adelorenzo/stable
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge remote-tracking branch 'origin/develop' into gc_improvements

Browse Source
This commit is contained in:
robcaulk
2022-12-01 14:32:19 +01:00
parent b929e0bb2b 95651fcd5a
commit aceee67e2b
87 changed files with 4730 additions and 2727 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
freqtrade/__init__.py
View File

@@ -1,5 +1,5 @@
""" Freqtrade bot """
__version__ = '2022.11.dev'
__version__ = '2022.12.dev'
if 'dev' in __version__:
try:

8
freqtrade/commands/analyze_commands.py
View File

@@ -60,10 +60,4 @@ def start_analysis_entries_exits(args: Dict[str, Any]) -> None:
logger.info('Starting freqtrade in analysis mode')
process_entry_exit_reasons(config['exportfilename'],
config['exchange']['pair_whitelist'],
config['analysis_groups'],
config['enter_reason_list'],
config['exit_reason_list'],
config['indicator_list']
)
process_entry_exit_reasons(config)

2
freqtrade/commands/arguments.py
View File

@@ -106,7 +106,7 @@ ARGS_HYPEROPT_SHOW = ["hyperopt_list_best", "hyperopt_list_profitable", "hyperop
"disableparamexport", "backtest_breakdown"]
ARGS_ANALYZE_ENTRIES_EXITS = ["exportfilename", "analysis_groups", "enter_reason_list",
"exit_reason_list", "indicator_list"]
"exit_reason_list", "indicator_list", "timerange"]
NO_CONF_REQURIED = ["convert-data", "convert-trade-data", "download-data", "list-timeframes",
"list-markets", "list-pairs", "list-strategies", "list-freqaimodels",

3
freqtrade/configuration/configuration.py
View File

@@ -462,6 +462,9 @@ class Configuration:
self._args_to_config(config, argname='indicator_list',
logstring='Analysis indicator list: {}')
self._args_to_config(config, argname='timerange',
logstring='Filter trades by timerange: {}')
def _process_runmode(self, config: Config) -> None:
self._args_to_config(config, argname='dry_run',

49
freqtrade/configuration/timerange.py
View File

@@ -3,11 +3,12 @@ This module contains the argument manager class
"""
import logging
import re
from datetime import datetime
from datetime import datetime, timezone
from typing import Optional
import arrow
from freqtrade.constants import DATETIME_PRINT_FORMAT
from freqtrade.exceptions import OperationalException
@@ -29,6 +30,52 @@ class TimeRange:
self.startts: int = startts
self.stopts: int = stopts
@property
def startdt(self) -> Optional[datetime]:
if self.startts:
return datetime.fromtimestamp(self.startts, tz=timezone.utc)
return None
@property
def stopdt(self) -> Optional[datetime]:
if self.stopts:
return datetime.fromtimestamp(self.stopts, tz=timezone.utc)
return None
@property
def timerange_str(self) -> str:
"""
Returns a string representation of the timerange as used by parse_timerange.
Follows the format yyyymmdd-yyyymmdd - leaving out the parts that are not set.
"""
start = ''
stop = ''
if startdt := self.startdt:
start = startdt.strftime('%Y%m%d')
if stopdt := self.stopdt:
stop = stopdt.strftime('%Y%m%d')
return f"{start}-{stop}"
@property
def start_fmt(self) -> str:
"""
Returns a string representation of the start date
"""
val = 'unbounded'
if (startdt := self.startdt) is not None:
val = startdt.strftime(DATETIME_PRINT_FORMAT)
return val
@property
def stop_fmt(self) -> str:
"""
Returns a string representation of the stop date
"""
val = 'unbounded'
if (stopdt := self.stopdt) is not None:
val = stopdt.strftime(DATETIME_PRINT_FORMAT)
return val
def __eq__(self, other):
"""Override the default Equals behavior"""
return (self.starttype == other.starttype and self.stoptype == other.stoptype

21
freqtrade/constants.py
View File

@@ -512,6 +512,7 @@ CONF_SCHEMA = {
'minimum': 0,
'maximum': 65535
},
'secure': {'type': 'boolean', 'default': False},
'ws_token': {'type': 'string'},
},
'required': ['name', 'host', 'ws_token']
@@ -577,9 +578,27 @@ CONF_SCHEMA = {
},
},
"model_training_parameters": {
"type": "object"
},
"rl_config": {
"type": "object",
"properties": {
"n_estimators": {"type": "integer", "default": 1000}
"train_cycles": {"type": "integer"},
"max_trade_duration_candles": {"type": "integer"},
"add_state_info": {"type": "boolean", "default": False},
"max_training_drawdown_pct": {"type": "number", "default": 0.02},
"cpu_count": {"type": "integer", "default": 1},
"model_type": {"type": "string", "default": "PPO"},
"policy_type": {"type": "string", "default": "MlpPolicy"},
"net_arch": {"type": "array", "default": [128, 128]},
"randomize_startinng_position": {"type": "boolean", "default": False},
"model_reward_parameters": {
"type": "object",
"properties": {
"rr": {"type": "number", "default": 1},
"profit_aim": {"type": "number", "default": 0.025}
}
}
},
},
},

7
freqtrade/data/converter.py
View File

@@ -3,7 +3,6 @@ Functions to convert data from one format to another
"""
import itertools
import logging
from datetime import datetime, timezone
from operator import itemgetter
from typing import Dict, List
@@ -138,11 +137,9 @@ def trim_dataframe(df: DataFrame, timerange, df_date_col: str = 'date',
df = df.iloc[startup_candles:, :]
else:
if timerange.starttype == 'date':
start = datetime.fromtimestamp(timerange.startts, tz=timezone.utc)
df = df.loc[df[df_date_col] >= start, :]
df = df.loc[df[df_date_col] >= timerange.startdt, :]
if timerange.stoptype == 'date':
stop = datetime.fromtimestamp(timerange.stopts, tz=timezone.utc)
df = df.loc[df[df_date_col] <= stop, :]
df = df.loc[df[df_date_col] <= timerange.stopdt, :]
return df

100
freqtrade/data/entryexitanalysis.py
View File

@@ -1,11 +1,12 @@
import logging
from pathlib import Path
from typing import List, Optional
import joblib
import pandas as pd
from tabulate import tabulate
from freqtrade.configuration import TimeRange
from freqtrade.constants import Config
from freqtrade.data.btanalysis import (get_latest_backtest_filename, load_backtest_data,
load_backtest_stats)
from freqtrade.exceptions import OperationalException
@@ -152,37 +153,55 @@ def _do_group_table_output(bigdf, glist):
logger.warning("Invalid group mask specified.")
def _print_results(analysed_trades, stratname, analysis_groups,
enter_reason_list, exit_reason_list,
indicator_list, columns=None):
if columns is None:
columns = ['pair', 'open_date', 'close_date', 'profit_abs', 'enter_reason', 'exit_reason']
def _select_rows_within_dates(df, timerange=None, df_date_col: str = 'date'):
if timerange:
if timerange.starttype == 'date':
df = df.loc[(df[df_date_col] >= timerange.startdt)]
if timerange.stoptype == 'date':
df = df.loc[(df[df_date_col] < timerange.stopdt)]
return df
bigdf = pd.DataFrame()
def _select_rows_by_tags(df, enter_reason_list, exit_reason_list):
if enter_reason_list and "all" not in enter_reason_list:
df = df.loc[(df['enter_reason'].isin(enter_reason_list))]
if exit_reason_list and "all" not in exit_reason_list:
df = df.loc[(df['exit_reason'].isin(exit_reason_list))]
return df
def prepare_results(analysed_trades, stratname,
enter_reason_list, exit_reason_list,
timerange=None):
res_df = pd.DataFrame()
for pair, trades in analysed_trades[stratname].items():
bigdf = pd.concat([bigdf, trades], ignore_index=True)
res_df = pd.concat([res_df, trades], ignore_index=True)
if bigdf.shape[0] > 0 and ('enter_reason' in bigdf.columns):
res_df = _select_rows_within_dates(res_df, timerange)
if res_df is not None and res_df.shape[0] > 0 and ('enter_reason' in res_df.columns):
res_df = _select_rows_by_tags(res_df, enter_reason_list, exit_reason_list)
return res_df
def print_results(res_df, analysis_groups, indicator_list):
if res_df.shape[0] > 0:
if analysis_groups:
_do_group_table_output(bigdf, analysis_groups)
if enter_reason_list and "all" not in enter_reason_list:
bigdf = bigdf.loc[(bigdf['enter_reason'].isin(enter_reason_list))]
if exit_reason_list and "all" not in exit_reason_list:
bigdf = bigdf.loc[(bigdf['exit_reason'].isin(exit_reason_list))]
_do_group_table_output(res_df, analysis_groups)
if "all" in indicator_list:
print(bigdf)
print(res_df)
elif indicator_list is not None:
available_inds = []
for ind in indicator_list:
if ind in bigdf:
if ind in res_df:
available_inds.append(ind)
ilist = ["pair", "enter_reason", "exit_reason"] + available_inds
_print_table(bigdf[ilist], sortcols=['exit_reason'], show_index=False)
_print_table(res_df[ilist], sortcols=['exit_reason'], show_index=False)
else:
print("\\_ No trades to show")
print("\\No trades to show")
def _print_table(df, sortcols=None, show_index=False):
@@ -201,27 +220,34 @@ def _print_table(df, sortcols=None, show_index=False):
)
def process_entry_exit_reasons(backtest_dir: Path,
pairlist: List[str],
analysis_groups: Optional[List[str]] = ["0", "1", "2"],
enter_reason_list: Optional[List[str]] = ["all"],
exit_reason_list: Optional[List[str]] = ["all"],
indicator_list: Optional[List[str]] = []):
def process_entry_exit_reasons(config: Config):
try:
backtest_stats = load_backtest_stats(backtest_dir)
analysis_groups = config.get('analysis_groups', [])
enter_reason_list = config.get('enter_reason_list', ["all"])
exit_reason_list = config.get('exit_reason_list', ["all"])
indicator_list = config.get('indicator_list', [])
timerange = TimeRange.parse_timerange(None if config.get(
'timerange') is None else str(config.get('timerange')))
backtest_stats = load_backtest_stats(config['exportfilename'])
for strategy_name, results in backtest_stats['strategy'].items():
trades = load_backtest_data(backtest_dir, strategy_name)
trades = load_backtest_data(config['exportfilename'], strategy_name)
if not trades.empty:
signal_candles = _load_signal_candles(backtest_dir)
analysed_trades_dict = _process_candles_and_indicators(pairlist, strategy_name,
trades, signal_candles)
_print_results(analysed_trades_dict,
strategy_name,
analysis_groups,
enter_reason_list,
exit_reason_list,
indicator_list)
signal_candles = _load_signal_candles(config['exportfilename'])
analysed_trades_dict = _process_candles_and_indicators(
config['exchange']['pair_whitelist'], strategy_name,
trades, signal_candles)
res_df = prepare_results(analysed_trades_dict, strategy_name,
enter_reason_list, exit_reason_list,
timerange=timerange)
print_results(res_df,
analysis_groups,
indicator_list)
except ValueError as e:
raise OperationalException(e) from e

6
freqtrade/data/history/history_utils.py
View File

@@ -1,6 +1,6 @@
import logging
import operator
from datetime import datetime, timezone
from datetime import datetime
from pathlib import Path
from typing import Dict, List, Optional, Tuple
@@ -160,9 +160,9 @@ def _load_cached_data_for_updating(
end = None
if timerange:
if timerange.starttype == 'date':
start = datetime.fromtimestamp(timerange.startts, tz=timezone.utc)
start = timerange.startdt
if timerange.stoptype == 'date':
end = datetime.fromtimestamp(timerange.stopts, tz=timezone.utc)
end = timerange.stopdt
# Intentionally don't pass timerange in - since we need to load the full dataset.
data = data_handler.ohlcv_load(pair, timeframe=timeframe,

6
freqtrade/data/history/idatahandler.py
View File

@@ -366,13 +366,11 @@ class IDataHandler(ABC):
"""
if timerange.starttype == 'date':
start = datetime.fromtimestamp(timerange.startts, tz=timezone.utc)
if pairdata.iloc[0]['date'] > start:
if pairdata.iloc[0]['date'] > timerange.startdt:
logger.warning(f"{pair}, {candle_type}, {timeframe}, "
f"data starts at {pairdata.iloc[0]['date']:%Y-%m-%d %H:%M:%S}")
if timerange.stoptype == 'date':
stop = datetime.fromtimestamp(timerange.stopts, tz=timezone.utc)
if pairdata.iloc[-1]['date'] < stop:
if pairdata.iloc[-1]['date'] < timerange.stopdt:
logger.warning(f"{pair}, {candle_type}, {timeframe}, "
f"data ends at {pairdata.iloc[-1]['date']:%Y-%m-%d %H:%M:%S}")

2582
freqtrade/exchange/binance_leverage_tiers.json
View File

File diff suppressed because it is too large Load Diff

2
freqtrade/exchange/bybit.py
View File

@@ -20,7 +20,7 @@ class Bybit(Exchange):
"""
_ft_has: Dict = {
"ohlcv_candle_limit": 200,
"ohlcv_candle_limit": 1000,
"ccxt_futures_name": "linear",
"ohlcv_has_history": False,
}

16
freqtrade/exchange/kraken.py
View File

@@ -218,3 +218,19 @@ class Kraken(Exchange):
fees = sum(df['open_fund'] * df['open_mark'] * amount * time_in_ratio)
return fees if is_short else -fees
def _trades_contracts_to_amount(self, trades: List) -> List:
"""
Fix "last" id issue for kraken data downloads
This whole override can probably be removed once the following
issue is closed in ccxt: https://github.com/ccxt/ccxt/issues/15827
"""
super()._trades_contracts_to_amount(trades)
if (
len(trades) > 0
and isinstance(trades[-1].get('info'), list)
and len(trades[-1].get('info', [])) > 7
):
trades[-1]['id'] = trades[-1].get('info', [])[-1]
return trades

135
freqtrade/freqai/RL/Base4ActionRLEnv.py Normal file
View File

@@ -0,0 +1,135 @@
import logging
from enum import Enum
from gym import spaces
from freqtrade.freqai.RL.BaseEnvironment import BaseEnvironment, Positions
logger = logging.getLogger(__name__)
class Actions(Enum):
Neutral = 0
Exit = 1
Long_enter = 2
Short_enter = 3
class Base4ActionRLEnv(BaseEnvironment):
"""
Base class for a 4 action environment
"""
def set_action_space(self):
self.action_space = spaces.Discrete(len(Actions))
def step(self, action: int):
"""
Logic for a single step (incrementing one candle in time)
by the agent
:param: action: int = the action type that the agent plans
to take for the current step.
:returns:
observation = current state of environment
step_reward = the reward from `calculate_reward()`
_done = if the agent "died" or if the candles finished
info = dict passed back to openai gym lib
"""
self._done = False
self._current_tick += 1
if self._current_tick == self._end_tick:
self._done = True
self._update_unrealized_total_profit()
step_reward = self.calculate_reward(action)
self.total_reward += step_reward
trade_type = None
if self.is_tradesignal(action):
"""
Action: Neutral, position: Long -> Close Long
Action: Neutral, position: Short -> Close Short
Action: Long, position: Neutral -> Open Long
Action: Long, position: Short -> Close Short and Open Long
Action: Short, position: Neutral -> Open Short
Action: Short, position: Long -> Close Long and Open Short
"""
if action == Actions.Neutral.value:
self._position = Positions.Neutral
trade_type = "neutral"
self._last_trade_tick = None
elif action == Actions.Long_enter.value:
self._position = Positions.Long
trade_type = "long"
self._last_trade_tick = self._current_tick
elif action == Actions.Short_enter.value:
self._position = Positions.Short
trade_type = "short"
self._last_trade_tick = self._current_tick
elif action == Actions.Exit.value:
self._update_total_profit()
self._position = Positions.Neutral
trade_type = "neutral"
self._last_trade_tick = None
else:
print("case not defined")
if trade_type is not None:
self.trade_history.append(
{'price': self.current_price(), 'index': self._current_tick,
'type': trade_type})
if self._total_profit < 1 - self.rl_config.get('max_training_drawdown_pct', 0.8):
self._done = True
self._position_history.append(self._position)
info = dict(
tick=self._current_tick,
total_reward=self.total_reward,
total_profit=self._total_profit,
position=self._position.value
)
observation = self._get_observation()
self._update_history(info)
return observation, step_reward, self._done, info
def is_tradesignal(self, action: int) -> bool:
"""
Determine if the signal is a trade signal
e.g.: agent wants a Actions.Long_exit while it is in a Positions.short
"""
return not ((action == Actions.Neutral.value and self._position == Positions.Neutral) or
(action == Actions.Neutral.value and self._position == Positions.Short) or
(action == Actions.Neutral.value and self._position == Positions.Long) or
(action == Actions.Short_enter.value and self._position == Positions.Short) or
(action == Actions.Short_enter.value and self._position == Positions.Long) or
(action == Actions.Exit.value and self._position == Positions.Neutral) or
(action == Actions.Long_enter.value and self._position == Positions.Long) or
(action == Actions.Long_enter.value and self._position == Positions.Short))
def _is_valid(self, action: int) -> bool:
"""
Determine if the signal is valid.
e.g.: agent wants a Actions.Long_exit while it is in a Positions.short
"""
# Agent should only try to exit if it is in position
if action == Actions.Exit.value:
if self._position not in (Positions.Short, Positions.Long):
return False
# Agent should only try to enter if it is not in position
if action in (Actions.Short_enter.value, Actions.Long_enter.value):
if self._position != Positions.Neutral:
return False
return True

145
freqtrade/freqai/RL/Base5ActionRLEnv.py Normal file
View File

@@ -0,0 +1,145 @@
import logging
from enum import Enum
from gym import spaces
from freqtrade.freqai.RL.BaseEnvironment import BaseEnvironment, Positions
logger = logging.getLogger(__name__)
class Actions(Enum):
Neutral = 0
Long_enter = 1
Long_exit = 2
Short_enter = 3
Short_exit = 4
class Base5ActionRLEnv(BaseEnvironment):
"""
Base class for a 5 action environment
"""
def set_action_space(self):
self.action_space = spaces.Discrete(len(Actions))
def step(self, action: int):
"""
Logic for a single step (incrementing one candle in time)
by the agent
:param: action: int = the action type that the agent plans
to take for the current step.
:returns:
observation = current state of environment
step_reward = the reward from `calculate_reward()`
_done = if the agent "died" or if the candles finished
info = dict passed back to openai gym lib
"""
self._done = False
self._current_tick += 1
if self._current_tick == self._end_tick:
self._done = True
self._update_unrealized_total_profit()
step_reward = self.calculate_reward(action)
self.total_reward += step_reward
trade_type = None
if self.is_tradesignal(action):
"""
Action: Neutral, position: Long -> Close Long
Action: Neutral, position: Short -> Close Short
Action: Long, position: Neutral -> Open Long
Action: Long, position: Short -> Close Short and Open Long
Action: Short, position: Neutral -> Open Short
Action: Short, position: Long -> Close Long and Open Short
"""
if action == Actions.Neutral.value:
self._position = Positions.Neutral
trade_type = "neutral"
self._last_trade_tick = None
elif action == Actions.Long_enter.value:
self._position = Positions.Long
trade_type = "long"
self._last_trade_tick = self._current_tick
elif action == Actions.Short_enter.value:
self._position = Positions.Short
trade_type = "short"
self._last_trade_tick = self._current_tick
elif action == Actions.Long_exit.value:
self._update_total_profit()
self._position = Positions.Neutral
trade_type = "neutral"
self._last_trade_tick = None
elif action == Actions.Short_exit.value:
self._update_total_profit()
self._position = Positions.Neutral
trade_type = "neutral"
self._last_trade_tick = None
else:
print("case not defined")
if trade_type is not None:
self.trade_history.append(
{'price': self.current_price(), 'index': self._current_tick,
'type': trade_type})
if (self._total_profit < self.max_drawdown or
self._total_unrealized_profit < self.max_drawdown):
self._done = True
self._position_history.append(self._position)
info = dict(
tick=self._current_tick,
total_reward=self.total_reward,
total_profit=self._total_profit,
position=self._position.value
)
observation = self._get_observation()
self._update_history(info)
return observation, step_reward, self._done, info
def is_tradesignal(self, action: int) -> bool:
"""
Determine if the signal is a trade signal
e.g.: agent wants a Actions.Long_exit while it is in a Positions.short
"""
return not ((action == Actions.Neutral.value and self._position == Positions.Neutral) or
(action == Actions.Neutral.value and self._position == Positions.Short) or
(action == Actions.Neutral.value and self._position == Positions.Long) or
(action == Actions.Short_enter.value and self._position == Positions.Short) or
(action == Actions.Short_enter.value and self._position == Positions.Long) or
(action == Actions.Short_exit.value and self._position == Positions.Long) or
(action == Actions.Short_exit.value and self._position == Positions.Neutral) or
(action == Actions.Long_enter.value and self._position == Positions.Long) or
(action == Actions.Long_enter.value and self._position == Positions.Short) or
(action == Actions.Long_exit.value and self._position == Positions.Short) or
(action == Actions.Long_exit.value and self._position == Positions.Neutral))
def _is_valid(self, action: int) -> bool:
# trade signal
"""
Determine if the signal is valid.
e.g.: agent wants a Actions.Long_exit while it is in a Positions.short
"""
# Agent should only try to exit if it is in position
if action in (Actions.Short_exit.value, Actions.Long_exit.value):
if self._position not in (Positions.Short, Positions.Long):
return False
# Agent should only try to enter if it is not in position
if action in (Actions.Short_enter.value, Actions.Long_enter.value):
if self._position != Positions.Neutral:
return False
return True

