stable/freqtrade/optimize/hyperopt_interface.py

196 lines
7.9 KiB
Python
Raw Normal View History

"""
IHyperOpt interface
2019-11-13 08:38:06 +00:00
This module defines the interface to apply for hyperopt
"""
2019-08-14 10:25:49 +00:00
import logging
import math
from abc import ABC
2021-09-16 04:28:10 +00:00
from typing import Dict, List
from skopt.space import Categorical, Dimension, Integer
2019-08-14 10:25:49 +00:00
from freqtrade.exchange import timeframe_to_minutes
from freqtrade.misc import round_dict
from freqtrade.optimize.space import SKDecimal
from freqtrade.strategy import IStrategy
2020-09-28 17:39:41 +00:00
2021-03-27 09:40:48 +00:00
2019-08-14 10:25:49 +00:00
logger = logging.getLogger(__name__)
2019-09-03 16:54:28 +00:00
class IHyperOpt(ABC):
"""
2019-11-13 08:38:06 +00:00
Interface for freqtrade hyperopt
Defines the mandatory structure must follow any custom hyperopt
2019-08-14 10:25:49 +00:00
Class attributes you can use:
2021-04-03 14:54:47 +00:00
timeframe -> int: value of the timeframe to use for the strategy
"""
2020-06-15 04:35:31 +00:00
ticker_interval: str # DEPRECATED
timeframe: str
strategy: IStrategy
def __init__(self, config: dict) -> None:
self.config = config
# Assign ticker_interval to be used in hyperopt
2020-06-15 04:35:31 +00:00
IHyperOpt.ticker_interval = str(config['timeframe']) # DEPRECATED
IHyperOpt.timeframe = str(config['timeframe'])
def generate_roi_table(self, params: Dict) -> Dict[int, float]:
"""
Create a ROI table.
Generates the ROI table that will be used by Hyperopt.
You may override it in your custom Hyperopt class.
"""
roi_table = {}
roi_table[0] = params['roi_p1'] + params['roi_p2'] + params['roi_p3']
roi_table[params['roi_t3']] = params['roi_p1'] + params['roi_p2']
roi_table[params['roi_t3'] + params['roi_t2']] = params['roi_p1']
roi_table[params['roi_t3'] + params['roi_t2'] + params['roi_t1']] = 0
return roi_table
def roi_space(self) -> List[Dimension]:
"""
2019-08-14 10:25:49 +00:00
Create a ROI space.
Defines values to search for each ROI steps.
This method implements adaptive roi hyperspace with varied
ranges for parameters which automatically adapts to the
2021-04-03 14:54:47 +00:00
timeframe used.
2019-08-14 10:25:49 +00:00
It's used by Freqtrade by default, if no custom roi_space method is defined.
"""
# Default scaling coefficients for the roi hyperspace. Can be changed
# to adjust resulting ranges of the ROI tables.
# Increase if you need wider ranges in the roi hyperspace, decrease if shorter
# ranges are needed.
roi_t_alpha = 1.0
roi_p_alpha = 1.0
2021-04-03 15:10:39 +00:00
timeframe_min = timeframe_to_minutes(self.timeframe)
