scattergl(
x=dates,
y=pg,
name=pair,
)
fig.append_trace(pair_profit, 3, 1)
2018-12-15 12:54:35 +00:00
plot(fig, filename=str(Path('user_data').joinpath('freqtrade-profit-plot.html')))
2018-01-12 09:55:49 +00:00
def define_index(min_date: int, max_date: int, ticker_interval: str) -> int:
"""
Return the index of a specific date
"""
interval_seconds = timeframe_to_seconds(ticker_interval)
return int((max_date - min_date) / interval_seconds)
2018-03-05 04:24:01 +00:00
def plot_parse_args(args: List[str]) -> Namespace:
"""
Parse args passed to the script
:param args: Cli arguments
:return: args: Array with all arguments
"""
arguments = Arguments(args, 'Graph profits')
2019-06-18 22:53:38 +00:00
arguments.common_options()
arguments.main_options()
arguments.common_optimize_options()
arguments.backtesting_options()
arguments.common_scripts_options()
return arguments.parse_args()
def main(sysargv: List[str]) -> None:
"""
This function will initiate the bot and start the trading loop.
:return: None
"""
logger.info('Starting Plot Dataframe')
plot_profit(
plot_parse_args(sysargv)
)
2018-01-12 09:55:49 +00:00
if __name__ == '__main__':
main(sys.argv[1:])