307
freqtrade/freqai/RL/BaseEnvironment.py Normal file
View File

@@ -0,0 +1,307 @@
import logging
import random
from abc import abstractmethod
from enum import Enum
from typing import Optional
import gym
import numpy as np
import pandas as pd
from gym import spaces
from gym.utils import seeding
from pandas import DataFrame
from freqtrade.data.dataprovider import DataProvider
logger = logging.getLogger(__name__)
class Positions(Enum):
Short = 0
Long = 1
Neutral = 0.5
def opposite(self):
return Positions.Short if self == Positions.Long else Positions.Long
class BaseEnvironment(gym.Env):
"""
Base class for environments. This class is agnostic to action count.
Inherited classes customize this to include varying action counts/types,
See RL/Base5ActionRLEnv.py and RL/Base4ActionRLEnv.py
"""
def __init__(self, df: DataFrame = DataFrame(), prices: DataFrame = DataFrame(),
reward_kwargs: dict = {}, window_size=10, starting_point=True,
id: str = 'baseenv-1', seed: int = 1, config: dict = {},
dp: Optional[DataProvider] = None):
"""
Initializes the training/eval environment.
:param df: dataframe of features
:param prices: dataframe of prices to be used in the training environment
:param window_size: size of window (temporal) to pass to the agent
:param reward_kwargs: extra config settings assigned by user in `rl_config`
:param starting_point: start at edge of window or not
:param id: string id of the environment (used in backend for multiprocessed env)
:param seed: Sets the seed of the environment higher in the gym.Env object
:param config: Typical user configuration file
:param dp: dataprovider from freqtrade
"""
self.config = config
self.rl_config = config['freqai']['rl_config']
self.add_state_info = self.rl_config.get('add_state_info', False)
self.id = id
self.seed(seed)
self.reset_env(df, prices, window_size, reward_kwargs, starting_point)
self.max_drawdown = 1 - self.rl_config.get('max_training_drawdown_pct', 0.8)
self.compound_trades = config['stake_amount'] == 'unlimited'
if self.config.get('fee', None) is not None:
self.fee = self.config['fee']
elif dp is not None:
self.fee = dp._exchange.get_fee(symbol=dp.current_whitelist()[0]) # type: ignore
else:
self.fee = 0.0015
def reset_env(self, df: DataFrame, prices: DataFrame, window_size: int,
reward_kwargs: dict, starting_point=True):
"""
Resets the environment when the agent fails (in our case, if the drawdown
exceeds the user set max_training_drawdown_pct)
:param df: dataframe of features
:param prices: dataframe of prices to be used in the training environment
:param window_size: size of window (temporal) to pass to the agent
:param reward_kwargs: extra config settings assigned by user in `rl_config`
:param starting_point: start at edge of window or not
"""
self.df = df
self.signal_features = self.df
self.prices = prices
self.window_size = window_size
self.starting_point = starting_point
self.rr = reward_kwargs["rr"]
self.profit_aim = reward_kwargs["profit_aim"]
# # spaces
if self.add_state_info:
self.total_features = self.signal_features.shape[1] + 3
else:
self.total_features = self.signal_features.shape[1]
self.shape = (window_size, self.total_features)
self.set_action_space()
self.observation_space = spaces.Box(
low=-1, high=1, shape=self.shape, dtype=np.float32)
# episode
self._start_tick: int = self.window_size
self._end_tick: int = len(self.prices) - 1
self._done: bool = False
self._current_tick: int = self._start_tick
self._last_trade_tick: Optional[int] = None
self._position = Positions.Neutral
self._position_history: list = [None]
self.total_reward: float = 0
self._total_profit: float = 1
self._total_unrealized_profit: float = 1
self.history: dict = {}
self.trade_history: list = []
@abstractmethod
def set_action_space(self):
"""
Unique to the environment action count. Must be inherited.
"""
def seed(self, seed: int = 1):
self.np_random, seed = seeding.np_random(seed)
return [seed]
def reset(self):
self._done = False
if self.starting_point is True:
if self.rl_config.get('randomize_starting_position', False):
length_of_data = int(self._end_tick / 4)
start_tick = random.randint(self.window_size + 1, length_of_data)
self._start_tick = start_tick
self._position_history = (self._start_tick * [None]) + [self._position]
else:
self._position_history = (self.window_size * [None]) + [self._position]
self._current_tick = self._start_tick
self._last_trade_tick = None
self._position = Positions.Neutral
self.total_reward = 0.
self._total_profit = 1. # unit
self.history = {}
self.trade_history = []
self.portfolio_log_returns = np.zeros(len(self.prices))
self._profits = [(self._start_tick, 1)]
self.close_trade_profit = []
self._total_unrealized_profit = 1
return self._get_observation()
@abstractmethod
def step(self, action: int):
"""
Step depeneds on action types, this must be inherited.
"""
return
def _get_observation(self):
"""
This may or may not be independent of action types, user can inherit
this in their custom "MyRLEnv"
"""
features_window = self.signal_features[(
self._current_tick - self.window_size):self._current_tick]
if self.add_state_info:
features_and_state = DataFrame(np.zeros((len(features_window), 3)),
columns=['current_profit_pct',
'position',
'trade_duration'],
index=features_window.index)
features_and_state['current_profit_pct'] = self.get_unrealized_profit()
features_and_state['position'] = self._position.value
features_and_state['trade_duration'] = self.get_trade_duration()
features_and_state = pd.concat([features_window, features_and_state], axis=1)
return features_and_state
else:
return features_window
def get_trade_duration(self):
"""
Get the trade duration if the agent is in a trade
"""
if self._last_trade_tick is None:
return 0
else:
return self._current_tick - self._last_trade_tick
def get_unrealized_profit(self):
"""
Get the unrealized profit if the agent is in a trade
"""
if self._last_trade_tick is None:
return 0.
if self._position == Positions.Neutral:
return 0.
elif self._position == Positions.Short:
current_price = self.add_exit_fee(self.prices.iloc[self._current_tick].open)
last_trade_price = self.add_entry_fee(self.prices.iloc[self._last_trade_tick].open)
return (last_trade_price - current_price) / last_trade_price
elif self._position == Positions.Long:
current_price = self.add_entry_fee(self.prices.iloc[self._current_tick].open)
last_trade_price = self.add_exit_fee(self.prices.iloc[self._last_trade_tick].open)
return (current_price - last_trade_price) / last_trade_price
else:
return 0.
@abstractmethod
def is_tradesignal(self, action: int) -> bool:
"""
Determine if the signal is a trade signal. This is
unique to the actions in the environment, and therefore must be
inherited.
"""
return True
def _is_valid(self, action: int) -> bool:
"""
Determine if the signal is valid.This is
unique to the actions in the environment, and therefore must be
inherited.
"""
return True
def add_entry_fee(self, price):
return price * (1 + self.fee)
def add_exit_fee(self, price):
return price / (1 + self.fee)
def _update_history(self, info):
if not self.history:
self.history = {key: [] for key in info.keys()}
for key, value in info.items():
self.history[key].append(value)
@abstractmethod
def calculate_reward(self, action: int) -> float:
"""
An example reward function. This is the one function that users will likely
wish to inject their own creativity into.
:param action: int = The action made by the agent for the current candle.
:return:
float = the reward to give to the agent for current step (used for optimization
of weights in NN)
"""
def _update_unrealized_total_profit(self):
"""
Update the unrealized total profit incase of episode end.
"""
if self._position in (Positions.Long, Positions.Short):
pnl = self.get_unrealized_profit()
if self.compound_trades:
# assumes unit stake and compounding
unrl_profit = self._total_profit * (1 + pnl)
else:
# assumes unit stake and no compounding
unrl_profit = self._total_profit + pnl
self._total_unrealized_profit = unrl_profit
def _update_total_profit(self):
pnl = self.get_unrealized_profit()
if self.compound_trades:
# assumes unit stake and compounding
self._total_profit = self._total_profit * (1 + pnl)
else:
# assumes unit stake and no compounding
self._total_profit += pnl
def current_price(self) -> float:
return self.prices.iloc[self._current_tick].open
# Keeping around incase we want to start building more complex environment
# templates in the future.
# def most_recent_return(self):
# """
# Calculate the tick to tick return if in a trade.
# Return is generated from rising prices in Long
# and falling prices in Short positions.
# The actions Sell/Buy or Hold during a Long position trigger the sell/buy-fee.
# """
# # Long positions
# if self._position == Positions.Long:
# current_price = self.prices.iloc[self._current_tick].open
# previous_price = self.prices.iloc[self._current_tick - 1].open
# if (self._position_history[self._current_tick - 1] == Positions.Short
# or self._position_history[self._current_tick - 1] == Positions.Neutral):
# previous_price = self.add_entry_fee(previous_price)
# return np.log(current_price) - np.log(previous_price)
# # Short positions
# if self._position == Positions.Short:
# current_price = self.prices.iloc[self._current_tick].open
# previous_price = self.prices.iloc[self._current_tick - 1].open
# if (self._position_history[self._current_tick - 1] == Positions.Long
# or self._position_history[self._current_tick - 1] == Positions.Neutral):
# previous_price = self.add_exit_fee(previous_price)
# return np.log(previous_price) - np.log(current_price)
# return 0
# def update_portfolio_log_returns(self, action):
# self.portfolio_log_returns[self._current_tick] = self.most_recent_return(action)

400
freqtrade/freqai/RL/BaseReinforcementLearningModel.py Normal file
View File

@@ -0,0 +1,400 @@
import importlib
import logging
from abc import abstractmethod
from datetime import datetime, timezone
from pathlib import Path
from typing import Any, Callable, Dict, Optional, Tuple, Type, Union
import gym
import numpy as np
import numpy.typing as npt
import pandas as pd
import torch as th
import torch.multiprocessing
from pandas import DataFrame
from stable_baselines3.common.callbacks import EvalCallback
from stable_baselines3.common.monitor import Monitor
from stable_baselines3.common.utils import set_random_seed
from stable_baselines3.common.vec_env import SubprocVecEnv
from freqtrade.exceptions import OperationalException
from freqtrade.freqai.data_kitchen import FreqaiDataKitchen
from freqtrade.freqai.freqai_interface import IFreqaiModel
from freqtrade.freqai.RL.Base5ActionRLEnv import Actions, Base5ActionRLEnv
from freqtrade.freqai.RL.BaseEnvironment import Positions
from freqtrade.persistence import Trade
logger = logging.getLogger(__name__)
torch.multiprocessing.set_sharing_strategy('file_system')
SB3_MODELS = ['PPO', 'A2C', 'DQN']
SB3_CONTRIB_MODELS = ['TRPO', 'ARS', 'RecurrentPPO', 'MaskablePPO']
class BaseReinforcementLearningModel(IFreqaiModel):
"""
User created Reinforcement Learning Model prediction class
"""
def __init__(self, **kwargs) -> None:
super().__init__(config=kwargs['config'])
self.max_threads = min(self.freqai_info['rl_config'].get(
'cpu_count', 1), max(int(self.max_system_threads / 2), 1))
th.set_num_threads(self.max_threads)
self.reward_params = self.freqai_info['rl_config']['model_reward_parameters']
self.train_env: Union[SubprocVecEnv, gym.Env] = None
self.eval_env: Union[SubprocVecEnv, gym.Env] = None
self.eval_callback: Optional[EvalCallback] = None
self.model_type = self.freqai_info['rl_config']['model_type']
self.rl_config = self.freqai_info['rl_config']
self.continual_learning = self.freqai_info.get('continual_learning', False)
if self.model_type in SB3_MODELS:
import_str = 'stable_baselines3'
elif self.model_type in SB3_CONTRIB_MODELS:
import_str = 'sb3_contrib'
else:
raise OperationalException(f'{self.model_type} not available in stable_baselines3 or '
f'sb3_contrib. please choose one of {SB3_MODELS} or '
f'{SB3_CONTRIB_MODELS}')
mod = importlib.import_module(import_str, self.model_type)
self.MODELCLASS = getattr(mod, self.model_type)
self.policy_type = self.freqai_info['rl_config']['policy_type']
self.unset_outlier_removal()
self.net_arch = self.rl_config.get('net_arch', [128, 128])
self.dd.model_type = "stable_baselines"
def unset_outlier_removal(self):
"""
If user has activated any function that may remove training points, this
function will set them to false and warn them
"""
if self.ft_params.get('use_SVM_to_remove_outliers', False):
self.ft_params.update({'use_SVM_to_remove_outliers': False})
logger.warning('User tried to use SVM with RL. Deactivating SVM.')
if self.ft_params.get('use_DBSCAN_to_remove_outliers', False):
self.ft_params.update({'use_DBSCAN_to_remove_outliers': False})
logger.warning('User tried to use DBSCAN with RL. Deactivating DBSCAN.')
if self.freqai_info['data_split_parameters'].get('shuffle', False):
self.freqai_info['data_split_parameters'].update({'shuffle': False})
logger.warning('User tried to shuffle training data. Setting shuffle to False')
def train(
self, unfiltered_df: DataFrame, pair: str, dk: FreqaiDataKitchen, **kwargs
) -> Any:
"""
Filter the training data and train a model to it. Train makes heavy use of the datakitchen
for storing, saving, loading, and analyzing the data.
:param unfiltered_df: Full dataframe for the current training period
:param metadata: pair metadata from strategy.
:returns:
:model: Trained model which can be used to inference (self.predict)
"""
logger.info("--------------------Starting training " f"{pair} --------------------")
features_filtered, labels_filtered = dk.filter_features(
unfiltered_df,
dk.training_features_list,
dk.label_list,
training_filter=True,
)
data_dictionary: Dict[str, Any] = dk.make_train_test_datasets(
features_filtered, labels_filtered)
dk.fit_labels() # FIXME useless for now, but just satiating append methods
# normalize all data based on train_dataset only
prices_train, prices_test = self.build_ohlc_price_dataframes(dk.data_dictionary, pair, dk)
data_dictionary = dk.normalize_data(data_dictionary)
# data cleaning/analysis
self.data_cleaning_train(dk)
logger.info(
f'Training model on {len(dk.data_dictionary["train_features"].columns)}'
f' features and {len(data_dictionary["train_features"])} data points'
)
self.set_train_and_eval_environments(data_dictionary, prices_train, prices_test, dk)
model = self.fit(data_dictionary, dk)
logger.info(f"--------------------done training {pair}--------------------")
return model
def set_train_and_eval_environments(self, data_dictionary: Dict[str, DataFrame],
prices_train: DataFrame, prices_test: DataFrame,
dk: FreqaiDataKitchen):
"""
User can override this if they are using a custom MyRLEnv
:param data_dictionary: dict = common data dictionary containing train and test
features/labels/weights.
:param prices_train/test: DataFrame = dataframe comprised of the prices to be used in the
environment during training or testing
:param dk: FreqaiDataKitchen = the datakitchen for the current pair
"""
train_df = data_dictionary["train_features"]
test_df = data_dictionary["test_features"]
self.train_env = self.MyRLEnv(df=train_df,
prices=prices_train,
window_size=self.CONV_WIDTH,
reward_kwargs=self.reward_params,
config=self.config,
dp=self.data_provider)
self.eval_env = Monitor(self.MyRLEnv(df=test_df,
prices=prices_test,
window_size=self.CONV_WIDTH,
reward_kwargs=self.reward_params,
config=self.config,
dp=self.data_provider))
self.eval_callback = EvalCallback(self.eval_env, deterministic=True,
render=False, eval_freq=len(train_df),
best_model_save_path=str(dk.data_path))
@abstractmethod
def fit(self, data_dictionary: Dict[str, Any], dk: FreqaiDataKitchen, **kwargs):
"""
Agent customizations and abstract Reinforcement Learning customizations
go in here. Abstract method, so this function must be overridden by
user class.
"""
return
def get_state_info(self, pair: str) -> Tuple[float, float, int]:
"""
State info during dry/live (not backtesting) which is fed back
into the model.
:param pair: str = COIN/STAKE to get the environment information for
:return:
:market_side: float = representing short, long, or neutral for
pair
:current_profit: float = unrealized profit of the current trade
:trade_duration: int = the number of candles that the trade has
been open for
"""
open_trades = Trade.get_trades_proxy(is_open=True)
market_side = 0.5
current_profit: float = 0
trade_duration = 0
for trade in open_trades:
if trade.pair == pair:
if self.data_provider._exchange is None: # type: ignore
logger.error('No exchange available.')
return 0, 0, 0
else:
current_rate = self.data_provider._exchange.get_rate( # type: ignore
pair, refresh=False, side="exit", is_short=trade.is_short)
now = datetime.now(timezone.utc).timestamp()
trade_duration = int((now - trade.open_date_utc.timestamp()) / self.base_tf_seconds)
current_profit = trade.calc_profit_ratio(current_rate)
if trade.is_short:
market_side = 0
else:
market_side = 1
return market_side, current_profit, int(trade_duration)
def predict(
self, unfiltered_df: DataFrame, dk: FreqaiDataKitchen, **kwargs
) -> Tuple[DataFrame, npt.NDArray[np.int_]]:
"""
Filter the prediction features data and predict with it.
:param unfiltered_dataframe: Full dataframe for the current backtest period.
:return:
:pred_df: dataframe containing the predictions
:do_predict: np.array of 1s and 0s to indicate places where freqai needed to remove
data (NaNs) or felt uncertain about data (PCA and DI index)
"""
dk.find_features(unfiltered_df)
filtered_dataframe, _ = dk.filter_features(
unfiltered_df, dk.training_features_list, training_filter=False
)
filtered_dataframe = dk.normalize_data_from_metadata(filtered_dataframe)
dk.data_dictionary["prediction_features"] = filtered_dataframe
# optional additional data cleaning/analysis
self.data_cleaning_predict(dk)
pred_df = self.rl_model_predict(
dk.data_dictionary["prediction_features"], dk, self.model)
pred_df.fillna(0, inplace=True)
return (pred_df, dk.do_predict)
def rl_model_predict(self, dataframe: DataFrame,
dk: FreqaiDataKitchen, model: Any) -> DataFrame:
"""
A helper function to make predictions in the Reinforcement learning module.
:param dataframe: DataFrame = the dataframe of features to make the predictions on
:param dk: FreqaiDatakitchen = data kitchen for the current pair
:param model: Any = the trained model used to inference the features.
"""
output = pd.DataFrame(np.zeros(len(dataframe)), columns=dk.label_list)
def _predict(window):
observations = dataframe.iloc[window.index]
if self.live and self.rl_config.get('add_state_info', False):
market_side, current_profit, trade_duration = self.get_state_info(dk.pair)
observations['current_profit_pct'] = current_profit
observations['position'] = market_side
observations['trade_duration'] = trade_duration
res, _ = model.predict(observations, deterministic=True)
return res
output = output.rolling(window=self.CONV_WIDTH).apply(_predict)
return output
def build_ohlc_price_dataframes(self, data_dictionary: dict,
pair: str, dk: FreqaiDataKitchen) -> Tuple[DataFrame,
DataFrame]:
"""
Builds the train prices and test prices for the environment.
"""
pair = pair.replace(':', '')
train_df = data_dictionary["train_features"]
test_df = data_dictionary["test_features"]
# price data for model training and evaluation
tf = self.config['timeframe']
ohlc_list = [f'%-{pair}raw_open_{tf}', f'%-{pair}raw_low_{tf}',
f'%-{pair}raw_high_{tf}', f'%-{pair}raw_close_{tf}']
rename_dict = {f'%-{pair}raw_open_{tf}': 'open', f'%-{pair}raw_low_{tf}': 'low',
f'%-{pair}raw_high_{tf}': ' high', f'%-{pair}raw_close_{tf}': 'close'}
prices_train = train_df.filter(ohlc_list, axis=1)
if prices_train.empty:
raise OperationalException('Reinforcement learning module didnt find the raw prices '
'assigned in populate_any_indicators. Please assign them '
'with:\n'
'informative[f"%-{pair}raw_close"] = informative["close"]\n'
'informative[f"%-{pair}raw_open"] = informative["open"]\n'
'informative[f"%-{pair}raw_high"] = informative["high"]\n'
'informative[f"%-{pair}raw_low"] = informative["low"]\n')
prices_train.rename(columns=rename_dict, inplace=True)
prices_train.reset_index(drop=True)
prices_test = test_df.filter(ohlc_list, axis=1)
prices_test.rename(columns=rename_dict, inplace=True)
prices_test.reset_index(drop=True)
return prices_train, prices_test
def load_model_from_disk(self, dk: FreqaiDataKitchen) -> Any:
"""
Can be used by user if they are trying to limit_ram_usage *and*
perform continual learning.
For now, this is unused.
"""
exists = Path(dk.data_path / f"{dk.model_filename}_model").is_file()
if exists:
model = self.MODELCLASS.load(dk.data_path / f"{dk.model_filename}_model")
else:
logger.info('No model file on disk to continue learning from.')
return model
def _on_stop(self):
"""
Hook called on bot shutdown. Close SubprocVecEnv subprocesses for clean shutdown.
"""
if self.train_env:
self.train_env.close()
if self.eval_env:
self.eval_env.close()
# Nested class which can be overridden by user to customize further
class MyRLEnv(Base5ActionRLEnv):
"""
User can override any function in BaseRLEnv and gym.Env. Here the user
sets a custom reward based on profit and trade duration.
"""
def calculate_reward(self, action: int) -> float:
"""
An example reward function. This is the one function that users will likely
wish to inject their own creativity into.
:param action: int = The action made by the agent for the current candle.
:return:
float = the reward to give to the agent for current step (used for optimization
of weights in NN)
"""
# first, penalize if the action is not valid
if not self._is_valid(action):
return -2
pnl = self.get_unrealized_profit()
factor = 100.
# reward agent for entering trades
if (action in (Actions.Long_enter.value, Actions.Short_enter.value)
and self._position == Positions.Neutral):
return 25
# discourage agent from not entering trades
if action == Actions.Neutral.value and self._position == Positions.Neutral:
return -1
max_trade_duration = self.rl_config.get('max_trade_duration_candles', 300)
if self._last_trade_tick:
trade_duration = self._current_tick - self._last_trade_tick
else:
trade_duration = 0
if trade_duration <= max_trade_duration:
factor *= 1.5
elif trade_duration > max_trade_duration:
factor *= 0.5
# discourage sitting in position
if (self._position in (Positions.Short, Positions.Long) and
action == Actions.Neutral.value):
return -1 * trade_duration / max_trade_duration
# close long
if action == Actions.Long_exit.value and self._position == Positions.Long:
if pnl > self.profit_aim * self.rr:
factor *= self.rl_config['model_reward_parameters'].get('win_reward_factor', 2)
return float(pnl * factor)
# close short
if action == Actions.Short_exit.value and self._position == Positions.Short:
if pnl > self.profit_aim * self.rr:
factor *= self.rl_config['model_reward_parameters'].get('win_reward_factor', 2)
return float(pnl * factor)
return 0.
def make_env(MyRLEnv: Type[gym.Env], env_id: str, rank: int,
seed: int, train_df: DataFrame, price: DataFrame,
reward_params: Dict[str, int], window_size: int, monitor: bool = False,
config: Dict[str, Any] = {}) -> Callable:
"""
Utility function for multiprocessed env.
:param env_id: (str) the environment ID
:param num_env: (int) the number of environment you wish to have in subprocesses
:param seed: (int) the inital seed for RNG
:param rank: (int) index of the subprocess
:return: (Callable)
"""
def _init() -> gym.Env:
env = MyRLEnv(df=train_df, prices=price, window_size=window_size,
reward_kwargs=reward_params, id=env_id, seed=seed + rank, config=config)
if monitor:
env = Monitor(env)
return env
set_random_seed(seed)
return _init

