Merge branch 'fixHyperoptFreqai' of https://github.com/wagnercosta/freqtrade into fixHyperoptFreqai

This commit is contained in:
Wagner Costa Santos 2022-09-06 15:43:08 -03:00
commit e0490b3efc
9 changed files with 84 additions and 33 deletions

Binary file not shown.

Before

Width:  |  Height:  |  Size: 191 KiB

After

Width:  |  Height:  |  Size: 185 KiB

View File

@ -113,14 +113,14 @@ Mandatory parameters are marked as **Required**, which means that they are requi
| `use_SVM_to_remove_outliers` | Train a support vector machine to detect and remove outliers from the training data set, as well as from incoming data points. See details about how it works [here](#removing-outliers-using-a-support-vector-machine-svm). <br> **Datatype:** Boolean.
| `svm_params` | All parameters available in Sklearn's `SGDOneClassSVM()`. See details about some select parameters [here](#removing-outliers-using-a-support-vector-machine-svm). <br> **Datatype:** Dictionary.
| `use_DBSCAN_to_remove_outliers` | Cluster data using DBSCAN to identify and remove outliers from training and prediction data. See details about how it works [here](#removing-outliers-with-dbscan). <br> **Datatype:** Boolean.
| `outlier_protection_percentage` | If more than `outlier_protection_percentage` fraction of points are removed as outliers, FreqAI will log a warning message and ignore outlier detection while keeping the original dataset intact. <br> **Datatype:** float. Default: `30`
| `reverse_train_test_order` | If true, FreqAI will train on the latest data split and test on historical split of the data. This allows the model to be trained up to the most recent data point, while avoiding overfitting. However, users should be careful to understand unorthodox nature of this parameter before employing it. <br> **Datatype:** bool. Default: False
| `outlier_protection_percentage` | If more than `outlier_protection_percentage` % of points are detected as outliers by the SVM or DBSCAN, FreqAI will log a warning message and ignore outlier detection while keeping the original dataset intact. If the outlier protection is triggered, no predictions will be made based on the training data. <br> **Datatype:** Float. Default: `30`
| `reverse_train_test_order` | If true, FreqAI will train on the latest data split and test on historical split of the data. This allows the model to be trained up to the most recent data point, while avoiding overfitting. However, users should be careful to understand unorthodox nature of this parameter before employing it. <br> **Datatype:** Boolean. Default: False
| | **Data split parameters**
| `data_split_parameters` | Include any additional parameters available from Scikit-learn `test_train_split()`, which are shown [here](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html) (external website). <br> **Datatype:** Dictionary.
| `test_size` | Fraction of data that should be used for testing instead of training. <br> **Datatype:** Positive float < 1.
| `shuffle` | Shuffle the training data points during training. Typically, for time-series forecasting, this is set to `False`. <br>
| `shuffle` | Shuffle the training data points during training. Typically, for time-series forecasting, this is set to `False`. <br> **Datatype:** Boolean.
| | **Model training parameters**
| `model_training_parameters` | A flexible dictionary that includes all parameters available by the user selected model library. For example, if the user uses `LightGBMRegressor`, this dictionary can contain any parameter available by the `LightGBMRegressor` [here](https://lightgbm.readthedocs.io/en/latest/pythonapi/lightgbm.LGBMRegressor.html) (external website). If the user selects a different model, this dictionary can contain any parameter from that model. <br> **Datatype:** Dictionary.**Datatype:** Boolean.
| `model_training_parameters` | A flexible dictionary that includes all parameters available by the user selected model library. For example, if the user uses `LightGBMRegressor`, this dictionary can contain any parameter available by the `LightGBMRegressor` [here](https://lightgbm.readthedocs.io/en/latest/pythonapi/lightgbm.LGBMRegressor.html) (external website). If the user selects a different model, this dictionary can contain any parameter from that model. <br> **Datatype:** Dictionary.
| `n_estimators` | The number of boosted trees to fit in regression. <br> **Datatype:** Integer.
| `learning_rate` | Boosting learning rate during regression. <br> **Datatype:** Float.
| `n_jobs`, `thread_count`, `task_type` | Set the number of threads for parallel processing and the `task_type` (`gpu` or `cpu`). Different model libraries use different parameter names. <br> **Datatype:** Float.
@ -749,7 +749,7 @@ Given a number of data points $N$, and a distance $\varepsilon$, DBSCAN clusters
![dbscan](assets/freqai_dbscan.jpg)
FreqAI uses `sklearn.cluster.DBSCAN` (details are available on scikit-learn's webpage [here](#https://scikit-learn.org/stable/modules/generated/sklearn.cluster.DBSCAN.html)) with `min_samples` ($N$) taken as double the no. of user-defined features, and `eps` ($\varepsilon$) taken as the longest distance in the *k-distance graph* computed from the nearest neighbors in the pairwise distances of all data points in the feature set.
FreqAI uses `sklearn.cluster.DBSCAN` (details are available on scikit-learn's webpage [here](#https://scikit-learn.org/stable/modules/generated/sklearn.cluster.DBSCAN.html)) with `min_samples` ($N$) taken as 1/4 of the no. of time points in the feature set, and `eps` ($\varepsilon$) taken as the elbow point in the *k-distance graph* computed from the nearest neighbors in the pairwise distances of all data points in the feature set.
## Additional information
@ -774,5 +774,5 @@ Code review, software architecture brainstorming:
@xmatthias
Beta testing and bug reporting:
@bloodhunter4rc, Salah Lamkadem @ikonx, @ken11o2, @longyu, @paranoidandy, @smidelis, @smarm
@bloodhunter4rc, Salah Lamkadem @ikonx, @ken11o2, @longyu, @paranoidandy, @smidelis, @smarm,
Juha Nykänen @suikula, Wagner Costa @wagnercosta