2019-08-14 10:25:49 +00:00
# We define here limits for the ROI space parameters automagically adapted to the
# timeframe used by the bot:
2019-08-14 10:25:49 +00:00
#
# * 'roi_t' (limits for the time intervals in the ROI tables) components
# are scaled linearly.
# * 'roi_p' (limits for the ROI value steps) components are scaled logarithmically.
#
# The scaling is designed so that it maps exactly to the legacy Freqtrade roi_space()
2021-04-03 14:54:47 +00:00
# method for the 5m timeframe.
2019-12-11 06:12:37 +00:00
roi_t_scale = timeframe_min / 5
roi_p_scale = math.log1p(timeframe_min) / math.log1p(5)
2019-08-14 10:25:49 +00:00
roi_limits = {
'roi_t1_min': int(10 * roi_t_scale * roi_t_alpha),
'roi_t1_max': int(120 * roi_t_scale * roi_t_alpha),
'roi_t2_min': int(10 * roi_t_scale * roi_t_alpha),
'roi_t2_max': int(60 * roi_t_scale * roi_t_alpha),
'roi_t3_min': int(10 * roi_t_scale * roi_t_alpha),
'roi_t3_max': int(40 * roi_t_scale * roi_t_alpha),
'roi_p1_min': 0.01 * roi_p_scale * roi_p_alpha,
'roi_p1_max': 0.04 * roi_p_scale * roi_p_alpha,
'roi_p2_min': 0.01 * roi_p_scale * roi_p_alpha,
'roi_p2_max': 0.07 * roi_p_scale * roi_p_alpha,
'roi_p3_min': 0.01 * roi_p_scale * roi_p_alpha,
'roi_p3_max': 0.20 * roi_p_scale * roi_p_alpha,
}
logger.debug(f"Using roi space limits: {roi_limits}")
2019-08-14 10:25:49 +00:00
p = {
'roi_t1': roi_limits['roi_t1_min'],
'roi_t2': roi_limits['roi_t2_min'],
'roi_t3': roi_limits['roi_t3_min'],
'roi_p1': roi_limits['roi_p1_min'],
'roi_p2': roi_limits['roi_p2_min'],
'roi_p3': roi_limits['roi_p3_min'],
}
2021-04-15 19:38:20 +00:00
logger.info(f"Min roi table: {round_dict(self.generate_roi_table(p), 3)}")
2019-08-14 10:25:49 +00:00
p = {
'roi_t1': roi_limits['roi_t1_max'],
'roi_t2': roi_limits['roi_t2_max'],
'roi_t3': roi_limits['roi_t3_max'],
'roi_p1': roi_limits['roi_p1_max'],
'roi_p2': roi_limits['roi_p2_max'],
'roi_p3': roi_limits['roi_p3_max'],
}
2021-04-15 19:38:20 +00:00
logger.info(f"Max roi table: {round_dict(self.generate_roi_table(p), 3)}")
return [
2019-08-14 10:25:49 +00:00
Integer(roi_limits['roi_t1_min'], roi_limits['roi_t1_max'], name='roi_t1'),
Integer(roi_limits['roi_t2_min'], roi_limits['roi_t2_max'], name='roi_t2'),
Integer(roi_limits['roi_t3_min'], roi_limits['roi_t3_max'], name='roi_t3'),
2021-04-15 19:38:20 +00:00
SKDecimal(roi_limits['roi_p1_min'], roi_limits['roi_p1_max'], decimals=3,
name='roi_p1'),
2021-04-15 19:38:20 +00:00
SKDecimal(roi_limits['roi_p2_min'], roi_limits['roi_p2_max'], decimals=3,
name='roi_p2'),
2021-04-15 19:38:20 +00:00
SKDecimal(roi_limits['roi_p3_min'], roi_limits['roi_p3_max'], decimals=3,
name='roi_p3'),
]
def stoploss_space(self) -> List[Dimension]:
"""
2019-08-14 10:25:49 +00:00
Create a stoploss space.
2019-08-14 10:25:49 +00:00
Defines range of stoploss values to search.
You may override it in your custom Hyperopt class.
"""
return [
SKDecimal(-0.35, -0.02, decimals=3, name='stoploss'),
]
2019-08-14 10:25:49 +00:00
def generate_trailing_params(self, params: Dict) -> Dict:
"""
Create dict with trailing stop parameters.
"""
return {
'trailing_stop': params['trailing_stop'],
'trailing_stop_positive': params['trailing_stop_positive'],
'trailing_stop_positive_offset': (params['trailing_stop_positive'] +
params['trailing_stop_positive_offset_p1']),
'trailing_only_offset_is_reached': params['trailing_only_offset_is_reached'],
}
def trailing_space(self) -> List[Dimension]:
2019-11-07 22:55:14 +00:00
"""
Create a trailing stoploss space.
You may override it in your custom Hyperopt class.
"""
return [
# It was decided to always set trailing_stop is to True if the 'trailing' hyperspace
# is used. Otherwise hyperopt will vary other parameters that won't have effect if
# trailing_stop is set False.
# This parameter is included into the hyperspace dimensions rather than assigning
# it explicitly in the code in order to have it printed in the results along with
# other 'trailing' hyperspace parameters.
Categorical([True], name='trailing_stop'),
SKDecimal(0.01, 0.35, decimals=3, name='trailing_stop_positive'),
# 'trailing_stop_positive_offset' should be greater than 'trailing_stop_positive',
# so this intermediate parameter is used as the value of the difference between
# them. The value of the 'trailing_stop_positive_offset' is constructed in the
# generate_trailing_params() method.
2020-02-07 23:49:06 +00:00
# This is similar to the hyperspace dimensions used for constructing the ROI tables.
SKDecimal(0.001, 0.1, decimals=3, name='trailing_stop_positive_offset_p1'),
2019-11-07 22:55:14 +00:00
Categorical([True, False], name='trailing_only_offset_is_reached'),
]
2019-08-14 10:25:49 +00:00
# This is needed for proper unpickling the class attribute ticker_interval
# which is set to the actual value by the resolver.
# Why do I still need such shamanic mantras in modern python?
def __getstate__(self):
state = self.__dict__.copy()
state['timeframe'] = self.timeframe
2019-08-14 10:25:49 +00:00
return state
def __setstate__(self, state):
self.__dict__.update(state)
IHyperOpt.ticker_interval = state['timeframe']
IHyperOpt.timeframe = state['timeframe']