0
freqtrade/freqai/RL/__init__.py Normal file
View File

84
freqtrade/freqai/data_drawer.py
View File

@@ -1,9 +1,10 @@
import collections
import importlib
import logging
import re
import shutil
import threading
from datetime import datetime, timezone
from datetime import datetime, timedelta, timezone
from pathlib import Path
from typing import Any, Dict, Tuple, TypedDict
@@ -81,6 +82,7 @@ class FreqaiDataDrawer:
self.historic_predictions_bkp_path = Path(
self.full_path / "historic_predictions.backup.pkl")
self.pair_dictionary_path = Path(self.full_path / "pair_dictionary.json")
self.global_metadata_path = Path(self.full_path / "global_metadata.json")
self.metric_tracker_path = Path(self.full_path / "metric_tracker.json")
self.follow_mode = follow_mode
if follow_mode:
@@ -98,6 +100,7 @@ class FreqaiDataDrawer:
self.empty_pair_dict: pair_info = {
"model_filename": "", "trained_timestamp": 0,
"data_path": "", "extras": {}}
self.model_type = self.freqai_info.get('model_save_type', 'joblib')
def update_metric_tracker(self, metric: str, value: float, pair: str) -> None:
"""
@@ -125,6 +128,17 @@ class FreqaiDataDrawer:
self.update_metric_tracker('cpu_load5min', load5 / cpus, pair)
self.update_metric_tracker('cpu_load15min', load15 / cpus, pair)
def load_global_metadata_from_disk(self):
"""
Locate and load a previously saved global metadata in present model folder.
"""
exists = self.global_metadata_path.is_file()
if exists:
with open(self.global_metadata_path, "r") as fp:
metatada_dict = rapidjson.load(fp, number_mode=rapidjson.NM_NATIVE)
return metatada_dict
return {}
def load_drawer_from_disk(self):
"""
Locate and load a previously saved data drawer full of all pair model metadata in
@@ -225,6 +239,15 @@ class FreqaiDataDrawer:
rapidjson.dump(self.follower_dict, fp, default=self.np_encoder,
number_mode=rapidjson.NM_NATIVE)
def save_global_metadata_to_disk(self, metadata: Dict[str, Any]):
"""
Save global metadata json to disk
"""
with self.save_lock:
with open(self.global_metadata_path, 'w') as fp:
rapidjson.dump(metadata, fp, default=self.np_encoder,
number_mode=rapidjson.NM_NATIVE)
def create_follower_dict(self):
"""
Create or dictionary for each follower to maintain unique persistent prediction targets
@@ -476,10 +499,12 @@ class FreqaiDataDrawer:
save_path = Path(dk.data_path)
# Save the trained model
if not dk.keras:
if self.model_type == 'joblib':
dump(model, save_path / f"{dk.model_filename}_model.joblib")
else:
elif self.model_type == 'keras':
model.save(save_path / f"{dk.model_filename}_model.h5")
elif 'stable_baselines' in self.model_type:
model.save(save_path / f"{dk.model_filename}_model.zip")
if dk.svm_model is not None:
dump(dk.svm_model, save_path / f"{dk.model_filename}_svm_model.joblib")
@@ -506,11 +531,10 @@ class FreqaiDataDrawer:
dk.pca, open(dk.data_path / f"{dk.model_filename}_pca_object.pkl", "wb")
)
# if self.live:
# store as much in ram as possible to increase performance
self.model_dictionary[coin] = model
self.pair_dict[coin]["model_filename"] = dk.model_filename
self.pair_dict[coin]["data_path"] = str(dk.data_path)
if coin not in self.meta_data_dictionary:
self.meta_data_dictionary[coin] = {}
self.meta_data_dictionary[coin]["train_df"] = dk.data_dictionary["train_features"]
@@ -542,14 +566,6 @@ class FreqaiDataDrawer:
if dk.live:
dk.model_filename = self.pair_dict[coin]["model_filename"]
dk.data_path = Path(self.pair_dict[coin]["data_path"])
if self.freqai_info.get("follow_mode", False):
# follower can be on a different system which is rsynced from the leader:
dk.data_path = Path(
self.config["user_data_dir"]
/ "models"
/ dk.data_path.parts[-2]
/ dk.data_path.parts[-1]
)
if coin in self.meta_data_dictionary:
dk.data = self.meta_data_dictionary[coin]["meta_data"]
@@ -568,12 +584,16 @@ class FreqaiDataDrawer:
# try to access model in memory instead of loading object from disk to save time
if dk.live and coin in self.model_dictionary:
model = self.model_dictionary[coin]
elif not dk.keras:
elif self.model_type == 'joblib':
model = load(dk.data_path / f"{dk.model_filename}_model.joblib")
else:
elif self.model_type == 'keras':
from tensorflow import keras
model = keras.models.load_model(dk.data_path / f"{dk.model_filename}_model.h5")
elif self.model_type == 'stable_baselines':
mod = importlib.import_module(
'stable_baselines3', self.freqai_info['rl_config']['model_type'])
MODELCLASS = getattr(mod, self.freqai_info['rl_config']['model_type'])
model = MODELCLASS.load(dk.data_path / f"{dk.model_filename}_model")
if Path(dk.data_path / f"{dk.model_filename}_svm_model.joblib").is_file():
dk.svm_model = load(dk.data_path / f"{dk.model_filename}_svm_model.joblib")
@@ -583,6 +603,10 @@ class FreqaiDataDrawer:
f"Unable to load model, ensure model exists at " f"{dk.data_path} "
)
# load it into ram if it was loaded from disk
if coin not in self.model_dictionary:
self.model_dictionary[coin] = model
if self.config["freqai"]["feature_parameters"]["principal_component_analysis"]:
dk.pca = cloudpickle.load(
open(dk.data_path / f"{dk.model_filename}_pca_object.pkl", "rb")
@@ -693,3 +717,31 @@ class FreqaiDataDrawer:
).reset_index(drop=True)
return corr_dataframes, base_dataframes
def get_timerange_from_live_historic_predictions(self) -> TimeRange:
"""
Returns timerange information based on historic predictions file
:return: timerange calculated from saved live data
"""
if not self.historic_predictions_path.is_file():
raise OperationalException(
'Historic predictions not found. Historic predictions data is required '
'to run backtest with the freqai-backtest-live-models option '
)
self.load_historic_predictions_from_disk()
all_pairs_end_dates = []
for pair in self.historic_predictions:
pair_historic_data = self.historic_predictions[pair]
all_pairs_end_dates.append(pair_historic_data.date_pred.max())
global_metadata = self.load_global_metadata_from_disk()
start_date = datetime.fromtimestamp(int(global_metadata["start_dry_live_date"]))
end_date = max(all_pairs_end_dates)
# add 1 day to string timerange to ensure BT module will load all dataframe data
end_date = end_date + timedelta(days=1)
backtesting_timerange = TimeRange(
'date', 'date', int(start_date.timestamp()), int(end_date.timestamp())
)
return backtesting_timerange

186
freqtrade/freqai/data_kitchen.py
View File

@@ -1,7 +1,7 @@
import copy
import logging
import shutil
from datetime import datetime, timedelta, timezone
from datetime import datetime, timezone
from math import cos, sin
from pathlib import Path
from typing import Any, Dict, List, Tuple
@@ -9,6 +9,7 @@ from typing import Any, Dict, List, Tuple
import numpy as np
import numpy.typing as npt
import pandas as pd
import psutil
from pandas import DataFrame
from scipy import stats
from sklearn import linear_model
@@ -86,12 +87,7 @@ class FreqaiDataKitchen:
if not self.live:
self.full_path = self.get_full_models_path(self.config)
if self.backtest_live_models:
if self.pair:
self.set_timerange_from_ready_models()
(self.training_timeranges,
self.backtesting_timeranges) = self.split_timerange_live_models()
else:
if not self.backtest_live_models:
self.full_timerange = self.create_fulltimerange(
self.config["timerange"], self.freqai_config.get("train_period_days", 0)
)
@@ -102,7 +98,10 @@ class FreqaiDataKitchen:
)
self.data['extra_returns_per_train'] = self.freqai_config.get('extra_returns_per_train', {})
self.thread_count = self.freqai_config.get("data_kitchen_thread_count", -1)
if not self.freqai_config.get("data_kitchen_thread_count", 0):
self.thread_count = max(int(psutil.cpu_count() * 2 - 2), 1)
else:
self.thread_count = self.freqai_config["data_kitchen_thread_count"]
self.train_dates: DataFrame = pd.DataFrame()
self.unique_classes: Dict[str, list] = {}
self.unique_class_list: list = []
@@ -433,9 +432,7 @@ class FreqaiDataKitchen:
timerange_train.stopts = timerange_train.startts + train_period_days
first = False
start = datetime.fromtimestamp(timerange_train.startts, tz=timezone.utc)
stop = datetime.fromtimestamp(timerange_train.stopts, tz=timezone.utc)
tr_training_list.append(start.strftime("%Y%m%d") + "-" + stop.strftime("%Y%m%d"))
tr_training_list.append(timerange_train.timerange_str)
tr_training_list_timerange.append(copy.deepcopy(timerange_train))
# associated backtest period
@@ -447,9 +444,7 @@ class FreqaiDataKitchen:
if timerange_backtest.stopts > config_timerange.stopts:
timerange_backtest.stopts = config_timerange.stopts
start = datetime.fromtimestamp(timerange_backtest.startts, tz=timezone.utc)
stop = datetime.fromtimestamp(timerange_backtest.stopts, tz=timezone.utc)
tr_backtesting_list.append(start.strftime("%Y%m%d") + "-" + stop.strftime("%Y%m%d"))
tr_backtesting_list.append(timerange_backtest.timerange_str)
tr_backtesting_list_timerange.append(copy.deepcopy(timerange_backtest))
# ensure we are predicting on exactly same amount of data as requested by user defined
@@ -460,29 +455,6 @@ class FreqaiDataKitchen:
# print(tr_training_list, tr_backtesting_list)
return tr_training_list_timerange, tr_backtesting_list_timerange
def split_timerange_live_models(
self
) -> Tuple[list, list]:
tr_backtesting_list_timerange = []
asset = self.pair.split("/")[0]
if asset not in self.backtest_live_models_data["assets_end_dates"]:
raise OperationalException(
f"Model not available for pair {self.pair}. "
"Please, try again after removing this pair from the configuration file."
)
asset_data = self.backtest_live_models_data["assets_end_dates"][asset]
backtesting_timerange = self.backtest_live_models_data["backtesting_timerange"]
model_end_dates = [x for x in asset_data]
model_end_dates.append(backtesting_timerange.stopts)
model_end_dates.sort()
for index, item in enumerate(model_end_dates):
if len(model_end_dates) > (index + 1):
tr_to_add = TimeRange("date", "date", item, model_end_dates[index + 1])
tr_backtesting_list_timerange.append(tr_to_add)
return tr_backtesting_list_timerange, tr_backtesting_list_timerange
def slice_dataframe(self, timerange: TimeRange, df: DataFrame) -> DataFrame:
"""
Given a full dataframe, extract the user desired window
@@ -491,11 +463,9 @@ class FreqaiDataKitchen:
it is sliced down to just the present training period.
"""
start = datetime.fromtimestamp(timerange.startts, tz=timezone.utc)
stop = datetime.fromtimestamp(timerange.stopts, tz=timezone.utc)
df = df.loc[df["date"] >= start, :]
df = df.loc[df["date"] >= timerange.startdt, :]
if not self.live:
df = df.loc[df["date"] < stop, :]
df = df.loc[df["date"] < timerange.stopdt, :]
return df
@@ -980,7 +950,8 @@ class FreqaiDataKitchen:
return weights
def get_predictions_to_append(self, predictions: DataFrame,
do_predict: npt.ArrayLike) -> DataFrame:
do_predict: npt.ArrayLike,
dataframe_backtest: DataFrame) -> DataFrame:
"""
Get backtest prediction from current backtest period
"""
@@ -1002,7 +973,9 @@ class FreqaiDataKitchen:
if self.freqai_config["feature_parameters"].get("DI_threshold", 0) > 0:
append_df["DI_values"] = self.DI_values
return append_df
dataframe_backtest.reset_index(drop=True, inplace=True)
merged_df = pd.concat([dataframe_backtest["date"], append_df], axis=1)
return merged_df
def append_predictions(self, append_df: DataFrame) -> None:
"""
@@ -1012,23 +985,18 @@ class FreqaiDataKitchen:
if self.full_df.empty:
self.full_df = append_df
else:
self.full_df = pd.concat([self.full_df, append_df], axis=0)
self.full_df = pd.concat([self.full_df, append_df], axis=0, ignore_index=True)
def fill_predictions(self, dataframe):
"""
Back fill values to before the backtesting range so that the dataframe matches size
when it goes back to the strategy. These rows are not included in the backtest.
"""
len_filler = len(dataframe) - len(self.full_df.index) # startup_candle_count
filler_df = pd.DataFrame(
np.zeros((len_filler, len(self.full_df.columns))), columns=self.full_df.columns
)
self.full_df = pd.concat([filler_df, self.full_df], axis=0, ignore_index=True)
to_keep = [col for col in dataframe.columns if not col.startswith("&")]
self.return_dataframe = pd.concat([dataframe[to_keep], self.full_df], axis=1)
self.return_dataframe = pd.merge(dataframe[to_keep],
self.full_df, how='left', on='date')
self.return_dataframe[self.full_df.columns] = (
self.return_dataframe[self.full_df.columns].fillna(value=0))
self.full_df = DataFrame()
return
@@ -1058,9 +1026,7 @@ class FreqaiDataKitchen:
backtest_timerange.startts = (
backtest_timerange.startts - backtest_period_days * SECONDS_IN_DAY
)
start = datetime.fromtimestamp(backtest_timerange.startts, tz=timezone.utc)
stop = datetime.fromtimestamp(backtest_timerange.stopts, tz=timezone.utc)
full_timerange = start.strftime("%Y%m%d") + "-" + stop.strftime("%Y%m%d")
full_timerange = backtest_timerange.timerange_str
config_path = Path(self.config["config_files"][0])
if not self.full_path.is_dir():
@@ -1327,22 +1293,22 @@ class FreqaiDataKitchen:
self, append_df: DataFrame
) -> None:
"""
Save prediction dataframe from backtesting to h5 file format
Save prediction dataframe from backtesting to feather file format
:param append_df: dataframe for backtesting period
"""
full_predictions_folder = Path(self.full_path / self.backtest_predictions_folder)
if not full_predictions_folder.is_dir():
full_predictions_folder.mkdir(parents=True, exist_ok=True)
append_df.to_hdf(self.backtesting_results_path, key='append_df', mode='w')
append_df.to_feather(self.backtesting_results_path)
def get_backtesting_prediction(
self
) -> DataFrame:
"""
Get prediction dataframe from h5 file format
Get prediction dataframe from feather file format
"""
append_df = pd.read_hdf(self.backtesting_results_path)
append_df = pd.read_feather(self.backtesting_results_path)
return append_df
def check_if_backtest_prediction_is_valid(
@@ -1358,19 +1324,20 @@ class FreqaiDataKitchen:
"""
path_to_predictionfile = Path(self.full_path /
self.backtest_predictions_folder /
f"{self.model_filename}_prediction.h5")
f"{self.model_filename}_prediction.feather")
self.backtesting_results_path = path_to_predictionfile
file_exists = path_to_predictionfile.is_file()
if file_exists:
append_df = self.get_backtesting_prediction()
if len(append_df) == len_backtest_df:
if len(append_df) == len_backtest_df and 'date' in append_df:
logger.info(f"Found backtesting prediction file at {path_to_predictionfile}")
return True
else:
logger.info("A new backtesting prediction file is required. "
"(Number of predictions is different from dataframe length).")
"(Number of predictions is different from dataframe length or "
"old prediction file version).")
return False
else:
logger.info(
@@ -1378,17 +1345,6 @@ class FreqaiDataKitchen:
)
return False
def set_timerange_from_ready_models(self):
backtesting_timerange, \
assets_end_dates = (
self.get_timerange_and_assets_end_dates_from_ready_models(self.full_path))
self.backtest_live_models_data = {
"backtesting_timerange": backtesting_timerange,
"assets_end_dates": assets_end_dates
}
return
def get_full_models_path(self, config: Config) -> Path:
"""
Returns default FreqAI model path
@@ -1399,88 +1355,6 @@ class FreqaiDataKitchen:
config["user_data_dir"] / "models" / str(freqai_config.get("identifier"))
)
def get_timerange_and_assets_end_dates_from_ready_models(
self, models_path: Path) -> Tuple[TimeRange, Dict[str, Any]]:
"""
Returns timerange information based on a FreqAI model directory
:param models_path: FreqAI model path
:return: a Tuple with (Timerange calculated from directory and
a Dict with pair and model end training dates info)
"""
all_models_end_dates = []
assets_end_dates: Dict[str, Any] = self.get_assets_timestamps_training_from_ready_models(
models_path)
for key in assets_end_dates:
for model_end_date in assets_end_dates[key]:
if model_end_date not in all_models_end_dates:
all_models_end_dates.append(model_end_date)
if len(all_models_end_dates) == 0:
raise OperationalException(
'At least 1 saved model is required to '
'run backtest with the freqai-backtest-live-models option'
)
if len(all_models_end_dates) == 1:
logger.warning(
"Only 1 model was found. Backtesting will run with the "
"timerange from the end of the training date to the current date"
)
finish_timestamp = int(datetime.now(tz=timezone.utc).timestamp())
if len(all_models_end_dates) > 1:
# After last model end date, use the same period from previous model
# to finish the backtest
all_models_end_dates.sort(reverse=True)
finish_timestamp = all_models_end_dates[0] + \
(all_models_end_dates[0] - all_models_end_dates[1])
all_models_end_dates.append(finish_timestamp)
all_models_end_dates.sort()
start_date = (datetime(*datetime.fromtimestamp(min(all_models_end_dates),
timezone.utc).timetuple()[:3], tzinfo=timezone.utc))
end_date = (datetime(*datetime.fromtimestamp(max(all_models_end_dates),
timezone.utc).timetuple()[:3], tzinfo=timezone.utc))
# add 1 day to string timerange to ensure BT module will load all dataframe data
end_date = end_date + timedelta(days=1)
backtesting_timerange = TimeRange(
'date', 'date', int(start_date.timestamp()), int(end_date.timestamp())
)
return backtesting_timerange, assets_end_dates
def get_assets_timestamps_training_from_ready_models(
self, models_path: Path) -> Dict[str, Any]:
"""
Scan the models path and returns all assets end training dates (timestamp)
:param models_path: FreqAI model path
:return: a Dict with asset and model end training dates info
"""
assets_end_dates: Dict[str, Any] = {}
if not models_path.is_dir():
raise OperationalException(
'Model folders not found. Saved models are required '
'to run backtest with the freqai-backtest-live-models option'
)
for model_dir in models_path.iterdir():
if str(model_dir.name).startswith("sub-train"):
model_end_date = int(model_dir.name.split("_")[1])
asset = model_dir.name.split("_")[0].replace("sub-train-", "")
model_file_name = (
f"cb_{str(model_dir.name).replace('sub-train-', '').lower()}"
"_model.joblib"
)
model_path_file = Path(model_dir / model_file_name)
if model_path_file.is_file():
if asset not in assets_end_dates:
assets_end_dates[asset] = []
assets_end_dates[asset].append(model_end_date)
return assets_end_dates
def remove_special_chars_from_feature_names(self, dataframe: pd.DataFrame) -> pd.DataFrame:
"""
Remove all special characters from feature strings (:)