View File

@ -824,6 +824,8 @@ Options:
- Merge the dataframe without lookahead bias
- Forward-fill (optional)
For a full sample, please refer to the [complete data provider example](#complete-data-provider-sample) below.
All columns of the informative dataframe will be available on the returning dataframe in a renamed fashion:
!!! Example "Column renaming"

View File

@ -142,12 +142,15 @@ class FreqtradeBot(LoggingMixin):
:return: None
"""
logger.info('Cleaning up modules ...')
try:
# Wrap db activities in shutdown to avoid problems if database is gone,
# and raises further exceptions.
if self.config['cancel_open_orders_on_exit']:
self.cancel_all_open_orders()
self.check_for_open_trades()
finally:
self.strategy.ft_bot_cleanup()
self.rpc.cleanup()
@ -283,7 +286,7 @@ class FreqtradeBot(LoggingMixin):
pair=trade.pair,
amount=trade.amount,
is_short=trade.is_short,
open_date=trade.open_date_utc
open_date=trade.date_last_filled_utc
)
trade.funding_fees = funding_fees
else:
@ -728,10 +731,11 @@ class FreqtradeBot(LoggingMixin):
fee = self.exchange.get_fee(symbol=pair, taker_or_maker='maker')
base_currency = self.exchange.get_pair_base_currency(pair)
open_date = datetime.now(timezone.utc)
funding_fees = self.exchange.get_funding_fees(
pair=pair, amount=amount, is_short=is_short, open_date=open_date)
# This is a new trade
if trade is None:
funding_fees = self.exchange.get_funding_fees(
pair=pair, amount=amount, is_short=is_short, open_date=open_date)
trade = Trade(
pair=pair,
base_currency=base_currency,
@ -1486,7 +1490,7 @@ class FreqtradeBot(LoggingMixin):
pair=trade.pair,
amount=trade.amount,
is_short=trade.is_short,
open_date=trade.open_date_utc,
open_date=trade.date_last_filled_utc,
)
exit_type = 'exit'
exit_reason = exit_tag or exit_check.exit_reason

