mirror of
https://github.com/freqtrade/freqtrade.git
synced 2024-11-10 02:12:01 +00:00
85 lines
2.3 KiB
Python
85 lines
2.3 KiB
Python
from datetime import timedelta
|
|
from typing import NamedTuple, Optional
|
|
|
|
from pandas import DataFrame
|
|
|
|
from freqtrade.enums import ExitType
|
|
from freqtrade.exchange import timeframe_to_minutes
|
|
from freqtrade.util.datetime_helpers import dt_utc
|
|
|
|
|
|
tests_start_time = dt_utc(2018, 10, 3)
|
|
tests_timeframe = "1h"
|
|
|
|
|
|
class BTrade(NamedTuple):
|
|
"""
|
|
Minimalistic Trade result used for functional backtesting
|
|
"""
|
|
|
|
exit_reason: ExitType
|
|
open_tick: int
|
|
close_tick: int
|
|
enter_tag: Optional[str] = None
|
|
is_short: bool = False
|
|
|
|
|
|
class BTContainer(NamedTuple):
|
|
"""
|
|
Minimal BacktestContainer defining Backtest inputs and results.
|
|
"""
|
|
|
|
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
|
|
use_exit_signal: bool = False
|
|
use_custom_stoploss: bool = False
|
|
custom_entry_price: Optional[float] = None
|
|
custom_exit_price: Optional[float] = None
|
|
leverage: float = 1.0
|
|
timeout: Optional[int] = None
|
|
adjust_entry_price: Optional[float] = None
|
|
|
|
|
|
def _get_frame_time_from_offset(offset):
|
|
minutes = offset * timeframe_to_minutes(tests_timeframe)
|
|
return tests_start_time + timedelta(minutes=minutes)
|
|
|
|
|
|
def _build_backtest_dataframe(data):
|
|
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]
|
|
columns = columns + ["enter_tag"] if len(data[0]) == 11 else columns
|
|
|
|
frame = DataFrame.from_records(data, columns=columns)
|
|
frame["date"] = frame["date"].apply(_get_frame_time_from_offset)
|
|
# Ensure floats are in place
|
|
for column in ["open", "high", "low", "close", "volume"]:
|
|
frame[column] = frame[column].astype("float64")
|
|
|
|
# Ensure all candles make kindof sense
|
|
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
|