146
freqtrade/freqai/freqai_interface.py
View File

@@ -5,15 +5,17 @@ from abc import ABC, abstractmethod
from collections import deque
from datetime import datetime, timezone
from pathlib import Path
from typing import Any, Dict, List, Literal, Tuple
from typing import Any, Dict, List, Literal, Optional, Tuple
import numpy as np
import pandas as pd
import psutil
from numpy.typing import NDArray
from pandas import DataFrame
from freqtrade.configuration import TimeRange
from freqtrade.constants import DATETIME_PRINT_FORMAT, Config
from freqtrade.constants import Config
from freqtrade.data.dataprovider import DataProvider
from freqtrade.enums import RunMode
from freqtrade.exceptions import OperationalException
from freqtrade.exchange import timeframe_to_seconds
@@ -67,6 +69,7 @@ class IFreqaiModel(ABC):
self.save_backtest_models: bool = self.freqai_info.get("save_backtest_models", True)
if self.save_backtest_models:
logger.info('Backtesting module configured to save all models.')
self.dd = FreqaiDataDrawer(Path(self.full_path), self.config, self.follow_mode)
# set current candle to arbitrary historical date
self.current_candle: datetime = datetime.fromtimestamp(637887600, tz=timezone.utc)
@@ -98,6 +101,9 @@ class IFreqaiModel(ABC):
self.get_corr_dataframes: bool = True
self._threads: List[threading.Thread] = []
self._stop_event = threading.Event()
self.metadata: Dict[str, Any] = self.dd.load_global_metadata_from_disk()
self.data_provider: Optional[DataProvider] = None
self.max_system_threads = max(int(psutil.cpu_count() * 2 - 2), 1)
record_params(config, self.full_path)
@@ -126,11 +132,13 @@ class IFreqaiModel(ABC):
self.live = strategy.dp.runmode in (RunMode.DRY_RUN, RunMode.LIVE)
self.dd.set_pair_dict_info(metadata)
self.data_provider = strategy.dp
if self.live:
self.inference_timer('start')
self.dk = FreqaiDataKitchen(self.config, self.live, metadata["pair"])
dk = self.start_live(dataframe, metadata, strategy, self.dk)
dataframe = dk.remove_features_from_df(dk.return_dataframe)
# For backtesting, each pair enters and then gets trained for each window along the
# sliding window defined by "train_period_days" (training window) and "live_retrain_hours"
@@ -139,20 +147,24 @@ class IFreqaiModel(ABC):
# the concatenated results for the full backtesting period back to the strategy.
elif not self.follow_mode:
self.dk = FreqaiDataKitchen(self.config, self.live, metadata["pair"])
if self.dk.backtest_live_models:
logger.info(
f"Backtesting {len(self.dk.backtesting_timeranges)} timeranges (live models)")
else:
logger.info(f"Training {len(self.dk.training_timeranges)} timeranges")
dataframe = self.dk.use_strategy_to_populate_indicators(
strategy, prediction_dataframe=dataframe, pair=metadata["pair"]
)
dk = self.start_backtesting(dataframe, metadata, self.dk)
if not self.config.get("freqai_backtest_live_models", False):
logger.info(f"Training {len(self.dk.training_timeranges)} timeranges")
dk = self.start_backtesting(dataframe, metadata, self.dk)
dataframe = dk.remove_features_from_df(dk.return_dataframe)
else:
logger.info(
"Backtesting using historic predictions (live models)")
dk = self.start_backtesting_from_historic_predictions(
dataframe, metadata, self.dk)
dataframe = dk.return_dataframe
dataframe = dk.remove_features_from_df(dk.return_dataframe)
self.clean_up()
if self.live:
self.inference_timer('stop', metadata["pair"])
return dataframe
def clean_up(self):
@@ -164,6 +176,13 @@ class IFreqaiModel(ABC):
self.model = None
self.dk = None
def _on_stop(self):
"""
Callback for Subclasses to override to include logic for shutting down resources
when SIGINT is sent.
"""
return
def shutdown(self):
"""
Cleans up threads on Shutdown, set stop event. Join threads to wait
@@ -172,6 +191,9 @@ class IFreqaiModel(ABC):
logger.info("Stopping FreqAI")
self._stop_event.set()
self.data_provider = None
self._on_stop()
logger.info("Waiting on Training iteration")
for _thread in self._threads:
_thread.join()
@@ -301,10 +323,11 @@ class IFreqaiModel(ABC):
self.model = self.dd.load_data(pair, dk)
pred_df, do_preds = self.predict(dataframe_backtest, dk)
append_df = dk.get_predictions_to_append(pred_df, do_preds)
append_df = dk.get_predictions_to_append(pred_df, do_preds, dataframe_backtest)
dk.append_predictions(append_df)
dk.save_backtesting_prediction(append_df)
self.backtesting_fit_live_predictions(dk)
dk.fill_predictions(dataframe)
return dk
@@ -617,6 +640,8 @@ class IFreqaiModel(ABC):
self.dd.historic_predictions[pair] = pred_df
hist_preds_df = self.dd.historic_predictions[pair]
self.set_start_dry_live_date(strat_df)
for label in hist_preds_df.columns:
if hist_preds_df[label].dtype == object:
continue
@@ -629,7 +654,7 @@ class IFreqaiModel(ABC):
hist_preds_df['DI_values'] = 0
for return_str in dk.data['extra_returns_per_train']:
hist_preds_df[return_str] = 0
hist_preds_df[return_str] = dk.data['extra_returns_per_train'][return_str]
hist_preds_df['close_price'] = strat_df['close']
hist_preds_df['date_pred'] = strat_df['date']
@@ -657,7 +682,8 @@ class IFreqaiModel(ABC):
for label in full_labels:
if self.dd.historic_predictions[dk.pair][label].dtype == object:
continue
f = spy.stats.norm.fit(self.dd.historic_predictions[dk.pair][label].tail(num_candles))
f = spy.stats.norm.fit(
self.dd.historic_predictions[dk.pair][label].tail(num_candles))
dk.data["labels_mean"][label], dk.data["labels_std"][label] = f[0], f[1]
return
@@ -788,14 +814,8 @@ class IFreqaiModel(ABC):
:return: if the data exists or not
"""
if self.config.get("freqai_backtest_live_models", False) and len(dataframe_backtest) == 0:
tr_backtest_startts_str = datetime.fromtimestamp(
tr_backtest.startts,
tz=timezone.utc).strftime(DATETIME_PRINT_FORMAT)
tr_backtest_stopts_str = datetime.fromtimestamp(
tr_backtest.stopts,
tz=timezone.utc).strftime(DATETIME_PRINT_FORMAT)
logger.info(f"No data found for pair {pair} from {tr_backtest_startts_str} "
f" from {tr_backtest_startts_str} to {tr_backtest_stopts_str}. "
logger.info(f"No data found for pair {pair} from "
f"from { tr_backtest.start_fmt} to {tr_backtest.stop_fmt}. "
"Probably more than one training within the same candle period.")
return False
return True
@@ -810,20 +830,88 @@ class IFreqaiModel(ABC):
:param pair: the current pair
:param total_trains: total trains (total number of slides for the sliding window)
"""
tr_train_startts_str = datetime.fromtimestamp(
tr_train.startts,
tz=timezone.utc).strftime(DATETIME_PRINT_FORMAT)
tr_train_stopts_str = datetime.fromtimestamp(
tr_train.stopts,
tz=timezone.utc).strftime(DATETIME_PRINT_FORMAT)
if not self.config.get("freqai_backtest_live_models", False):
logger.info(
f"Training {pair}, {self.pair_it}/{self.total_pairs} pairs"
f" from {tr_train_startts_str} "
f"to {tr_train_stopts_str}, {train_it}/{total_trains} "
f" from {tr_train.start_fmt} "
f"to {tr_train.stop_fmt}, {train_it}/{total_trains} "
"trains"
)
def backtesting_fit_live_predictions(self, dk: FreqaiDataKitchen):
"""
Apply fit_live_predictions function in backtesting with a dummy historic_predictions
The loop is required to simulate dry/live operation, as it is not possible to predict
the type of logic implemented by the user.
:param dk: datakitchen object
"""
fit_live_predictions_candles = self.freqai_info.get("fit_live_predictions_candles", 0)
if fit_live_predictions_candles:
logger.info("Applying fit_live_predictions in backtesting")
label_columns = [col for col in dk.full_df.columns if (
col.startswith("&") and
not (col.startswith("&") and col.endswith("_mean")) and
not (col.startswith("&") and col.endswith("_std")) and
col not in self.dk.data["extra_returns_per_train"])
]
for index in range(len(dk.full_df)):
if index >= fit_live_predictions_candles:
self.dd.historic_predictions[self.dk.pair] = (
dk.full_df.iloc[index - fit_live_predictions_candles:index])
self.fit_live_predictions(self.dk, self.dk.pair)
for label in label_columns:
if dk.full_df[label].dtype == object:
continue
if "labels_mean" in self.dk.data:
dk.full_df.at[index, f"{label}_mean"] = (
self.dk.data["labels_mean"][label])
if "labels_std" in self.dk.data:
dk.full_df.at[index, f"{label}_std"] = self.dk.data["labels_std"][label]
for extra_col in self.dk.data["extra_returns_per_train"]:
dk.full_df.at[index, f"{extra_col}"] = (
self.dk.data["extra_returns_per_train"][extra_col])
return
def update_metadata(self, metadata: Dict[str, Any]):
"""
Update global metadata and save the updated json file
:param metadata: new global metadata dict
"""
self.dd.save_global_metadata_to_disk(metadata)
self.metadata = metadata
def set_start_dry_live_date(self, live_dataframe: DataFrame):
key_name = "start_dry_live_date"
if key_name not in self.metadata:
metadata = self.metadata
metadata[key_name] = int(
pd.to_datetime(live_dataframe.tail(1)["date"].values[0]).timestamp())
self.update_metadata(metadata)
def start_backtesting_from_historic_predictions(
self, dataframe: DataFrame, metadata: dict, dk: FreqaiDataKitchen
) -> FreqaiDataKitchen:
"""
:param dataframe: DataFrame = strategy passed dataframe
:param metadata: Dict = pair metadata
:param dk: FreqaiDataKitchen = Data management/analysis tool associated to present pair only
:return:
FreqaiDataKitchen = Data management/analysis tool associated to present pair only
"""
pair = metadata["pair"]
dk.return_dataframe = dataframe
saved_dataframe = self.dd.historic_predictions[pair]
columns_to_drop = list(set(saved_dataframe.columns).intersection(
dk.return_dataframe.columns))
dk.return_dataframe = dk.return_dataframe.drop(columns=list(columns_to_drop))
dk.return_dataframe = pd.merge(
dk.return_dataframe, saved_dataframe, how='left', left_on='date', right_on="date_pred")
# dk.return_dataframe = dk.return_dataframe[saved_dataframe.columns].fillna(0)
return dk
# Following methods which are overridden by user made prediction models.
# See freqai/prediction_models/CatboostPredictionModel.py for an example.

141
freqtrade/freqai/prediction_models/ReinforcementLearner.py Normal file
View File

@@ -0,0 +1,141 @@
import logging
from pathlib import Path
from typing import Any, Dict
import torch as th
from freqtrade.freqai.data_kitchen import FreqaiDataKitchen
from freqtrade.freqai.RL.Base5ActionRLEnv import Actions, Base5ActionRLEnv, Positions
from freqtrade.freqai.RL.BaseReinforcementLearningModel import BaseReinforcementLearningModel
logger = logging.getLogger(__name__)
class ReinforcementLearner(BaseReinforcementLearningModel):
"""
Reinforcement Learning Model prediction model.
Users can inherit from this class to make their own RL model with custom
environment/training controls. Define the file as follows:
```
from freqtrade.freqai.prediction_models.ReinforcementLearner import ReinforcementLearner
class MyCoolRLModel(ReinforcementLearner):
```
Save the file to `user_data/freqaimodels`, then run it with:
freqtrade trade --freqaimodel MyCoolRLModel --config config.json --strategy SomeCoolStrat
Here the users can override any of the functions
available in the `IFreqaiModel` inheritance tree. Most importantly for RL, this
is where the user overrides `MyRLEnv` (see below), to define custom
`calculate_reward()` function, or to override any other parts of the environment.
This class also allows users to override any other part of the IFreqaiModel tree.
For example, the user can override `def fit()` or `def train()` or `def predict()`
to take fine-tuned control over these processes.
Another common override may be `def data_cleaning_predict()` where the user can
take fine-tuned control over the data handling pipeline.
"""
def fit(self, data_dictionary: Dict[str, Any], dk: FreqaiDataKitchen, **kwargs):
"""
User customizable fit method
:param data_dictionary: dict = common data dictionary containing all train/test
features/labels/weights.
:param dk: FreqaiDatakitchen = data kitchen for current pair.
:return:
model Any = trained model to be used for inference in dry/live/backtesting
"""
train_df = data_dictionary["train_features"]
total_timesteps = self.freqai_info["rl_config"]["train_cycles"] * len(train_df)
policy_kwargs = dict(activation_fn=th.nn.ReLU,
net_arch=self.net_arch)
if dk.pair not in self.dd.model_dictionary or not self.continual_learning:
model = self.MODELCLASS(self.policy_type, self.train_env, policy_kwargs=policy_kwargs,
tensorboard_log=Path(
dk.full_path / "tensorboard" / dk.pair.split('/')[0]),
**self.freqai_info['model_training_parameters']
)
else:
logger.info('Continual training activated - starting training from previously '
'trained agent.')
model = self.dd.model_dictionary[dk.pair]
model.set_env(self.train_env)
model.learn(
total_timesteps=int(total_timesteps),
callback=self.eval_callback
)
if Path(dk.data_path / "best_model.zip").is_file():
logger.info('Callback found a best model.')
best_model = self.MODELCLASS.load(dk.data_path / "best_model")
return best_model
logger.info('Couldnt find best model, using final model instead.')
return model
class MyRLEnv(Base5ActionRLEnv):
"""
User can override any function in BaseRLEnv and gym.Env. Here the user
sets a custom reward based on profit and trade duration.
"""
def calculate_reward(self, action: int) -> float:
"""
An example reward function. This is the one function that users will likely
wish to inject their own creativity into.
:param action: int = The action made by the agent for the current candle.
:return:
float = the reward to give to the agent for current step (used for optimization
of weights in NN)
"""
# first, penalize if the action is not valid
if not self._is_valid(action):
return -2
pnl = self.get_unrealized_profit()
factor = 100.
# reward agent for entering trades
if (action in (Actions.Long_enter.value, Actions.Short_enter.value)
and self._position == Positions.Neutral):
return 25
# discourage agent from not entering trades
if action == Actions.Neutral.value and self._position == Positions.Neutral:
return -1
max_trade_duration = self.rl_config.get('max_trade_duration_candles', 300)
trade_duration = self._current_tick - self._last_trade_tick # type: ignore
if trade_duration <= max_trade_duration:
factor *= 1.5
elif trade_duration > max_trade_duration:
factor *= 0.5
# discourage sitting in position
if (self._position in (Positions.Short, Positions.Long) and
action == Actions.Neutral.value):
return -1 * trade_duration / max_trade_duration
# close long
if action == Actions.Long_exit.value and self._position == Positions.Long:
if pnl > self.profit_aim * self.rr:
factor *= self.rl_config['model_reward_parameters'].get('win_reward_factor', 2)
return float(pnl * factor)
# close short
if action == Actions.Short_exit.value and self._position == Positions.Short:
if pnl > self.profit_aim * self.rr:
factor *= self.rl_config['model_reward_parameters'].get('win_reward_factor', 2)
return float(pnl * factor)
return 0.

