From 9488e8992dea615420b4712be52774b8995bec5e Mon Sep 17 00:00:00 2001 From: froggleston Date: Sun, 22 May 2022 23:24:52 +0100 Subject: [PATCH] First commit for integrating buy_reasons into FT --- freqtrade/commands/__init__.py | 1 + freqtrade/commands/arguments.py | 14 +- freqtrade/commands/cli_options.py | 31 +++ freqtrade/configuration/configuration.py | 15 ++ freqtrade/data/entryexitanalysis.py | 258 +++++++++++++++++++++++ 5 files changed, 317 insertions(+), 2 deletions(-) create mode 100755 freqtrade/data/entryexitanalysis.py diff --git a/freqtrade/commands/__init__.py b/freqtrade/commands/__init__.py index 0e637c487..d93ed1e09 100644 --- a/freqtrade/commands/__init__.py +++ b/freqtrade/commands/__init__.py @@ -6,6 +6,7 @@ Contains all start-commands, subcommands and CLI Interface creation. Note: Be careful with file-scoped imports in these subfiles. as they are parsed on startup, nothing containing optional modules should be loaded. """ +from freqtrade.commands.analyze_commands import start_analysis_entries_exits from freqtrade.commands.arguments import Arguments from freqtrade.commands.build_config_commands import start_new_config from freqtrade.commands.data_commands import (start_convert_data, start_convert_trades, diff --git a/freqtrade/commands/arguments.py b/freqtrade/commands/arguments.py index 815e28175..4dd0141fa 100644 --- a/freqtrade/commands/arguments.py +++ b/freqtrade/commands/arguments.py @@ -101,6 +101,9 @@ ARGS_HYPEROPT_SHOW = ["hyperopt_list_best", "hyperopt_list_profitable", "hyperop "print_json", "hyperoptexportfilename", "hyperopt_show_no_header", "disableparamexport", "backtest_breakdown"] +ARGS_ANALYZE_ENTRIES_EXITS = ["analysis_groups", "enter_reason_list", + "exit_reason_list", "indicator_list"] + NO_CONF_REQURIED = ["convert-data", "convert-trade-data", "download-data", "list-timeframes", "list-markets", "list-pairs", "list-strategies", "list-data", "hyperopt-list", "hyperopt-show", "backtest-filter", @@ -182,8 +185,9 @@ class Arguments: self.parser = argparse.ArgumentParser(description='Free, open source crypto trading bot') self._build_args(optionlist=['version'], parser=self.parser) - from freqtrade.commands import (start_backtesting, start_backtesting_show, - start_convert_data, start_convert_db, start_convert_trades, + from freqtrade.commands import (start_analysis_entries_exits, start_backtesting, + start_backtesting_show, start_convert_data, + start_convert_db, start_convert_trades, start_create_userdir, start_download_data, start_edge, start_hyperopt, start_hyperopt_list, start_hyperopt_show, start_install_ui, start_list_data, start_list_exchanges, @@ -415,3 +419,9 @@ class Arguments: parents=[_common_parser]) webserver_cmd.set_defaults(func=start_webserver) self._build_args(optionlist=ARGS_WEBSERVER, parser=webserver_cmd) + + # Add backtesting analysis subcommand + analysis_cmd = subparsers.add_parser('analysis', help='Analysis module.', + parents=[_common_parser, _strategy_parser]) + analysis_cmd.set_defaults(func=start_analysis_entries_exits) + self._build_args(optionlist=ARGS_ANALYZE_ENTRIES_EXITS, parser=analysis_cmd) diff --git a/freqtrade/commands/cli_options.py b/freqtrade/commands/cli_options.py index aac9f5713..f925bd699 100644 --- a/freqtrade/commands/cli_options.py +++ b/freqtrade/commands/cli_options.py @@ -614,4 +614,35 @@ AVAILABLE_CLI_OPTIONS = { "that do not contain any parameters."), action="store_true", ), + "analysis_groups": Arg( + "--analysis_groups", + help=("grouping output - ", + "0: simple wins/losses by enter tag, ", + "1: by enter_tag, ", + "2: by enter_tag and exit_tag, ", + "3: by pair and enter_tag, ", + "4: by pair, enter_ and exit_tag (this can get quite large)"), + nargs='?', + default="0,1,2", + ), + "enter_reason_list": Arg( + "--enter_reason_list", + help=("Comma separated list of entry signals to analyse. Default: all. ", + "e.g. 'entry_tag_a,entry_tag_b'"), + nargs='?', + default='all', + ), + "exit_reason_list": Arg( + "--exit_reason_list", + help=("Comma separated list of exit signals to analyse. Default: all. ", + "e.g. 'exit_tag_a,roi,stop_loss,trailing_stop_loss'"), + nargs='?', + default='all', + ), + "indicator_list": Arg( + "--indicator_list", + help=("Comma separated list of indicators to analyse. ", + "e.g. 