View File

@ -686,7 +686,7 @@ class Backtesting:
self.futures_data[trade.pair],
amount=trade.amount,
is_short=trade.is_short,
open_date=trade.open_date_utc,
open_date=trade.date_last_filled_utc,
close_date=exit_candle_time,
)

View File

@ -212,17 +212,18 @@ def migrate_orders_table(engine, table_back_name: str, cols_order: List):
ft_fee_base = get_column_def(cols_order, 'ft_fee_base', 'null')
average = get_column_def(cols_order, 'average', 'null')
stop_price = get_column_def(cols_order, 'stop_price', 'null')
funding_fee = get_column_def(cols_order, 'funding_fee', '0.0')
# sqlite does not support literals for booleans
with engine.begin() as connection:
connection.execute(text(f"""
insert into orders (id, ft_trade_id, ft_order_side, ft_pair, ft_is_open, order_id,
status, symbol, order_type, side, price, amount, filled, average, remaining, cost,
stop_price, order_date, order_filled_date, order_update_date, ft_fee_base)
stop_price, order_date, order_filled_date, order_update_date, ft_fee_base, funding_fee)
select id, ft_trade_id, ft_order_side, ft_pair, ft_is_open, order_id,
status, symbol, order_type, side, price, amount, filled, {average} average, remaining,
cost, {stop_price} stop_price, order_date, order_filled_date,
order_update_date, {ft_fee_base} ft_fee_base
order_update_date, {ft_fee_base} ft_fee_base, {funding_fee} funding_fee
from {table_back_name}
"""))
@ -307,9 +308,10 @@ def check_migrate(engine, decl_base, previous_tables) -> None:
# Check if migration necessary
# Migrates both trades and orders table!
# if ('orders' not in previous_tables
# or not has_column(cols_orders, 'stop_price')):
# or not has_column(cols_orders, 'funding_fee')):
migrating = False
if not has_column(cols_trades, 'contract_size'):
# if not has_column(cols_trades, 'contract_size'):
if not has_column(cols_orders, 'funding_fee'):
migrating = True
logger.info(f"Running database migration for trades - "
f"backup: {table_back_name}, {order_table_bak_name}")