51
freqtrade/freqai/prediction_models/ReinforcementLearner_multiproc.py Normal file
View File

@@ -0,0 +1,51 @@
import logging
from typing import Any, Dict # , Tuple
# import numpy.typing as npt
from pandas import DataFrame
from stable_baselines3.common.callbacks import EvalCallback
from stable_baselines3.common.vec_env import SubprocVecEnv
from freqtrade.freqai.data_kitchen import FreqaiDataKitchen
from freqtrade.freqai.prediction_models.ReinforcementLearner import ReinforcementLearner
from freqtrade.freqai.RL.BaseReinforcementLearningModel import make_env
logger = logging.getLogger(__name__)
class ReinforcementLearner_multiproc(ReinforcementLearner):
"""
Demonstration of how to build vectorized environments
"""
def set_train_and_eval_environments(self, data_dictionary: Dict[str, Any],
prices_train: DataFrame, prices_test: DataFrame,
dk: FreqaiDataKitchen):
"""
User can override this if they are using a custom MyRLEnv
:param data_dictionary: dict = common data dictionary containing train and test
features/labels/weights.
:param prices_train/test: DataFrame = dataframe comprised of the prices to be used in
the environment during training
or testing
:param dk: FreqaiDataKitchen = the datakitchen for the current pair
"""
train_df = data_dictionary["train_features"]
test_df = data_dictionary["test_features"]
env_id = "train_env"
self.train_env = SubprocVecEnv([make_env(self.MyRLEnv, env_id, i, 1, train_df, prices_train,
self.reward_params, self.CONV_WIDTH, monitor=True,
config=self.config) for i
in range(self.max_threads)])
eval_env_id = 'eval_env'
self.eval_env = SubprocVecEnv([make_env(self.MyRLEnv, eval_env_id, i, 1,
test_df, prices_test,
self.reward_params, self.CONV_WIDTH, monitor=True,
config=self.config) for i
in range(self.max_threads)])
self.eval_callback = EvalCallback(self.eval_env, deterministic=True,
render=False, eval_freq=len(train_df),
best_model_save_path=str(dk.data_path))

9
freqtrade/freqai/utils.py
View File

@@ -14,6 +14,7 @@ from freqtrade.data.history.history_utils import refresh_backtest_ohlcv_data
from freqtrade.exceptions import OperationalException
from freqtrade.exchange import timeframe_to_seconds
from freqtrade.exchange.exchange import market_is_active
from freqtrade.freqai.data_drawer import FreqaiDataDrawer
from freqtrade.freqai.data_kitchen import FreqaiDataKitchen
from freqtrade.plugins.pairlist.pairlist_helpers import dynamic_expand_pairlist
@@ -229,8 +230,6 @@ def get_timerange_backtest_live_models(config: Config) -> str:
"""
dk = FreqaiDataKitchen(config)
models_path = dk.get_full_models_path(config)
timerange, _ = dk.get_timerange_and_assets_end_dates_from_ready_models(models_path)
start_date = datetime.fromtimestamp(timerange.startts, tz=timezone.utc)
end_date = datetime.fromtimestamp(timerange.stopts, tz=timezone.utc)
tr = f"{start_date.strftime('%Y%m%d')}-{end_date.strftime('%Y%m%d')}"
return tr
dd = FreqaiDataDrawer(models_path, config)
timerange = dd.get_timerange_from_live_historic_predictions()
return timerange.timerange_str

45
freqtrade/freqtradebot.py
View File

@@ -191,10 +191,10 @@ class FreqtradeBot(LoggingMixin):
# Check whether markets have to be reloaded and reload them when it's needed
self.exchange.reload_markets()
self.update_closed_trades_without_assigned_fees()
self.update_trades_without_assigned_fees()
# Query trades from persistence layer
trades = Trade.get_open_trades()
trades: List[Trade] = Trade.get_open_trades()
self.active_pair_whitelist = self._refresh_active_whitelist(trades)
@@ -354,7 +354,7 @@ class FreqtradeBot(LoggingMixin):
if self.trading_mode == TradingMode.FUTURES:
self._schedule.run_pending()
def update_closed_trades_without_assigned_fees(self) -> None:
def update_trades_without_assigned_fees(self) -> None:
"""
Update closed trades without close fees assigned.
Only acts when Orders are in the database, otherwise the last order-id is unknown.
@@ -381,15 +381,16 @@ class FreqtradeBot(LoggingMixin):
trades = Trade.get_open_trades_without_assigned_fees()
for trade in trades:
if trade.is_open and not trade.fee_updated(trade.entry_side):
order = trade.select_order(trade.entry_side, False)
open_order = trade.select_order(trade.entry_side, True)
if order and open_order is None:
logger.info(
f"Updating {trade.entry_side}-fee on trade {trade}"
f"for order {order.order_id}."
)
self.update_trade_state(trade, order.order_id, send_msg=False)
with self._exit_lock:
if trade.is_open and not trade.fee_updated(trade.entry_side):
order = trade.select_order(trade.entry_side, False)
open_order = trade.select_order(trade.entry_side, True)
if order and open_order is None:
logger.info(
f"Updating {trade.entry_side}-fee on trade {trade}"
f"for order {order.order_id}."
)
self.update_trade_state(trade, order.order_id, send_msg=False)
def handle_insufficient_funds(self, trade: Trade):
"""
@@ -826,6 +827,8 @@ class FreqtradeBot(LoggingMixin):
co = self.exchange.cancel_stoploss_order_with_result(
trade.stoploss_order_id, trade.pair, trade.amount)
trade.update_order(co)
# Reset stoploss order id.
trade.stoploss_order_id = None
except InvalidOrderException:
logger.exception(f"Could not cancel stoploss order {trade.stoploss_order_id}")
return trade
@@ -982,7 +985,7 @@ class FreqtradeBot(LoggingMixin):
# SELL / exit positions / close trades logic and methods
#
def exit_positions(self, trades: List[Any]) -> int:
def exit_positions(self, trades: List[Trade]) -> int:
"""
Tries to execute exit orders for open trades (positions)
"""
@@ -1010,7 +1013,7 @@ class FreqtradeBot(LoggingMixin):
def handle_trade(self, trade: Trade) -> bool:
"""
Sells/exits_short the current pair if the threshold is reached and updates the trade record.
Exits the current pair if the threshold is reached and updates the trade record.
:return: True if trade has been sold/exited_short, False otherwise
"""
if not trade.is_open:
@@ -1133,10 +1136,8 @@ class FreqtradeBot(LoggingMixin):
trade.exit_reason = ExitType.STOPLOSS_ON_EXCHANGE.value
self.update_trade_state(trade, trade.stoploss_order_id, stoploss_order,
stoploss_order=True)
# Lock pair for one candle to prevent immediate rebuys
self.strategy.lock_pair(trade.pair, datetime.now(timezone.utc),
reason='Auto lock')
self._notify_exit(trade, "stoploss", True)
self.handle_protections(trade.pair, trade.trade_direction)
return True
if trade.open_order_id or not trade.is_open:
@@ -1150,7 +1151,7 @@ class FreqtradeBot(LoggingMixin):
stoploss = (
self.edge.stoploss(pair=trade.pair)
if self.edge else
self.strategy.stoploss / trade.leverage
trade.stop_loss_pct / trade.leverage
)
if trade.is_short:
stop_price = trade.open_rate * (1 - stoploss)
@@ -1169,7 +1170,6 @@ class FreqtradeBot(LoggingMixin):
if self.create_stoploss_order(trade=trade, stop_price=trade.stoploss_or_liquidation):
return False
else:
trade.stoploss_order_id = None
logger.warning('Stoploss order was cancelled, but unable to recreate one.')
# Finally we check if stoploss on exchange should be moved up because of trailing.
@@ -1595,11 +1595,6 @@ class FreqtradeBot(LoggingMixin):
trade.close_rate_requested = limit
trade.exit_reason = exit_reason
if not sub_trade_amt:
# Lock pair for one candle to prevent immediate re-trading
self.strategy.lock_pair(trade.pair, datetime.now(timezone.utc),
reason='Auto lock')
self._notify_exit(trade, order_type, sub_trade=bool(sub_trade_amt), order=order_obj)
# In case of market sell orders the order can be closed immediately
if order.get('status', 'unknown') in ('closed', 'expired'):
@@ -1809,6 +1804,8 @@ class FreqtradeBot(LoggingMixin):
self._notify_enter(trade, order, fill=True, sub_trade=sub_trade)
def handle_protections(self, pair: str, side: LongShort) -> None:
# Lock pair for one candle to prevent immediate rebuys
self.strategy.lock_pair(pair, datetime.now(timezone.utc), reason='Auto lock')
prot_trig = self.protections.stop_per_pair(pair, side=side)
if prot_trig:
msg = {'type': RPCMessageType.PROTECTION_TRIGGER, }

23
freqtrade/misc.py
View File

@@ -10,7 +10,8 @@ from typing import Any, Dict, Iterator, List, Mapping, Union
from typing.io import IO
from urllib.parse import urlparse
import pandas
import orjson
import pandas as pd
import rapidjson
from freqtrade.constants import DECIMAL_PER_COIN_FALLBACK, DECIMALS_PER_COIN
@@ -256,29 +257,37 @@ def parse_db_uri_for_logging(uri: str):
return parsed_db_uri.geturl().replace(f':{pwd}@', ':*****@')
def dataframe_to_json(dataframe: pandas.DataFrame) -> str:
def dataframe_to_json(dataframe: pd.DataFrame) -> str:
"""
Serialize a DataFrame for transmission over the wire using JSON
:param dataframe: A pandas DataFrame
:returns: A JSON string of the pandas DataFrame
"""
return dataframe.to_json(orient='split')
# https://github.com/pandas-dev/pandas/issues/24889
# https://github.com/pandas-dev/pandas/issues/40443
# We need to convert to a dict to avoid mem leak
def default(z):
if isinstance(z, pd.Timestamp):
return z.timestamp() * 1e3
raise TypeError
return str(orjson.dumps(dataframe.to_dict(orient='split'), default=default), 'utf-8')
def json_to_dataframe(data: str) -> pandas.DataFrame:
def json_to_dataframe(data: str) -> pd.DataFrame:
"""
Deserialize JSON into a DataFrame
:param data: A JSON string
:returns: A pandas DataFrame from the JSON string
"""
dataframe = pandas.read_json(data, orient='split')
dataframe = pd.read_json(data, orient='split')
if 'date' in dataframe.columns:
dataframe['date'] = pandas.to_datetime(dataframe['date'], unit='ms', utc=True)
dataframe['date'] = pd.to_datetime(dataframe['date'], unit='ms', utc=True)
return dataframe
def remove_entry_exit_signals(dataframe: pandas.DataFrame):
def remove_entry_exit_signals(dataframe: pd.DataFrame):
"""
Remove Entry and Exit signals from a DataFrame

74
freqtrade/optimize/backtesting.py
View File

@@ -692,10 +692,11 @@ class Backtesting:
trade.orders.append(order)
return trade
def _get_exit_trade_entry(self, trade: LocalTrade, row: Tuple) -> Optional[LocalTrade]:
def _get_exit_trade_entry(
self, trade: LocalTrade, row: Tuple, is_first: bool) -> Optional[LocalTrade]:
exit_candle_time: datetime = row[DATE_IDX].to_pydatetime()
if self.trading_mode == TradingMode.FUTURES:
if is_first and self.trading_mode == TradingMode.FUTURES:
trade.funding_fees = self.exchange.calculate_funding_fees(
self.futures_data[trade.pair],
amount=trade.amount,
@@ -704,32 +705,7 @@ class Backtesting:
close_date=exit_candle_time,
)
if self.timeframe_detail and trade.pair in self.detail_data:
exit_candle_end = exit_candle_time + timedelta(minutes=self.timeframe_min)
detail_data = self.detail_data[trade.pair]
detail_data = detail_data.loc[
(detail_data['date'] >= exit_candle_time) &
(detail_data['date'] < exit_candle_end)
].copy()
if len(detail_data) == 0:
# Fall back to "regular" data if no detail data was found for this candle
return self._get_exit_trade_entry_for_candle(trade, row)
detail_data.loc[:, 'enter_long'] = row[LONG_IDX]
detail_data.loc[:, 'exit_long'] = row[ELONG_IDX]
detail_data.loc[:, 'enter_short'] = row[SHORT_IDX]
detail_data.loc[:, 'exit_short'] = row[ESHORT_IDX]
detail_data.loc[:, 'enter_tag'] = row[ENTER_TAG_IDX]
detail_data.loc[:, 'exit_tag'] = row[EXIT_TAG_IDX]
for det_row in detail_data[HEADERS].values.tolist():
res = self._get_exit_trade_entry_for_candle(trade, det_row)
if res:
return res
return None
else:
return self._get_exit_trade_entry_for_candle(trade, row)
return self._get_exit_trade_entry_for_candle(trade, row)
def get_valid_price_and_stake(
self, pair: str, row: Tuple, propose_rate: float, stake_amount: float,
@@ -1074,7 +1050,7 @@ class Backtesting:
def backtest_loop(
self, row: Tuple, pair: str, current_time: datetime, end_date: datetime,
max_open_trades: int, open_trade_count_start: int) -> int:
max_open_trades: int, open_trade_count_start: int, is_first: bool = True) -> int:
"""
NOTE: This method is used by Hyperopt at each iteration. Please keep it optimized.
@@ -1092,9 +1068,11 @@ class Backtesting:
# without positionstacking, we can only have one open trade per pair.
# max_open_trades must be respected
# don't open on the last row
# We only open trades on the main candle, not on detail candles
trade_dir = self.check_for_trade_entry(row)
if (
(self._position_stacking or len(LocalTrade.bt_trades_open_pp[pair]) == 0)
and is_first
and self.trade_slot_available(max_open_trades, open_trade_count_start)
and current_time != end_date
and trade_dir is not None
@@ -1120,7 +1098,7 @@ class Backtesting:
# 4. Create exit orders (if any)
if not trade.open_order_id:
self._get_exit_trade_entry(trade, row) # Place exit order if necessary
self._get_exit_trade_entry(trade, row, is_first) # Place exit order if necessary
# 5. Process exit orders.
order = trade.select_order(trade.exit_side, is_open=True)
@@ -1171,7 +1149,6 @@ class Backtesting:
self.progress.init_step(BacktestState.BACKTEST, int(
(end_date - start_date) / timedelta(minutes=self.timeframe_min)))
# Loop timerange and get candle for each pair at that point in time
while current_time <= end_date:
open_trade_count_start = LocalTrade.bt_open_open_trade_count
@@ -1185,9 +1162,37 @@ class Backtesting:
row_index += 1
indexes[pair] = row_index
self.dataprovider._set_dataframe_max_index(row_index)
current_detail_time: datetime = row[DATE_IDX].to_pydatetime()
if self.timeframe_detail and pair in self.detail_data:
exit_candle_end = current_detail_time + timedelta(minutes=self.timeframe_min)
open_trade_count_start = self.backtest_loop(
row, pair, current_time, end_date, max_open_trades, open_trade_count_start)
detail_data = self.detail_data[pair]
detail_data = detail_data.loc[
(detail_data['date'] >= current_detail_time) &
(detail_data['date'] < exit_candle_end)
].copy()
if len(detail_data) == 0:
# Fall back to "regular" data if no detail data was found for this candle
open_trade_count_start = self.backtest_loop(
row, pair, current_time, end_date, max_open_trades,
open_trade_count_start)
detail_data.loc[:, 'enter_long'] = row[LONG_IDX]
detail_data.loc[:, 'exit_long'] = row[ELONG_IDX]
detail_data.loc[:, 'enter_short'] = row[SHORT_IDX]
detail_data.loc[:, 'exit_short'] = row[ESHORT_IDX]
detail_data.loc[:, 'enter_tag'] = row[ENTER_TAG_IDX]
detail_data.loc[:, 'exit_tag'] = row[EXIT_TAG_IDX]
is_first = True
current_time_det = current_time
for det_row in detail_data[HEADERS].values.tolist():
open_trade_count_start = self.backtest_loop(
det_row, pair, current_time_det, end_date, max_open_trades,
open_trade_count_start, is_first)
current_time_det += timedelta(minutes=self.timeframe_detail_min)
is_first = False
else:
open_trade_count_start = self.backtest_loop(
row, pair, current_time, end_date, max_open_trades, open_trade_count_start)
# Move time one configured time_interval ahead.
self.progress.increment()
@@ -1286,8 +1291,7 @@ class Backtesting:
def _get_min_cached_backtest_date(self):
min_backtest_date = None
backtest_cache_age = self.config.get('backtest_cache', constants.BACKTEST_CACHE_DEFAULT)
if self.timerange.stopts == 0 or datetime.fromtimestamp(
self.timerange.stopts, tz=timezone.utc) > datetime.now(tz=timezone.utc):
if self.timerange.stopts == 0 or self.timerange.stopdt > datetime.now(tz=timezone.utc):
logger.warning('Backtest result caching disabled due to use of open-ended timerange.')
elif backtest_cache_age == 'day':
min_backtest_date = datetime.now(tz=timezone.utc) - timedelta(days=1)

17
freqtrade/optimize/hyperopt_tools.py
View File

@@ -17,6 +17,7 @@ from freqtrade.enums import HyperoptState
from freqtrade.exceptions import OperationalException
from freqtrade.misc import deep_merge_dicts, round_coin_value, round_dict, safe_value_fallback2
from freqtrade.optimize.hyperopt_epoch_filters import hyperopt_filter_epochs
from freqtrade.optimize.optimize_reports import generate_wins_draws_losses
logger = logging.getLogger(__name__)
@@ -325,8 +326,10 @@ class HyperoptTools():
# New mode, using backtest result for metrics
trials['results_metrics.winsdrawslosses'] = trials.apply(
lambda x: f"{x['results_metrics.wins']} {x['results_metrics.draws']:>4} "
f"{x['results_metrics.losses']:>4}", axis=1)
lambda x: generate_wins_draws_losses(
x['results_metrics.wins'], x['results_metrics.draws'],
x['results_metrics.losses']
), axis=1)
trials = trials[['Best', 'current_epoch', 'results_metrics.total_trades',
'results_metrics.winsdrawslosses',
@@ -337,7 +340,7 @@ class HyperoptTools():
'loss', 'is_initial_point', 'is_random', 'is_best']]
trials.columns = [
'Best', 'Epoch', 'Trades', ' Win Draw Loss', 'Avg profit',
'Best', 'Epoch', 'Trades', ' Win Draw Loss Win%', 'Avg profit',
'Total profit', 'Profit', 'Avg duration', 'max_drawdown', 'max_drawdown_account',
'max_drawdown_abs', 'Objective', 'is_initial_point', 'is_random', 'is_best'
]
@@ -467,9 +470,9 @@ class HyperoptTools():
base_metrics = ['Best', 'current_epoch', 'results_metrics.total_trades',
'results_metrics.profit_mean', 'results_metrics.profit_median',
'results_metrics.profit_total',
'Stake currency',
'results_metrics.profit_total', 'Stake currency',
'results_metrics.profit_total_abs', 'results_metrics.holding_avg',
'results_metrics.trade_count_long', 'results_metrics.trade_count_short',
'loss', 'is_initial_point', 'is_best']
perc_multi = 100
@@ -477,7 +480,9 @@ class HyperoptTools():
trials = trials[base_metrics + param_metrics]
base_columns = ['Best', 'Epoch', 'Trades', 'Avg profit', 'Median profit', 'Total profit',
'Stake currency', 'Profit', 'Avg duration', 'Objective',
'Stake currency', 'Profit', 'Avg duration',
'Trade count long', 'Trade count short',
'Objective',
'is_initial_point', 'is_best']
param_columns = list(results[0]['params_dict'].keys())
trials.columns = base_columns + param_columns

10
freqtrade/optimize/optimize_reports.py
View File

@@ -86,7 +86,7 @@ def _get_line_header(first_column: str, stake_currency: str,
'Win Draw Loss Win%']
def _generate_wins_draws_losses(wins, draws, losses):
def generate_wins_draws_losses(wins, draws, losses):
if wins > 0 and losses == 0:
wl_ratio = '100'
elif wins == 0:
@@ -600,7 +600,7 @@ def text_table_bt_results(pair_results: List[Dict[str, Any]], stake_currency: st
output = [[
t['key'], t['trades'], t['profit_mean_pct'], t['profit_sum_pct'], t['profit_total_abs'],
t['profit_total_pct'], t['duration_avg'],
_generate_wins_draws_losses(t['wins'], t['draws'], t['losses'])
generate_wins_draws_losses(t['wins'], t['draws'], t['losses'])
] for t in pair_results]
# Ignore type as floatfmt does allow tuples but mypy does not know that
return tabulate(output, headers=headers,
@@ -626,7 +626,7 @@ def text_table_exit_reason(exit_reason_stats: List[Dict[str, Any]], stake_curren
output = [[
t.get('exit_reason', t.get('sell_reason')), t['trades'],
_generate_wins_draws_losses(t['wins'], t['draws'], t['losses']),
generate_wins_draws_losses(t['wins'], t['draws'], t['losses']),
t['profit_mean_pct'], t['profit_sum_pct'],
round_coin_value(t['profit_total_abs'], stake_currency, False),
t['profit_total_pct'],
@@ -656,7 +656,7 @@ def text_table_tags(tag_type: str, tag_results: List[Dict[str, Any]], stake_curr
t['profit_total_abs'],
t['profit_total_pct'],
t['duration_avg'],
_generate_wins_draws_losses(
generate_wins_draws_losses(
t['wins'],
t['draws'],
t['losses'])] for t in tag_results]
@@ -715,7 +715,7 @@ def text_table_strategy(strategy_results, stake_currency: str) -> str:
output = [[
t['key'], t['trades'], t['profit_mean_pct'], t['profit_sum_pct'], t['profit_total_abs'],
t['profit_total_pct'], t['duration_avg'],
_generate_wins_draws_losses(t['wins'], t['draws'], t['losses']), drawdown]
generate_wins_draws_losses(t['wins'], t['draws'], t['losses']), drawdown]
for t, drawdown in zip(strategy_results, drawdown)]
# Ignore type as floatfmt does allow tuples but mypy does not know that
return tabulate(output, headers=headers,

2
freqtrade/persistence/pairlock_middleware.py
View File

@@ -87,7 +87,7 @@ class PairLocks():
Get the lock that expires the latest for the pair given.
"""
locks = PairLocks.get_pair_locks(pair, now, side=side)
locks = sorted(locks, key=lambda l: l.lock_end_time, reverse=True)
locks = sorted(locks, key=lambda lock: lock.lock_end_time, reverse=True)
return locks[0] if locks else None
@staticmethod

