freqtrade_origin/tests/optimize/__init__.py

85 lines
2.3 KiB
Python
Raw Normal View History

2023-05-14 15:46:56 +00:00
from datetime import timedelta
from typing import Dict, List, NamedTuple, Optional
from pandas import DataFrame
2022-03-25 05:55:37 +00:00
from freqtrade.enums import ExitType
from freqtrade.exchange import timeframe_to_minutes
2023-05-14 15:46:56 +00:00
from freqtrade.util.datetime_helpers import dt_utc
2020-09-28 17:43:15 +00:00
2023-05-14 15:46:56 +00:00
tests_start_time = dt_utc(2018, 10, 3)
2024-05-12 13:29:14 +00:00
tests_timeframe = "1h"
class BTrade(NamedTuple):
"""
Minimalistic Trade result used for functional backtesting
"""
2024-05-12 13:29:14 +00:00
2022-03-24 19:33:47 +00:00
exit_reason: ExitType
open_tick: int
close_tick: int
2021-09-26 13:20:59 +00:00
enter_tag: Optional[str] = None
2022-03-16 18:26:08 +00:00
is_short: bool = False
class BTContainer(NamedTuple):
"""
Minimal BacktestContainer defining Backtest inputs and results.
"""
2024-05-12 13:29:14 +00:00
2020-02-10 09:35:48 +00:00
data: List[List[float]]
stop_loss: float
roi: Dict[str, float]
trades: List[BTrade]
profit_perc: float
trailing_stop: bool = False
trailing_only_offset_is_reached: bool = False
trailing_stop_positive: Optional[float] = None
trailing_stop_positive_offset: float = 0.0
2022-04-03 17:27:30 +00:00
use_exit_signal: bool = False
use_custom_stoploss: bool = False
2022-02-05 14:12:29 +00:00
custom_entry_price: Optional[float] = None
2022-02-05 14:20:05 +00:00
custom_exit_price: Optional[float] = None
2021-11-30 19:42:18 +00:00
leverage: float = 1.0
2022-05-04 05:13:02 +00:00
timeout: Optional[int] = None
adjust_entry_price: Optional[float] = None
def _get_frame_time_from_offset(offset):
minutes = offset * timeframe_to_minutes(tests_timeframe)
2023-05-14 15:46:56 +00:00
return tests_start_time + timedelta(minutes=minutes)
def _build_backtest_dataframe(data):
2024-05-12 13:29:14 +00:00
columns = [
"date",
"open",
"high",
"low",
"close",
"volume",
"enter_long",
"exit_long",
"enter_short",
"exit_short",
]
if len(data[0]) == 8:
# No short columns
data = [d + [0, 0] for d in data]
2024-05-12 13:29:14 +00:00
columns = columns + ["enter_tag"] if len(data[0]) == 11 else columns
frame = DataFrame.from_records(data, columns=columns)
2024-05-12 13:29:14 +00:00
frame["date"] = frame["date"].apply(_get_frame_time_from_offset)
# Ensure floats are in place
2024-05-12 13:29:14 +00:00
for column in ["open", "high", "low", "close", "volume"]:
frame[column] = frame[column].astype("float64")
2021-10-30 14:10:28 +00:00
# Ensure all candles make kindof sense
2024-05-12 13:29:14 +00:00
assert all(frame["low"] <= frame["close"])
assert all(frame["low"] <= frame["open"])
assert all(frame["high"] >= frame["close"])
assert all(frame["high"] >= frame["open"])
return frame