View File

@ -65,6 +65,8 @@ class Order(_DECL_BASE):
order_filled_date = Column(DateTime, nullable=True)
order_update_date = Column(DateTime, nullable=True)
funding_fee = Column(Float, nullable=True)
ft_fee_base = Column(Float, nullable=True)
@property
@ -72,6 +74,13 @@ class Order(_DECL_BASE):
""" Order-date with UTC timezoneinfo"""
return self.order_date.replace(tzinfo=timezone.utc)
@property
def order_filled_utc(self) -> Optional[datetime]:
""" last order-date with UTC timezoneinfo"""
return (
self.order_filled_date.replace(tzinfo=timezone.utc) if self.order_filled_date else None
)
@property
def safe_price(self) -> float:
return self.average or self.price
@ -119,6 +128,10 @@ class Order(_DECL_BASE):
self.ft_is_open = True
if self.status in NON_OPEN_EXCHANGE_STATES:
self.ft_is_open = False
if self.trade:
# Assign funding fee up to this point
# (represents the funding fee since the last order)
self.funding_fee = self.trade.funding_fees
if (order.get('filled', 0.0) or 0.0) > 0:
self.order_filled_date = datetime.now(timezone.utc)
self.order_update_date = datetime.now(timezone.utc)
@ -179,6 +192,10 @@ class Order(_DECL_BASE):
self.remaining = 0
self.status = 'closed'
self.ft_is_open = False
# Assign funding fees to Order.
# Assumes backtesting will use date_last_filled_utc to calculate future funding fees.
self.funding_fee = trade.funding_fees
if (self.ft_order_side == trade.entry_side):
trade.open_rate = self.price
trade.recalc_trade_from_orders()
@ -346,6 +363,15 @@ class LocalTrade():
else:
return self.amount
@property
def date_last_filled_utc(self) -> datetime:
""" Date of the last filled order"""
orders = self.select_filled_orders()
if not orders:
return self.open_date_utc
return max([self.open_date_utc,
max(o.order_filled_utc for o in orders if o.order_filled_utc)])
@property
def open_date_utc(self):
return self.open_date.replace(tzinfo=timezone.utc)
@ -843,10 +869,14 @@ class LocalTrade():
close_profit = 0.0
close_profit_abs = 0.0
profit = None
for o in self.orders:
# Reset funding fees
self.funding_fees = 0.0
funding_fees = 0.0
ordercount = len(self.orders) - 1
for i, o in enumerate(self.orders):
if o.ft_is_open or not o.filled:
continue
funding_fees += (o.funding_fee or 0.0)
tmp_amount = FtPrecise(o.safe_amount_after_fee)
tmp_price = FtPrecise(o.safe_price)
@ -861,7 +891,11 @@ class LocalTrade():
avg_price = current_stake / current_amount
if is_exit:
# Process partial exits
# Process exits
if i == ordercount and is_closing:
# Apply funding fees only to the last closing order
self.funding_fees = funding_fees
exit_rate = o.safe_price
exit_amount = o.safe_amount_after_fee
profit = self.calc_profit(rate=exit_rate, amount=exit_amount,
@ -871,6 +905,7 @@ class LocalTrade():
exit_rate, amount=exit_amount, open_rate=avg_price)
else:
total_stake = total_stake + self._calc_open_trade_value(tmp_amount, price)
self.funding_fees = funding_fees
if close_profit:
self.close_profit = close_profit

View File

@ -261,11 +261,15 @@ class RPC:
profit_str += f" ({fiat_profit:.2f})"
fiat_profit_sum = fiat_profit if isnan(fiat_profit_sum) \
else fiat_profit_sum + fiat_profit
open_order = (trade.select_order_by_order_id(
trade.open_order_id) if trade.open_order_id else None)
detail_trade = [
f'{trade.id} {direction_str}',
trade.pair + ('*' if (trade.open_order_id is not None
and trade.close_rate_requested is None) else '')
+ ('**' if (trade.close_rate_requested is not None) else ''),
trade.pair + ('*' if (open_order
and open_order.ft_order_side == trade.entry_side) else '')
+ ('**' if (open_order and
open_order.ft_order_side == trade.exit_side is not None) else ''),
shorten_date(arrow.get(trade.open_date).humanize(only_distance=True)),
profit_str
]

View File

@ -615,21 +615,25 @@ def test_calc_open_close_trade_price(
is_short=is_short,
leverage=lev,
trading_mode=trading_mode,
funding_fees=funding_fees
)
entry_order = limit_order[trade.entry_side]
exit_order = limit_order[trade.exit_side]
trade.open_order_id = f'something-{is_short}-{lev}-{exchange}'
oobj = Order.parse_from_ccxt_object(entry_order, 'ADA/USDT', trade.entry_side)
trade.orders.append(oobj)
oobj.trade = trade
oobj.update_from_ccxt_object(entry_order)
trade.update_trade(oobj)
trade.funding_fees = funding_fees
oobj = Order.parse_from_ccxt_object(exit_order, 'ADA/USDT', trade.exit_side)
trade.orders.append(oobj)
oobj.trade = trade
oobj.update_from_ccxt_object(exit_order)
trade.update_trade(oobj)
assert trade.is_open is False
assert trade.funding_fees == funding_fees
assert pytest.approx(trade._calc_open_trade_value(trade.amount, trade.open_rate)) == open_value
assert pytest.approx(trade.calc_close_trade_value(trade.close_rate)) == close_value