7
freqtrade/persistence/trade_model.py
View File

@@ -90,6 +90,13 @@ class Order(_DECL_BASE):
def safe_filled(self) -> float:
return self.filled if self.filled is not None else self.amount or 0.0
@property
def safe_remaining(self) -> float:
return (
self.remaining if self.remaining is not None else
self.amount - (self.filled or 0.0)
)
@property
def safe_fee_base(self) -> float:
return self.ft_fee_base or 0.0

2
freqtrade/rpc/api_server/api_auth.py
View File

@@ -81,8 +81,6 @@ async def validate_ws_token(
except HTTPException:
pass
# No checks passed, deny the connection
logger.debug("Denying websocket request.")
# If it doesn't match, close the websocket connection
await ws.close(code=status.WS_1008_POLICY_VIOLATION)

108
freqtrade/rpc/api_server/api_ws.py
View File

@@ -1,16 +1,16 @@
import logging
import time
from typing import Any, Dict
from fastapi import APIRouter, Depends, WebSocketDisconnect
from fastapi.websockets import WebSocket, WebSocketState
from fastapi import APIRouter, Depends
from fastapi.websockets import WebSocket
from pydantic import ValidationError
from websockets.exceptions import WebSocketException
from freqtrade.enums import RPCMessageType, RPCRequestType
from freqtrade.rpc.api_server.api_auth import validate_ws_token
from freqtrade.rpc.api_server.deps import get_channel_manager, get_rpc
from freqtrade.rpc.api_server.ws import WebSocketChannel
from freqtrade.rpc.api_server.ws.channel import ChannelManager
from freqtrade.rpc.api_server.deps import get_message_stream, get_rpc
from freqtrade.rpc.api_server.ws.channel import WebSocketChannel, create_channel
from freqtrade.rpc.api_server.ws.message_stream import MessageStream
from freqtrade.rpc.api_server.ws_schemas import (WSAnalyzedDFMessage, WSMessageSchema,
WSRequestSchema, WSWhitelistMessage)
from freqtrade.rpc.rpc import RPC
@@ -22,23 +22,35 @@ logger = logging.getLogger(__name__)
router = APIRouter()
async def is_websocket_alive(ws: WebSocket) -> bool:
async def channel_reader(channel: WebSocketChannel, rpc: RPC):
"""
Check if a FastAPI Websocket is still open
Iterate over the messages from the channel and process the request
"""
if (
ws.application_state == WebSocketState.CONNECTED and
ws.client_state == WebSocketState.CONNECTED
):
return True
return False
async for message in channel:
await _process_consumer_request(message, channel, rpc)
async def channel_broadcaster(channel: WebSocketChannel, message_stream: MessageStream):
"""
Iterate over messages in the message stream and send them
"""
async for message, ts in message_stream:
if channel.subscribed_to(message.get('type')):
# Log a warning if this channel is behind
# on the message stream by a lot
if (time.time() - ts) > 60:
logger.warning(f"Channel {channel} is behind MessageStream by 1 minute,"
" this can cause a memory leak if you see this message"
" often, consider reducing pair list size or amount of"
" consumers.")
await channel.send(message, timeout=True)
async def _process_consumer_request(
request: Dict[str, Any],
channel: WebSocketChannel,
rpc: RPC,
channel_manager: ChannelManager
rpc: RPC
):
"""
Validate and handle a request from a websocket consumer
@@ -74,65 +86,29 @@ async def _process_consumer_request(
# Format response
response = WSWhitelistMessage(data=whitelist)
# Send it back
await channel_manager.send_direct(channel, response.dict(exclude_none=True))
await channel.send(response.dict(exclude_none=True))
elif type == RPCRequestType.ANALYZED_DF:
limit = None
if data:
# Limit the amount of candles per dataframe to 'limit' or 1500
limit = max(data.get('limit', 1500), 1500)
# Limit the amount of candles per dataframe to 'limit' or 1500
limit = min(data.get('limit', 1500), 1500) if data else None
# For every pair in the generator, send a separate message
for message in rpc._ws_request_analyzed_df(limit):
# Format response
response = WSAnalyzedDFMessage(data=message)
await channel_manager.send_direct(channel, response.dict(exclude_none=True))
await channel.send(response.dict(exclude_none=True))
@router.websocket("/message/ws")
async def message_endpoint(
ws: WebSocket,
websocket: WebSocket,
token: str = Depends(validate_ws_token),
rpc: RPC = Depends(get_rpc),
channel_manager=Depends(get_channel_manager),
token: str = Depends(validate_ws_token)
message_stream: MessageStream = Depends(get_message_stream)
):
"""
Message WebSocket endpoint, facilitates sending RPC messages
"""
try:
channel = await channel_manager.on_connect(ws)
if await is_websocket_alive(ws):
logger.info(f"Consumer connected - {channel}")
# Keep connection open until explicitly closed, and process requests
try:
while not channel.is_closed():
request = await channel.recv()
# Process the request here
await _process_consumer_request(request, channel, rpc, channel_manager)
except (WebSocketDisconnect, WebSocketException):
# Handle client disconnects
logger.info(f"Consumer disconnected - {channel}")
except RuntimeError:
# Handle cases like -
# RuntimeError('Cannot call "send" once a closed message has been sent')
pass
except Exception as e:
logger.info(f"Consumer connection failed - {channel}: {e}")
logger.debug(e, exc_info=e)
except RuntimeError:
# WebSocket was closed
# Do nothing
pass
except Exception as e:
logger.error(f"Failed to serve - {ws.client}")
# Log tracebacks to keep track of what errors are happening
logger.exception(e)
finally:
if channel:
await channel_manager.on_disconnect(ws)
if token:
async with create_channel(websocket) as channel:
await channel.run_channel_tasks(
channel_reader(channel, rpc),
channel_broadcaster(channel, message_stream)
)

4
freqtrade/rpc/api_server/deps.py
View File

@@ -41,8 +41,8 @@ def get_exchange(config=Depends(get_config)):
return ApiServer._exchange
def get_channel_manager():
return ApiServer._ws_channel_manager
def get_message_stream():
return ApiServer._message_stream
def is_webserver_mode(config=Depends(get_config)):

105
freqtrade/rpc/api_server/webserver.py
View File

@@ -1,22 +1,17 @@
import asyncio
import logging
from ipaddress import IPv4Address
from threading import Thread
from typing import Any, Dict, Optional
import orjson
import uvicorn
from fastapi import Depends, FastAPI
from fastapi.middleware.cors import CORSMiddleware
# Look into alternatives
from janus import Queue as ThreadedQueue
from starlette.responses import JSONResponse
from freqtrade.constants import Config
from freqtrade.exceptions import OperationalException
from freqtrade.rpc.api_server.uvicorn_threaded import UvicornServer
from freqtrade.rpc.api_server.ws import ChannelManager
from freqtrade.rpc.api_server.ws_schemas import WSMessageSchemaType
from freqtrade.rpc.api_server.ws.message_stream import MessageStream
from freqtrade.rpc.rpc import RPC, RPCException, RPCHandler
@@ -50,10 +45,8 @@ class ApiServer(RPCHandler):
_config: Config = {}
# Exchange - only available in webserver mode.
_exchange = None
# websocket message queue stuff
_ws_channel_manager: ChannelManager
_ws_thread = None
_ws_loop: Optional[asyncio.AbstractEventLoop] = None
# websocket message stuff
_message_stream: Optional[MessageStream] = None
def __new__(cls, *args, **kwargs):
"""
@@ -71,15 +64,11 @@ class ApiServer(RPCHandler):
return
self._standalone: bool = standalone
self._server = None
self._ws_queue: Optional[ThreadedQueue] = None
self._ws_background_task = None
ApiServer.__initialized = True
api_config = self._config['api_server']
ApiServer._ws_channel_manager = ChannelManager()
self.app = FastAPI(title="Freqtrade API",
docs_url='/docs' if api_config.get('enable_openapi', False) else None,
redoc_url=None,
@@ -105,21 +94,9 @@ class ApiServer(RPCHandler):
del ApiServer._rpc
if self._server and not self._standalone:
logger.info("Stopping API Server")
# self._server.force_exit, self._server.should_exit = True, True
self._server.cleanup()
if self._ws_thread and self._ws_loop:
logger.info("Stopping API Server background tasks")
if self._ws_background_task:
# Cancel the queue task
self._ws_background_task.cancel()
self._ws_thread.join()
self._ws_thread = None
self._ws_loop = None
self._ws_background_task = None
@classmethod
def shutdown(cls):
cls.__initialized = False
@@ -129,9 +106,11 @@ class ApiServer(RPCHandler):
cls._rpc = None
def send_msg(self, msg: Dict[str, Any]) -> None:
if self._ws_queue:
sync_q = self._ws_queue.sync_q
sync_q.put(msg)
"""
Publish the message to the message stream
"""
if ApiServer._message_stream:
ApiServer._message_stream.publish(msg)
def handle_rpc_exception(self, request, exc):
logger.exception(f"API Error calling: {exc}")
@@ -170,51 +149,30 @@ class ApiServer(RPCHandler):
)
app.add_exception_handler(RPCException, self.handle_rpc_exception)
app.add_event_handler(
event_type="startup",
func=self._api_startup_event
)
app.add_event_handler(
event_type="shutdown",
func=self._api_shutdown_event
)
def start_message_queue(self):
if self._ws_thread:
return
async def _api_startup_event(self):
"""
Creates the MessageStream class on startup
so it has access to the same event loop
as uvicorn
"""
if not ApiServer._message_stream:
ApiServer._message_stream = MessageStream()
# Create a new loop, as it'll be just for the background thread
self._ws_loop = asyncio.new_event_loop()
# Start the thread
self._ws_thread = Thread(target=self._ws_loop.run_forever)
self._ws_thread.start()
# Finally, submit the coro to the thread
self._ws_background_task = asyncio.run_coroutine_threadsafe(
self._broadcast_queue_data(), loop=self._ws_loop)
async def _broadcast_queue_data(self) -> None:
# Instantiate the queue in this coroutine so it's attached to our loop
self._ws_queue = ThreadedQueue()
async_queue = self._ws_queue.async_q
try:
while True:
logger.debug("Getting queue messages...")
# Get data from queue
message: WSMessageSchemaType = await async_queue.get()
logger.debug(f"Found message of type: {message.get('type')}")
async_queue.task_done()
# Broadcast it
await self._ws_channel_manager.broadcast(message)
except asyncio.CancelledError:
pass
# For testing, shouldn't happen when stable
except Exception as e:
logger.exception(f"Exception happened in background task: {e}")
finally:
# Disconnect channels and stop the loop on cancel
await self._ws_channel_manager.disconnect_all()
if self._ws_loop:
self._ws_loop.stop()
# Avoid adding more items to the queue if they aren't
# going to get broadcasted.
self._ws_queue = None
async def _api_shutdown_event(self):
"""
Removes the MessageStream class on shutdown
"""
if ApiServer._message_stream:
ApiServer._message_stream = None
def start_api(self):
"""
@@ -254,7 +212,6 @@ class ApiServer(RPCHandler):
if self._standalone:
self._server.run()
else:
self.start_message_queue()
self._server.run_in_thread()
except Exception:
logger.exception("Api server failed to start.")

3
freqtrade/rpc/api_server/ws/__init__.py
View File

@@ -3,4 +3,5 @@
from freqtrade.rpc.api_server.ws.types import WebSocketType
from freqtrade.rpc.api_server.ws.proxy import WebSocketProxy
from freqtrade.rpc.api_server.ws.serializer import HybridJSONWebSocketSerializer
from freqtrade.rpc.api_server.ws.channel import ChannelManager, WebSocketChannel
from freqtrade.rpc.api_server.ws.channel import WebSocketChannel
from freqtrade.rpc.api_server.ws.message_stream import MessageStream

263
freqtrade/rpc/api_server/ws/channel.py
View File

@@ -1,11 +1,13 @@
import asyncio
import logging
import time
from threading import RLock
from typing import Any, Dict, List, Optional, Type, Union
from collections import deque
from contextlib import asynccontextmanager
from typing import Any, AsyncIterator, Deque, Dict, List, Optional, Type, Union
from uuid import uuid4
from fastapi import WebSocket as FastAPIWebSocket
from fastapi import WebSocketDisconnect
from websockets.exceptions import ConnectionClosed
from freqtrade.rpc.api_server.ws.proxy import WebSocketProxy
from freqtrade.rpc.api_server.ws.serializer import (HybridJSONWebSocketSerializer,
@@ -21,31 +23,27 @@ class WebSocketChannel:
"""
Object to help facilitate managing a websocket connection
"""
def __init__(
self,
websocket: WebSocketType,
channel_id: Optional[str] = None,
drain_timeout: int = 3,
throttle: float = 0.01,
serializer_cls: Type[WebSocketSerializer] = HybridJSONWebSocketSerializer
):
self.channel_id = channel_id if channel_id else uuid4().hex[:8]
# The WebSocket object
self._websocket = WebSocketProxy(websocket)
self.drain_timeout = drain_timeout
self.throttle = throttle
self._subscriptions: List[str] = []
# 32 is the size of the receiving queue in websockets package
self.queue: asyncio.Queue[Dict[str, Any]] = asyncio.Queue(maxsize=32)
self._relay_task = asyncio.create_task(self.relay())
# Internal event to signify a closed websocket
self._closed = asyncio.Event()
# The async tasks created for the channel
self._channel_tasks: List[asyncio.Task] = []
# Deque for average send times
self._send_times: Deque[float] = deque([], maxlen=10)
# High limit defaults to 3 to start
self._send_high_limit = 3
# The subscribed message types
self._subscriptions: List[str] = []
# Wrap the WebSocket in the Serializing class
self._wrapped_ws = serializer_cls(self._websocket)
@@ -61,40 +59,58 @@ class WebSocketChannel:
def remote_addr(self):
return self._websocket.remote_addr
async def _send(self, data):
"""
Send data on the wrapped websocket
"""
await self._wrapped_ws.send(data)
@property
def avg_send_time(self):
return sum(self._send_times) / len(self._send_times)
async def send(self, data) -> bool:
def _calc_send_limit(self):
"""
Add the data to the queue to be sent.
:returns: True if data added to queue, False otherwise
Calculate the send high limit for this channel
"""
# This block only runs if the queue is full, it will wait
# until self.drain_timeout for the relay to drain the outgoing queue
# We can't use asyncio.wait_for here because the queue may have been created with a
# different eventloop
start = time.time()
while self.queue.full():
await asyncio.sleep(1)
if (time.time() - start) > self.drain_timeout:
return False
# Only update if we have enough data
if len(self._send_times) == self._send_times.maxlen:
# At least 1s or twice the average of send times, with a
# maximum of 3 seconds per message
self._send_high_limit = min(max(self.avg_send_time * 2, 1), 3)
# If for some reason the queue is still full, just return False
async def send(
self,
message: Union[WSMessageSchemaType, Dict[str, Any]],
timeout: bool = False
):
"""
Send a message on the wrapped websocket. If the sending
takes too long, it will raise a TimeoutError and
disconnect the connection.
:param message: The message to send
:param timeout: Enforce send high limit, defaults to False
"""
try:
self.queue.put_nowait(data)
except asyncio.QueueFull:
return False
_ = time.time()
# If the send times out, it will raise
# a TimeoutError and bubble up to the
# message_endpoint to close the connection
await asyncio.wait_for(
self._wrapped_ws.send(message),
timeout=self._send_high_limit if timeout else None
)
total_time = time.time() - _
self._send_times.append(total_time)
# If we got here everything is ok
return True
self._calc_send_limit()
except asyncio.TimeoutError:
logger.info(f"Connection for {self} timed out, disconnecting")
raise
# Explicitly give control back to event loop as
# websockets.send does not
await asyncio.sleep(0.01)
async def recv(self):
"""
Receive data on the wrapped websocket
Receive a message on the wrapped websocket
"""
return await self._wrapped_ws.recv()
@@ -104,18 +120,28 @@ class WebSocketChannel:
"""
return await self._websocket.ping()
async def accept(self):
"""
Accept the underlying websocket connection,
if the connection has been closed before we can
accept, just close the channel.
"""
try:
return await self._websocket.accept()
except RuntimeError:
await self.close()
async def close(self):
"""
Close the WebSocketChannel
"""
try:
await self.raw_websocket.close()
except Exception:
pass
self._closed.set()
self._relay_task.cancel()
try:
await self._websocket.close()
except RuntimeError:
pass
def is_closed(self) -> bool:
"""
@@ -139,99 +165,76 @@ class WebSocketChannel:
"""
return message_type in self._subscriptions
async def relay(self):
async def run_channel_tasks(self, *tasks, **kwargs):
"""
Relay messages from the channel's queue and send them out. This is started
as a task.
Create and await on the channel tasks unless an exception
was raised, then cancel them all.
:params *tasks: All coros or tasks to be run concurrently
:param **kwargs: Any extra kwargs to pass to gather
"""
while not self._closed.is_set():
message = await self.queue.get()
if not self.is_closed():
# Wrap the coros into tasks if they aren't already
self._channel_tasks = [
task if isinstance(task, asyncio.Task) else asyncio.create_task(task)
for task in tasks
]
try:
await self._send(message)
self.queue.task_done()
return await asyncio.gather(*self._channel_tasks, **kwargs)
except Exception:
# If an exception occurred, cancel the rest of the tasks
await self.cancel_channel_tasks()
# Limit messages per sec.
# Could cause problems with queue size if too low, and
# problems with network traffik if too high.
# 0.01 = 100/s
await asyncio.sleep(self.throttle)
except RuntimeError:
# The connection was closed, just exit the task
return
class ChannelManager:
def __init__(self):
self.channels = dict()
self._lock = RLock() # Re-entrant Lock
async def on_connect(self, websocket: WebSocketType):
async def cancel_channel_tasks(self):
"""
Wrap websocket connection into Channel and add to list
:param websocket: The WebSocket object to attach to the Channel
Cancel and wait on all channel tasks
"""
if isinstance(websocket, FastAPIWebSocket):
for task in self._channel_tasks:
task.cancel()
# Wait for tasks to finish cancelling
try:
await websocket.accept()
except RuntimeError:
# The connection was closed before we could accept it
return
await task
except (
asyncio.CancelledError,
asyncio.TimeoutError,
WebSocketDisconnect,
ConnectionClosed,
RuntimeError
):
pass
except Exception as e:
logger.info(f"Encountered unknown exception: {e}", exc_info=e)
ws_channel = WebSocketChannel(websocket)
self._channel_tasks = []
with self._lock:
self.channels[websocket] = ws_channel
return ws_channel
async def on_disconnect(self, websocket: WebSocketType):
async def __aiter__(self):
"""
Call close on the channel if it's not, and remove from channel list
Generator for received messages
"""
# We can not catch any errors here as websocket.recv is
# the first to catch any disconnects and bubble it up
# so the connection is garbage collected right away
while not self.is_closed():
yield await self.recv()
:param websocket: The WebSocket objet attached to the Channel
"""
with self._lock:
channel = self.channels.get(websocket)
if channel:
logger.info(f"Disconnecting channel {channel}")
if not channel.is_closed():
await channel.close()
del self.channels[websocket]
@asynccontextmanager
async def create_channel(
websocket: WebSocketType,
**kwargs
) -> AsyncIterator[WebSocketChannel]:
"""
Context manager for safely opening and closing a WebSocketChannel
"""
channel = WebSocketChannel(websocket, **kwargs)
try:
await channel.accept()
logger.info(f"Connected to channel - {channel}")
async def disconnect_all(self):
"""
Disconnect all Channels
"""
with self._lock:
for websocket in self.channels.copy().keys():
await self.on_disconnect(websocket)
async def broadcast(self, message: WSMessageSchemaType):
"""
Broadcast a message on all Channels
:param message: The message to send
"""
with self._lock:
for channel in self.channels.copy().values():
if channel.subscribed_to(message.get('type')):
await self.send_direct(channel, message)
async def send_direct(
self, channel: WebSocketChannel, message: Union[WSMessageSchemaType, Dict[str, Any]]):
"""
Send a message directly through direct_channel only
:param direct_channel: The WebSocketChannel object to send the message through
:param message: The message to send
"""
if not await channel.send(message):
await self.on_disconnect(channel.raw_websocket)
def has_channels(self):
"""
Flag for more than 0 channels
"""
return len(self.channels) > 0
yield channel
finally:
await channel.close()
logger.info(f"Disconnected from channel - {channel}")

31
freqtrade/rpc/api_server/ws/message_stream.py Normal file
View File

@@ -0,0 +1,31 @@
import asyncio
import time
class MessageStream:
"""
A message stream for consumers to subscribe to,
and for producers to publish to.
"""
def __init__(self):
self._loop = asyncio.get_running_loop()
self._waiter = self._loop.create_future()
def publish(self, message):
"""
Publish a message to this MessageStream
:param message: The message to publish
"""
waiter, self._waiter = self._waiter, self._loop.create_future()
waiter.set_result((message, time.time(), self._waiter))
async def __aiter__(self):
"""
Iterate over the messages in the message stream
"""
waiter = self._waiter
while True:
# Shield the future from being cancelled by a task waiting on it
message, ts, waiter = await asyncio.shield(waiter)
yield message, ts

8
freqtrade/rpc/api_server/ws/serializer.py
View File

@@ -1,5 +1,6 @@
import logging
from abc import ABC, abstractmethod
from typing import Any, Dict, Union
import orjson
import rapidjson
@@ -7,6 +8,7 @@ from pandas import DataFrame
from freqtrade.misc import dataframe_to_json, json_to_dataframe
from freqtrade.rpc.api_server.ws.proxy import WebSocketProxy
from freqtrade.rpc.api_server.ws_schemas import WSMessageSchemaType
logger = logging.getLogger(__name__)
@@ -24,17 +26,13 @@ class WebSocketSerializer(ABC):
def _deserialize(self, data):
raise NotImplementedError()
async def send(self, data: bytes):
async def send(self, data: Union[WSMessageSchemaType, Dict[str, Any]]):
await self._websocket.send(self._serialize(data))
async def recv(self) -> bytes:
data = await self._websocket.recv()
return self._deserialize(data)
async def close(self, code: int = 1000):
await self._websocket.close(code)
class HybridJSONWebSocketSerializer(WebSocketSerializer):
def _serialize(self, data) -> str:

4
freqtrade/rpc/external_message_consumer.py
View File

@@ -31,6 +31,7 @@ class Producer(TypedDict):
name: str
host: str
port: int
secure: bool
ws_token: str
@@ -180,7 +181,8 @@ class ExternalMessageConsumer:
host, port = producer['host'], producer['port']
token = producer['ws_token']
name = producer['name']
ws_url = f"ws://{host}:{port}/api/v1/message/ws?token={token}"
scheme = 'wss' if producer.get('secure', False) else 'ws'
ws_url = f"{scheme}://{host}:{port}/api/v1/message/ws?token={token}"
# This will raise InvalidURI if the url is bad
async with websockets.connect(

66
freqtrade/rpc/rpc.py
View File

@@ -218,9 +218,10 @@ class RPC:
stoploss_current_dist_pct=round(stoploss_current_dist_ratio * 100, 2),
stoploss_entry_dist=stoploss_entry_dist,
stoploss_entry_dist_ratio=round(stoploss_entry_dist_ratio, 8),
open_order='({} {} rem={:.8f})'.format(
order.order_type, order.side, order.remaining
) if order else None,
open_order=(
f'({order.order_type} {order.side} rem={order.safe_remaining:.8f})' if
order else None
),
))
results.append(trade_dict)
return results
@@ -739,6 +740,24 @@ class RPC:
self._freqtrade.wallets.update()
return {'result': f'Created sell order for trade {trade_id}.'}
def _force_entry_validations(self, pair: str, order_side: SignalDirection):
if not self._freqtrade.config.get('force_entry_enable', False):
raise RPCException('Force_entry not enabled.')
if self._freqtrade.state != State.RUNNING:
raise RPCException('trader is not running')
if order_side == SignalDirection.SHORT and self._freqtrade.trading_mode == TradingMode.SPOT:
raise RPCException("Can't go short on Spot markets.")
if pair not in self._freqtrade.exchange.get_markets(tradable_only=True):
raise RPCException('Symbol does not exist or market is not active.')
# Check if pair quote currency equals to the stake currency.
stake_currency = self._freqtrade.config.get('stake_currency')
if not self._freqtrade.exchange.get_pair_quote_currency(pair) == stake_currency:
raise RPCException(
f'Wrong pair selected. Only pairs with stake-currency {stake_currency} allowed.')
def _rpc_force_entry(self, pair: str, price: Optional[float], *,
order_type: Optional[str] = None,
order_side: SignalDirection = SignalDirection.LONG,
@@ -749,21 +768,8 @@ class RPC:
Handler for forcebuy <asset> <price>
Buys a pair trade at the given or current price
"""
self._force_entry_validations(pair, order_side)
if not self._freqtrade.config.get('force_entry_enable', False):
raise RPCException('Force_entry not enabled.')
if self._freqtrade.state != State.RUNNING:
raise RPCException('trader is not running')
if order_side == SignalDirection.SHORT and self._freqtrade.trading_mode == TradingMode.SPOT:
raise RPCException("Can't go short on Spot markets.")
# Check if pair quote currency equals to the stake currency.
stake_currency = self._freqtrade.config.get('stake_currency')
if not self._freqtrade.exchange.get_pair_quote_currency(pair) == stake_currency:
raise RPCException(
f'Wrong pair selected. Only pairs with stake-currency {stake_currency} allowed.')
# check if valid pair
# check if pair already has an open pair
@@ -773,6 +779,9 @@ class RPC:
is_short = trade.is_short
if not self._freqtrade.strategy.position_adjustment_enable:
raise RPCException(f'position for {pair} already open - id: {trade.id}')
if trade.open_order_id is not None:
raise RPCException(f'position for {pair} already open - id: {trade.id} '
f'and has open order {trade.open_order_id}')
else:
if Trade.get_open_trade_count() >= self._config['max_open_trades']:
raise RPCException("Maximum number of trades is reached.")
@@ -785,17 +794,18 @@ class RPC:
if not order_type:
order_type = self._freqtrade.strategy.order_types.get(
'force_entry', self._freqtrade.strategy.order_types['entry'])
if self._freqtrade.execute_entry(pair, stake_amount, price,
ordertype=order_type, trade=trade,
is_short=is_short,
enter_tag=enter_tag,
leverage_=leverage,
):
Trade.commit()
trade = Trade.get_trades([Trade.is_open.is_(True), Trade.pair == pair]).first()
return trade
else:
raise RPCException(f'Failed to enter position for {pair}.')
with self._freqtrade._exit_lock:
if self._freqtrade.execute_entry(pair, stake_amount, price,
ordertype=order_type, trade=trade,
is_short=is_short,
enter_tag=enter_tag,
leverage_=leverage,
):
Trade.commit()
trade = Trade.get_trades([Trade.is_open.is_(True), Trade.pair == pair]).first()
return trade
else:
raise RPCException(f'Failed to enter position for {pair}.')
def _rpc_delete(self, trade_id: int) -> Dict[str, Union[str, int]]:
"""

691
freqtrade/rpc/telegram.py
View File

@@ -79,6 +79,8 @@ def authorized_only(command_handler: Callable[..., None]) -> Callable[..., Any]:
)
try:
return command_handler(self, *args, **kwargs)
except RPCException as e:
self._send_msg(str(e))
except BaseException:
logger.exception('Exception occurred within Telegram module')
@@ -538,72 +540,67 @@ class Telegram(RPCHandler):
handler for `/status` and `/status <id>`.
"""
try:
# Check if there's at least one numerical ID provided.
# If so, try to get only these trades.
trade_ids = []
if context.args and len(context.args) > 0:
trade_ids = [int(i) for i in context.args if i.isnumeric()]
# Check if there's at least one numerical ID provided.
# If so, try to get only these trades.
trade_ids = []
if context.args and len(context.args) > 0:
trade_ids = [int(i) for i in context.args if i.isnumeric()]
results = self._rpc._rpc_trade_status(trade_ids=trade_ids)
position_adjust = self._config.get('position_adjustment_enable', False)
max_entries = self._config.get('max_entry_position_adjustment', -1)
for r in results:
r['open_date_hum'] = arrow.get(r['open_date']).humanize()
r['num_entries'] = len([o for o in r['orders'] if o['ft_is_entry']])
r['exit_reason'] = r.get('exit_reason', "")
lines = [
"*Trade ID:* `{trade_id}`" +
(" `(since {open_date_hum})`" if r['is_open'] else ""),
"*Current Pair:* {pair}",
"*Direction:* " + ("`Short`" if r.get('is_short') else "`Long`"),
"*Leverage:* `{leverage}`" if r.get('leverage') else "",
"*Amount:* `{amount} ({stake_amount} {quote_currency})`",
"*Enter Tag:* `{enter_tag}`" if r['enter_tag'] else "",
"*Exit Reason:* `{exit_reason}`" if r['exit_reason'] else "",
]
results = self._rpc._rpc_trade_status(trade_ids=trade_ids)
position_adjust = self._config.get('position_adjustment_enable', False)
max_entries = self._config.get('max_entry_position_adjustment', -1)
for r in results:
r['open_date_hum'] = arrow.get(r['open_date']).humanize()
r['num_entries'] = len([o for o in r['orders'] if o['ft_is_entry']])
r['exit_reason'] = r.get('exit_reason', "")
lines = [
"*Trade ID:* `{trade_id}`" +
(" `(since {open_date_hum})`" if r['is_open'] else ""),
"*Current Pair:* {pair}",
"*Direction:* " + ("`Short`" if r.get('is_short') else "`Long`"),
"*Leverage:* `{leverage}`" if r.get('leverage') else "",
"*Amount:* `{amount} ({stake_amount} {quote_currency})`",
"*Enter Tag:* `{enter_tag}`" if r['enter_tag'] else "",
"*Exit Reason:* `{exit_reason}`" if r['exit_reason'] else "",
]
if position_adjust:
max_buy_str = (f"/{max_entries + 1}" if (max_entries > 0) else "")
lines.append("*Number of Entries:* `{num_entries}`" + max_buy_str)
if position_adjust:
max_buy_str = (f"/{max_entries + 1}" if (max_entries > 0) else "")
lines.append("*Number of Entries:* `{num_entries}`" + max_buy_str)
lines.extend([
"*Open Rate:* `{open_rate:.8f}`",
"*Close Rate:* `{close_rate:.8f}`" if r['close_rate'] else "",
"*Open Date:* `{open_date}`",
"*Close Date:* `{close_date}`" if r['close_date'] else "",
"*Current Rate:* `{current_rate:.8f}`" if r['is_open'] else "",
("*Current Profit:* " if r['is_open'] else "*Close Profit: *")
+ "`{profit_ratio:.2%}`",
])
lines.extend([
"*Open Rate:* `{open_rate:.8f}`",
"*Close Rate:* `{close_rate:.8f}`" if r['close_rate'] else "",
"*Open Date:* `{open_date}`",
"*Close Date:* `{close_date}`" if r['close_date'] else "",
"*Current Rate:* `{current_rate:.8f}`" if r['is_open'] else "",
("*Current Profit:* " if r['is_open'] else "*Close Profit: *")
+ "`{profit_ratio:.2%}`",
])
if r['is_open']:
if r.get('realized_profit'):
lines.append("*Realized Profit:* `{realized_profit:.8f}`")
if (r['stop_loss_abs'] != r['initial_stop_loss_abs']
and r['initial_stop_loss_ratio'] is not None):
# Adding initial stoploss only if it is different from stoploss
lines.append("*Initial Stoploss:* `{initial_stop_loss_abs:.8f}` "
"`({initial_stop_loss_ratio:.2%})`")
if r['is_open']:
if r.get('realized_profit'):
lines.append("*Realized Profit:* `{realized_profit:.8f}`")
if (r['stop_loss_abs'] != r['initial_stop_loss_abs']
and r['initial_stop_loss_ratio'] is not None):
# Adding initial stoploss only if it is different from stoploss
lines.append("*Initial Stoploss:* `{initial_stop_loss_abs:.8f}` "
"`({initial_stop_loss_ratio:.2%})`")
# Adding stoploss and stoploss percentage only if it is not None
lines.append("*Stoploss:* `{stop_loss_abs:.8f}` " +
("`({stop_loss_ratio:.2%})`" if r['stop_loss_ratio'] else ""))
lines.append("*Stoploss distance:* `{stoploss_current_dist:.8f}` "
"`({stoploss_current_dist_ratio:.2%})`")
if r['open_order']:
lines.append(
"*Open Order:* `{open_order}`"
+ "- `{exit_order_status}`" if r['exit_order_status'] else "")
# Adding stoploss and stoploss percentage only if it is not None
lines.append("*Stoploss:* `{stop_loss_abs:.8f}` " +
("`({stop_loss_ratio:.2%})`" if r['stop_loss_ratio'] else ""))
lines.append("*Stoploss distance:* `{stoploss_current_dist:.8f}` "
"`({stoploss_current_dist_ratio:.2%})`")
if r['open_order']:
lines.append(
"*Open Order:* `{open_order}`"
+ "- `{exit_order_status}`" if r['exit_order_status'] else "")
lines_detail = self._prepare_order_details(
r['orders'], r['quote_currency'], r['is_open'])
lines.extend(lines_detail if lines_detail else "")
self.__send_status_msg(lines, r)
except RPCException as e:
self._send_msg(str(e))
lines_detail = self._prepare_order_details(
r['orders'], r['quote_currency'], r['is_open'])
lines.extend(lines_detail if lines_detail else "")
self.__send_status_msg(lines, r)
def __send_status_msg(self, lines: List[str], r: Dict[str, Any]) -> None:
"""
@@ -630,37 +627,34 @@ class Telegram(RPCHandler):
:param update: message update
:return: None
"""
try:
fiat_currency = self._config.get('fiat_display_currency', '')
statlist, head, fiat_profit_sum = self._rpc._rpc_status_table(
self._config['stake_currency'], fiat_currency)
fiat_currency = self._config.get('fiat_display_currency', '')
statlist, head, fiat_profit_sum = self._rpc._rpc_status_table(
self._config['stake_currency'], fiat_currency)
show_total = not isnan(fiat_profit_sum) and len(statlist) > 1
max_trades_per_msg = 50
"""
Calculate the number of messages of 50 trades per message
0.99 is used to make sure that there are no extra (empty) messages
As an example with 50 trades, there will be int(50/50 + 0.99) = 1 message
"""
messages_count = max(int(len(statlist) / max_trades_per_msg + 0.99), 1)
for i in range(0, messages_count):
trades = statlist[i * max_trades_per_msg:(i + 1) * max_trades_per_msg]
if show_total and i == messages_count - 1:
# append total line
trades.append(["Total", "", "", f"{fiat_profit_sum:.2f} {fiat_currency}"])
show_total = not isnan(fiat_profit_sum) and len(statlist) > 1
max_trades_per_msg = 50
"""
Calculate the number of messages of 50 trades per message
0.99 is used to make sure that there are no extra (empty) messages
As an example with 50 trades, there will be int(50/50 + 0.99) = 1 message
"""
messages_count = max(int(len(statlist) / max_trades_per_msg + 0.99), 1)
for i in range(0, messages_count):
trades = statlist[i * max_trades_per_msg:(i + 1) * max_trades_per_msg]
if show_total and i == messages_count - 1:
# append total line
trades.append(["Total", "", "", f"{fiat_profit_sum:.2f} {fiat_currency}"])
message = tabulate(trades,
headers=head,
tablefmt='simple')
if show_total and i == messages_count - 1:
# insert separators line between Total
lines = message.split("\n")
message = "\n".join(lines[:-1] + [lines[1]] + [lines[-1]])
self._send_msg(f"<pre>{message}</pre>", parse_mode=ParseMode.HTML,
reload_able=True, callback_path="update_status_table",
query=update.callback_query)
except RPCException as e:
self._send_msg(str(e))
message = tabulate(trades,
headers=head,
tablefmt='simple')
if show_total and i == messages_count - 1:
# insert separators line between Total
lines = message.split("\n")
message = "\n".join(lines[:-1] + [lines[1]] + [lines[-1]])
self._send_msg(f"<pre>{message}</pre>", parse_mode=ParseMode.HTML,
reload_able=True, callback_path="update_status_table",
query=update.callback_query)
@authorized_only
def _timeunit_stats(self, update: Update, context: CallbackContext, unit: str) -> None:
@@ -686,35 +680,32 @@ class Telegram(RPCHandler):
timescale = int(context.args[0]) if context.args else val.default
except (TypeError, ValueError, IndexError):
timescale = val.default
try:
stats = self._rpc._rpc_timeunit_profit(
timescale,
stake_cur,
fiat_disp_cur,
unit
)
stats_tab = tabulate(
[[f"{period['date']} ({period['trade_count']})",
f"{round_coin_value(period['abs_profit'], stats['stake_currency'])}",
f"{period['fiat_value']:.2f} {stats['fiat_display_currency']}",
f"{period['rel_profit']:.2%}",
] for period in stats['data']],
headers=[
f"{val.header} (count)",
f'{stake_cur}',
f'{fiat_disp_cur}',
'Profit %',
'Trades',
],
tablefmt='simple')
message = (
f'<b>{val.message} Profit over the last {timescale} {val.message2}</b>:\n'
f'<pre>{stats_tab}</pre>'
)
self._send_msg(message, parse_mode=ParseMode.HTML, reload_able=True,
callback_path=val.callback, query=update.callback_query)
except RPCException as e:
self._send_msg(str(e))
stats = self._rpc._rpc_timeunit_profit(
timescale,
stake_cur,
fiat_disp_cur,
unit
)
stats_tab = tabulate(
[[f"{period['date']} ({period['trade_count']})",
f"{round_coin_value(period['abs_profit'], stats['stake_currency'])}",
f"{period['fiat_value']:.2f} {stats['fiat_display_currency']}",
f"{period['rel_profit']:.2%}",
] for period in stats['data']],
headers=[
f"{val.header} (count)",
f'{stake_cur}',
f'{fiat_disp_cur}',
'Profit %',
'Trades',
],
tablefmt='simple')
message = (
f'<b>{val.message} Profit over the last {timescale} {val.message2}</b>:\n'
f'<pre>{stats_tab}</pre>'
)
self._send_msg(message, parse_mode=ParseMode.HTML, reload_able=True,
callback_path=val.callback, query=update.callback_query)
@authorized_only
def _daily(self, update: Update, context: CallbackContext) -> None:
@@ -878,79 +869,76 @@ class Telegram(RPCHandler):
@authorized_only
def _balance(self, update: Update, context: CallbackContext) -> None:
""" Handler for /balance """
try:
result = self._rpc._rpc_balance(self._config['stake_currency'],
self._config.get('fiat_display_currency', ''))
result = self._rpc._rpc_balance(self._config['stake_currency'],
self._config.get('fiat_display_currency', ''))
balance_dust_level = self._config['telegram'].get('balance_dust_level', 0.0)
if not balance_dust_level:
balance_dust_level = DUST_PER_COIN.get(self._config['stake_currency'], 1.0)
balance_dust_level = self._config['telegram'].get('balance_dust_level', 0.0)
if not balance_dust_level:
balance_dust_level = DUST_PER_COIN.get(self._config['stake_currency'], 1.0)
output = ''
if self._config['dry_run']:
output += "*Warning:* Simulated balances in Dry Mode.\n"
starting_cap = round_coin_value(
result['starting_capital'], self._config['stake_currency'])
output += f"Starting capital: `{starting_cap}`"
starting_cap_fiat = round_coin_value(
result['starting_capital_fiat'], self._config['fiat_display_currency']
) if result['starting_capital_fiat'] > 0 else ''
output += (f" `, {starting_cap_fiat}`.\n"
) if result['starting_capital_fiat'] > 0 else '.\n'
output = ''
if self._config['dry_run']:
output += "*Warning:* Simulated balances in Dry Mode.\n"
starting_cap = round_coin_value(
result['starting_capital'], self._config['stake_currency'])
output += f"Starting capital: `{starting_cap}`"
starting_cap_fiat = round_coin_value(
result['starting_capital_fiat'], self._config['fiat_display_currency']
) if result['starting_capital_fiat'] > 0 else ''
output += (f" `, {starting_cap_fiat}`.\n"
) if result['starting_capital_fiat'] > 0 else '.\n'
total_dust_balance = 0
total_dust_currencies = 0
for curr in result['currencies']:
curr_output = ''
if curr['est_stake'] > balance_dust_level:
if curr['is_position']:
curr_output = (
f"*{curr['currency']}:*\n"
f"\t`{curr['side']}: {curr['position']:.8f}`\n"
f"\t`Leverage: {curr['leverage']:.1f}`\n"
f"\t`Est. {curr['stake']}: "
f"{round_coin_value(curr['est_stake'], curr['stake'], False)}`\n")
else:
curr_output = (
f"*{curr['currency']}:*\n"
f"\t`Available: {curr['free']:.8f}`\n"
f"\t`Balance: {curr['balance']:.8f}`\n"
f"\t`Pending: {curr['used']:.8f}`\n"
f"\t`Est. {curr['stake']}: "
f"{round_coin_value(curr['est_stake'], curr['stake'], False)}`\n")
elif curr['est_stake'] <= balance_dust_level:
total_dust_balance += curr['est_stake']
total_dust_currencies += 1
# Handle overflowing message length
if len(output + curr_output) >= MAX_MESSAGE_LENGTH:
self._send_msg(output)
output = curr_output
total_dust_balance = 0
total_dust_currencies = 0
for curr in result['currencies']:
curr_output = ''
if curr['est_stake'] > balance_dust_level:
if curr['is_position']:
curr_output = (
f"*{curr['currency']}:*\n"
f"\t`{curr['side']}: {curr['position']:.8f}`\n"
f"\t`Leverage: {curr['leverage']:.1f}`\n"
f"\t`Est. {curr['stake']}: "
f"{round_coin_value(curr['est_stake'], curr['stake'], False)}`\n")
else:
output += curr_output
curr_output = (
f"*{curr['currency']}:*\n"
f"\t`Available: {curr['free']:.8f}`\n"
f"\t`Balance: {curr['balance']:.8f}`\n"
f"\t`Pending: {curr['used']:.8f}`\n"
f"\t`Est. {curr['stake']}: "
f"{round_coin_value(curr['est_stake'], curr['stake'], False)}`\n")
elif curr['est_stake'] <= balance_dust_level:
total_dust_balance += curr['est_stake']
total_dust_currencies += 1
if total_dust_balance > 0:
output += (
f"*{total_dust_currencies} Other "
f"{plural(total_dust_currencies, 'Currency', 'Currencies')} "
f"(< {balance_dust_level} {result['stake']}):*\n"
f"\t`Est. {result['stake']}: "
f"{round_coin_value(total_dust_balance, result['stake'], False)}`\n")
tc = result['trade_count'] > 0
stake_improve = f" `({result['starting_capital_ratio']:.2%})`" if tc else ''
fiat_val = f" `({result['starting_capital_fiat_ratio']:.2%})`" if tc else ''
# Handle overflowing message length
if len(output + curr_output) >= MAX_MESSAGE_LENGTH:
self._send_msg(output)
output = curr_output
else:
output += curr_output
output += ("\n*Estimated Value*:\n"
f"\t`{result['stake']}: "
f"{round_coin_value(result['total'], result['stake'], False)}`"
f"{stake_improve}\n"
f"\t`{result['symbol']}: "
f"{round_coin_value(result['value'], result['symbol'], False)}`"
f"{fiat_val}\n")
self._send_msg(output, reload_able=True, callback_path="update_balance",
query=update.callback_query)
except RPCException as e:
self._send_msg(str(e))
if total_dust_balance > 0:
output += (
f"*{total_dust_currencies} Other "
f"{plural(total_dust_currencies, 'Currency', 'Currencies')} "
f"(< {balance_dust_level} {result['stake']}):*\n"
f"\t`Est. {result['stake']}: "
f"{round_coin_value(total_dust_balance, result['stake'], False)}`\n")
tc = result['trade_count'] > 0
stake_improve = f" `({result['starting_capital_ratio']:.2%})`" if tc else ''
fiat_val = f" `({result['starting_capital_fiat_ratio']:.2%})`" if tc else ''
output += ("\n*Estimated Value*:\n"
f"\t`{result['stake']}: "
f"{round_coin_value(result['total'], result['stake'], False)}`"
f"{stake_improve}\n"
f"\t`{result['symbol']}: "
f"{round_coin_value(result['value'], result['symbol'], False)}`"
f"{fiat_val}\n")
self._send_msg(output, reload_able=True, callback_path="update_balance",
query=update.callback_query)
@authorized_only
def _start(self, update: Update, context: CallbackContext) -> None:
@@ -1125,26 +1113,23 @@ class Telegram(RPCHandler):
nrecent = int(context.args[0]) if context.args else 10
except (TypeError, ValueError, IndexError):
nrecent = 10
try:
trades = self._rpc._rpc_trade_history(
nrecent
)
trades_tab = tabulate(
[[arrow.get(trade['close_date']).humanize(),
trade['pair'] + " (#" + str(trade['trade_id']) + ")",
f"{(trade['close_profit']):.2%} ({trade['close_profit_abs']})"]
for trade in trades['trades']],
headers=[
'Close Date',
'Pair (ID)',
f'Profit ({stake_cur})',
],
tablefmt='simple')
message = (f"<b>{min(trades['trades_count'], nrecent)} recent trades</b>:\n"
+ (f"<pre>{trades_tab}</pre>" if trades['trades_count'] > 0 else ''))
self._send_msg(message, parse_mode=ParseMode.HTML)
except RPCException as e:
self._send_msg(str(e))
trades = self._rpc._rpc_trade_history(
nrecent
)
trades_tab = tabulate(
[[arrow.get(trade['close_date']).humanize(),
trade['pair'] + " (#" + str(trade['trade_id']) + ")",
f"{(trade['close_profit']):.2%} ({trade['close_profit_abs']})"]
for trade in trades['trades']],
headers=[
'Close Date',
'Pair (ID)',
f'Profit ({stake_cur})',
],
tablefmt='simple')
message = (f"<b>{min(trades['trades_count'], nrecent)} recent trades</b>:\n"
+ (f"<pre>{trades_tab}</pre>" if trades['trades_count'] > 0 else ''))
self._send_msg(message, parse_mode=ParseMode.HTML)
@authorized_only
def _delete_trade(self, update: Update, context: CallbackContext) -> None:
@@ -1155,18 +1140,14 @@ class Telegram(RPCHandler):
:param update: message update
:return: None
"""
try:
if not context.args or len(context.args) == 0:
raise RPCException("Trade-id not set.")
trade_id = int(context.args[0])
msg = self._rpc._rpc_delete(trade_id)
self._send_msg((
f"`{msg['result_msg']}`\n"
'Please make sure to take care of this asset on the exchange manually.'
))
except RPCException as e:
self._send_msg(str(e))
if not context.args or len(context.args) == 0:
raise RPCException("Trade-id not set.")
trade_id = int(context.args[0])
msg = self._rpc._rpc_delete(trade_id)
self._send_msg((
f"`{msg['result_msg']}`\n"
'Please make sure to take care of this asset on the exchange manually.'
))
@authorized_only
def _performance(self, update: Update, context: CallbackContext) -> None:
@@ -1177,27 +1158,24 @@ class Telegram(RPCHandler):
:param update: message update
:return: None
"""
try:
trades = self._rpc._rpc_performance()
output = "<b>Performance:</b>\n"
for i, trade in enumerate(trades):
stat_line = (
f"{i+1}.\t <code>{trade['pair']}\t"
f"{round_coin_value(trade['profit_abs'], self._config['stake_currency'])} "
f"({trade['profit_ratio']:.2%}) "
f"({trade['count']})</code>\n")
trades = self._rpc._rpc_performance()
output = "<b>Performance:</b>\n"
for i, trade in enumerate(trades):
stat_line = (
f"{i+1}.\t <code>{trade['pair']}\t"
f"{round_coin_value(trade['profit_abs'], self._config['stake_currency'])} "
f"({trade['profit_ratio']:.2%}) "
f"({trade['count']})</code>\n")
if len(output + stat_line) >= MAX_MESSAGE_LENGTH:
self._send_msg(output, parse_mode=ParseMode.HTML)
output = stat_line
else:
output += stat_line
if len(output + stat_line) >= MAX_MESSAGE_LENGTH:
self._send_msg(output, parse_mode=ParseMode.HTML)
output = stat_line
else:
output += stat_line
self._send_msg(output, parse_mode=ParseMode.HTML,
reload_able=True, callback_path="update_performance",
query=update.callback_query)
except RPCException as e:
self._send_msg(str(e))
self._send_msg(output, parse_mode=ParseMode.HTML,
reload_able=True, callback_path="update_performance",
query=update.callback_query)
@authorized_only
def _enter_tag_performance(self, update: Update, context: CallbackContext) -> None:
@@ -1208,31 +1186,28 @@ class Telegram(RPCHandler):
:param update: message update
:return: None
"""
try:
pair = None
if context.args and isinstance(context.args[0], str):
pair = context.args[0]
pair = None
if context.args and isinstance(context.args[0], str):
pair = context.args[0]
trades = self._rpc._rpc_enter_tag_performance(pair)
output = "<b>Entry Tag Performance:</b>\n"
for i, trade in enumerate(trades):
stat_line = (
f"{i+1}.\t <code>{trade['enter_tag']}\t"
f"{round_coin_value(trade['profit_abs'], self._config['stake_currency'])} "
f"({trade['profit_ratio']:.2%}) "
f"({trade['count']})</code>\n")
trades = self._rpc._rpc_enter_tag_performance(pair)
output = "<b>Entry Tag Performance:</b>\n"
for i, trade in enumerate(trades):
stat_line = (
f"{i+1}.\t <code>{trade['enter_tag']}\t"
f"{round_coin_value(trade['profit_abs'], self._config['stake_currency'])} "
f"({trade['profit_ratio']:.2%}) "
f"({trade['count']})</code>\n")
if len(output + stat_line) >= MAX_MESSAGE_LENGTH:
self._send_msg(output, parse_mode=ParseMode.HTML)
output = stat_line
else:
output += stat_line
if len(output + stat_line) >= MAX_MESSAGE_LENGTH:
self._send_msg(output, parse_mode=ParseMode.HTML)
output = stat_line
else:
output += stat_line
self._send_msg(output, parse_mode=ParseMode.HTML,
reload_able=True, callback_path="update_enter_tag_performance",
query=update.callback_query)
except RPCException as e:
self._send_msg(str(e))
self._send_msg(output, parse_mode=ParseMode.HTML,
reload_able=True, callback_path="update_enter_tag_performance",
query=update.callback_query)
@authorized_only
def _exit_reason_performance(self, update: Update, context: CallbackContext) -> None:
@@ -1243,31 +1218,28 @@ class Telegram(RPCHandler):
:param update: message update
:return: None
"""
try:
pair = None
if context.args and isinstance(context.args[0], str):
pair = context.args[0]
pair = None
if context.args and isinstance(context.args[0], str):
pair = context.args[0]
trades = self._rpc._rpc_exit_reason_performance(pair)
output = "<b>Exit Reason Performance:</b>\n"
for i, trade in enumerate(trades):
stat_line = (
f"{i+1}.\t <code>{trade['exit_reason']}\t"
f"{round_coin_value(trade['profit_abs'], self._config['stake_currency'])} "
f"({trade['profit_ratio']:.2%}) "
f"({trade['count']})</code>\n")
trades = self._rpc._rpc_exit_reason_performance(pair)
output = "<b>Exit Reason Performance:</b>\n"
for i, trade in enumerate(trades):
stat_line = (
f"{i+1}.\t <code>{trade['exit_reason']}\t"
f"{round_coin_value(trade['profit_abs'], self._config['stake_currency'])} "
f"({trade['profit_ratio']:.2%}) "
f"({trade['count']})</code>\n")
if len(output + stat_line) >= MAX_MESSAGE_LENGTH:
self._send_msg(output, parse_mode=ParseMode.HTML)
output = stat_line
else:
output += stat_line
if len(output + stat_line) >= MAX_MESSAGE_LENGTH:
self._send_msg(output, parse_mode=ParseMode.HTML)
output = stat_line
else:
output += stat_line
self._send_msg(output, parse_mode=ParseMode.HTML,
reload_able=True, callback_path="update_exit_reason_performance",
query=update.callback_query)
except RPCException as e:
self._send_msg(str(e))
self._send_msg(output, parse_mode=ParseMode.HTML,
reload_able=True, callback_path="update_exit_reason_performance",
query=update.callback_query)
@authorized_only
def _mix_tag_performance(self, update: Update, context: CallbackContext) -> None:
@@ -1278,31 +1250,28 @@ class Telegram(RPCHandler):
:param update: message update
:return: None
"""
try:
pair = None
if context.args and isinstance(context.args[0], str):
pair = context.args[0]
pair = None
if context.args and isinstance(context.args[0], str):
pair = context.args[0]
trades = self._rpc._rpc_mix_tag_performance(pair)
output = "<b>Mix Tag Performance:</b>\n"
for i, trade in enumerate(trades):
stat_line = (
f"{i+1}.\t <code>{trade['mix_tag']}\t"
f"{round_coin_value(trade['profit_abs'], self._config['stake_currency'])} "
f"({trade['profit']:.2%}) "
f"({trade['count']})</code>\n")
trades = self._rpc._rpc_mix_tag_performance(pair)
output = "<b>Mix Tag Performance:</b>\n"
for i, trade in enumerate(trades):
stat_line = (
f"{i+1}.\t <code>{trade['mix_tag']}\t"
f"{round_coin_value(trade['profit_abs'], self._config['stake_currency'])} "
f"({trade['profit']:.2%}) "
f"({trade['count']})</code>\n")
if len(output + stat_line) >= MAX_MESSAGE_LENGTH:
self._send_msg(output, parse_mode=ParseMode.HTML)
output = stat_line
else:
output += stat_line
if len(output + stat_line) >= MAX_MESSAGE_LENGTH:
self._send_msg(output, parse_mode=ParseMode.HTML)
output = stat_line
else:
output += stat_line
self._send_msg(output, parse_mode=ParseMode.HTML,
reload_able=True, callback_path="update_mix_tag_performance",
query=update.callback_query)
except RPCException as e:
self._send_msg(str(e))
self._send_msg(output, parse_mode=ParseMode.HTML,
reload_able=True, callback_path="update_mix_tag_performance",
query=update.callback_query)
@authorized_only
def _count(self, update: Update, context: CallbackContext) -> None:
@@ -1313,18 +1282,15 @@ class Telegram(RPCHandler):
:param update: message update
:return: None
"""
try:
counts = self._rpc._rpc_count()
message = tabulate({k: [v] for k, v in counts.items()},
headers=['current', 'max', 'total stake'],
tablefmt='simple')
message = "<pre>{}</pre>".format(message)
logger.debug(message)
self._send_msg(message, parse_mode=ParseMode.HTML,
reload_able=True, callback_path="update_count",
query=update.callback_query)
except RPCException as e:
self._send_msg(str(e))
counts = self._rpc._rpc_count()
message = tabulate({k: [v] for k, v in counts.items()},
headers=['current', 'max', 'total stake'],
tablefmt='simple')
message = "<pre>{}</pre>".format(message)
logger.debug(message)
self._send_msg(message, parse_mode=ParseMode.HTML,
reload_able=True, callback_path="update_count",
query=update.callback_query)
@authorized_only
def _locks(self, update: Update, context: CallbackContext) -> None:
@@ -1372,22 +1338,19 @@ class Telegram(RPCHandler):
Handler for /whitelist
Shows the currently active whitelist
"""
try:
whitelist = self._rpc._rpc_whitelist()
whitelist = self._rpc._rpc_whitelist()
if context.args:
if "sorted" in context.args:
whitelist['whitelist'] = sorted(whitelist['whitelist'])
if "baseonly" in context.args:
whitelist['whitelist'] = [pair.split("/")[0] for pair in whitelist['whitelist']]
if context.args:
if "sorted" in context.args:
whitelist['whitelist'] = sorted(whitelist['whitelist'])
if "baseonly" in context.args:
whitelist['whitelist'] = [pair.split("/")[0] for pair in whitelist['whitelist']]
message = f"Using whitelist `{whitelist['method']}` with {whitelist['length']} pairs\n"
message += f"`{', '.join(whitelist['whitelist'])}`"
message = f"Using whitelist `{whitelist['method']}` with {whitelist['length']} pairs\n"
message += f"`{', '.join(whitelist['whitelist'])}`"
logger.debug(message)
self._send_msg(message)
except RPCException as e:
self._send_msg(str(e))
logger.debug(message)
self._send_msg(message)
@authorized_only
def _blacklist(self, update: Update, context: CallbackContext) -> None:
@@ -1425,30 +1388,27 @@ class Telegram(RPCHandler):
Shows the latest logs
"""
try:
try:
limit = int(context.args[0]) if context.args else 10
except (TypeError, ValueError, IndexError):
limit = 10
logs = RPC._rpc_get_logs(limit)['logs']
msgs = ''
msg_template = "*{}* {}: {} \\- `{}`"
for logrec in logs:
msg = msg_template.format(escape_markdown(logrec[0], version=2),
escape_markdown(logrec[2], version=2),
escape_markdown(logrec[3], version=2),
escape_markdown(logrec[4], version=2))
if len(msgs + msg) + 10 >= MAX_MESSAGE_LENGTH:
# Send message immediately if it would become too long
self._send_msg(msgs, parse_mode=ParseMode.MARKDOWN_V2)
msgs = msg + '\n'
else:
# Append message to messages to send
msgs += msg + '\n'
if msgs:
limit = int(context.args[0]) if context.args else 10
except (TypeError, ValueError, IndexError):
limit = 10
logs = RPC._rpc_get_logs(limit)['logs']
msgs = ''
msg_template = "*{}* {}: {} \\- `{}`"
for logrec in logs:
msg = msg_template.format(escape_markdown(logrec[0], version=2),
escape_markdown(logrec[2], version=2),
escape_markdown(logrec[3], version=2),
escape_markdown(logrec[4], version=2))
if len(msgs + msg) + 10 >= MAX_MESSAGE_LENGTH:
# Send message immediately if it would become too long
self._send_msg(msgs, parse_mode=ParseMode.MARKDOWN_V2)
except RPCException as e:
self._send_msg(str(e))
msgs = msg + '\n'
else:
# Append message to messages to send
msgs += msg + '\n'
if msgs:
self._send_msg(msgs, parse_mode=ParseMode.MARKDOWN_V2)
@authorized_only
def _edge(self, update: Update, context: CallbackContext) -> None:
@@ -1456,21 +1416,17 @@ class Telegram(RPCHandler):
Handler for /edge
Shows information related to Edge
"""
try:
edge_pairs = self._rpc._rpc_edge()
if not edge_pairs:
message = '<b>Edge only validated following pairs:</b>'
self._send_msg(message, parse_mode=ParseMode.HTML)
edge_pairs = self._rpc._rpc_edge()
if not edge_pairs:
message = '<b>Edge only validated following pairs:</b>'
self._send_msg(message, parse_mode=ParseMode.HTML)
for chunk in chunks(edge_pairs, 25):
edge_pairs_tab = tabulate(chunk, headers='keys', tablefmt='simple')
message = (f'<b>Edge only validated following pairs:</b>\n'
f'<pre>{edge_pairs_tab}</pre>')
for chunk in chunks(edge_pairs, 25):
edge_pairs_tab = tabulate(chunk, headers='keys', tablefmt='simple')
message = (f'<b>Edge only validated following pairs:</b>\n'
f'<pre>{edge_pairs_tab}</pre>')
self._send_msg(message, parse_mode=ParseMode.HTML)
except RPCException as e:
self._send_msg(str(e))
self._send_msg(message, parse_mode=ParseMode.HTML)
@authorized_only
def _help(self, update: Update, context: CallbackContext) -> None:
@@ -1551,12 +1507,9 @@ class Telegram(RPCHandler):
Handler for /health
Shows the last process timestamp
"""
try:
health = self._rpc._health()
message = f"Last process: `{health['last_process_loc']}`"
self._send_msg(message)
except RPCException as e:
self._send_msg(str(e))
health = self._rpc._health()
message = f"Last process: `{health['last_process_loc']}`"
self._send_msg(message)
@authorized_only
def _version(self, update: Update, context: CallbackContext) -> None:

4
freqtrade/templates/FreqaiExampleHybridStrategy.py
View File

@@ -19,7 +19,7 @@ class FreqaiExampleHybridStrategy(IStrategy):
Launching this strategy would be:
freqtrade trade --strategy FreqaiExampleHyridStrategy --strategy-path freqtrade/templates
freqtrade trade --strategy FreqaiExampleHybridStrategy --strategy-path freqtrade/templates
--freqaimodel CatboostClassifier --config config_examples/config_freqai.example.json
or the user simply adds this to their config:
@@ -86,7 +86,7 @@ class FreqaiExampleHybridStrategy(IStrategy):
process_only_new_candles = True
stoploss = -0.05
use_exit_signal = True
startup_candle_count: int = 300
startup_candle_count: int = 30
can_short = True
# Hyperoptable parameters

2
freqtrade/templates/strategy_analysis_example.ipynb
View File

@@ -328,7 +328,7 @@
"# Show graph inline\n",
"# graph.show()\n",
"\n",
"# Render graph in a seperate window\n",
"# Render graph in a separate window\n",
"graph.show(renderer=\"browser\")\n"
]
},