'close,rsi,bb_lowerband,profit_abs'"), + nargs='?', + ), } diff --git a/freqtrade/configuration/configuration.py b/freqtrade/configuration/configuration.py index 96b585cd1..ea4bcace8 100644 --- a/freqtrade/configuration/configuration.py +++ b/freqtrade/configuration/configuration.py @@ -95,6 +95,8 @@ class Configuration: self._process_data_options(config) + self._process_analyze_options(config) + # Check if the exchange set by the user is supported check_exchange(config, config.get('experimental', {}).get('block_bad_exchanges', True)) @@ -433,6 +435,19 @@ class Configuration: self._args_to_config(config, argname='candle_types', logstring='Detected --candle-types: {}') + def _process_analyze_options(self, config: Dict[str, Any]) -> None: + self._args_to_config(config, argname='analysis_groups', + logstring='Analysis reason groups: {}') + + self._args_to_config(config, argname='enter_reason_list', + logstring='Analysis enter tag list: {}') + + self._args_to_config(config, argname='exit_reason_list', + logstring='Analysis exit tag list: {}') + + self._args_to_config(config, argname='indicator_list', + logstring='Analysis indicator list: {}') + def _process_runmode(self, config: Dict[str, Any]) -> None: self._args_to_config(config, argname='dry_run', diff --git a/freqtrade/data/entryexitanalysis.py b/freqtrade/data/entryexitanalysis.py new file mode 100755 index 000000000..62216d5ea --- /dev/null +++ b/freqtrade/data/entryexitanalysis.py @@ -0,0 +1,258 @@ +import joblib +import logging +import os + +from pathlib import Path +from typing import List, Optional + +import pandas as pd +from tabulate import tabulate + +from freqtrade.data.btanalysis import (load_backtest_data, get_latest_backtest_filename) +from freqtrade.exceptions import OperationalException + + +logger = logging.getLogger(__name__) + + +def _load_signal_candles(backtest_dir: Path): + scpf = Path(backtest_dir, + os.path.splitext( + get_latest_backtest_filename(backtest_dir))[0] + "_signals.pkl" + ) + try: + scp = open(scpf, "rb") + signal_candles = joblib.load(scp) + logger.info(f"Loaded signal candles: {str(scpf)}") + except Exception as e: + logger.error("Cannot load signal candles from pickled results: ", e) + + return signal_candles + + +def _process_candles_and_indicators(pairlist, strategy_name, trades, signal_candles): + analysed_trades_dict = {} + analysed_trades_dict[strategy_name] = {} + + try: + logger.info(f"Processing {strategy_name} : {len(pairlist)} pairs") + + for pair in pairlist: + if pair in signal_candles[strategy_name]: + analysed_trades_dict[strategy_name][pair] = _analyze_candles_and_indicators( + pair, + trades, + signal_candles[strategy_name][pair]) + except Exception: + pass + + return analysed_trades_dict + + +def _analyze_candles_and_indicators(pair, trades, signal_candles): + buyf = signal_candles + + if len(buyf) > 0: + buyf = buyf.set_index('date', drop=False) + trades_red = trades.loc[trades['pair'] == pair].copy() + + trades_inds = pd.DataFrame() + + if trades_red.shape[0] > 0 and buyf.shape[0] > 0: + for t, v in trades_red.open_date.items(): + allinds = buyf.loc[(buyf['date'] < v)] + if allinds.shape[0] > 0: + tmp_inds = allinds.iloc[[-1]] + + trades_red.loc[t, 'signal_date'] = tmp_inds['date'].values[0] + trades_red.loc[t, 'enter_reason'] = trades_red.loc[t, 'enter_tag'] + tmp_inds.index.rename('signal_date', inplace=True) + trades_inds = pd.concat([trades_inds, tmp_inds]) + + if 'signal_date' in trades_red: + trades_red['signal_date'] = pd.to_datetime(trades_red['signal_date'], utc=True) + trades_red.set_index('signal_date', inplace=True) + + try: + trades_red = pd.merge(trades_red, trades_inds, on='signal_date', how='outer') + except Exception as e: + print(e) + return trades_red + else: + return pd.DataFrame() + + +def _do_group_table_output(bigdf, glist): + if "0" in glist: + wins = bigdf.loc[bigdf['profit_abs'] >= 0] \ + .groupby(['enter_reason']) \ + .agg({'profit_abs': ['sum']}) + + wins.columns = ['profit_abs_wins'] + loss = bigdf.loc[bigdf['profit_abs'] < 0] \ + .groupby(['enter_reason']) \ + .agg({'profit_abs': ['sum']}) + loss.columns = ['profit_abs_loss'] + + new = bigdf.groupby(['enter_reason']).agg({'profit_abs': [ + 'count', + lambda x: sum(x > 0), + lambda x: sum(x <= 0)]}) + + new = pd.merge(new, wins, left_index=True, right_index=True) + new = pd.merge(new, loss, left_index=True, right_index=True) + + new['profit_tot'] = new['profit_abs_wins'] - abs(new['profit_abs_loss']) + + new['wl_ratio_pct'] = (new.iloc[:, 1] / new.iloc[:, 0] * 100) + new['avg_win'] = (new['profit_abs_wins'] / new.iloc[:, 1]) + new['avg_loss'] = (new['profit_abs_loss'] / new.iloc[:, 2]) + + new.columns = ['total_num_buys', 'wins', 'losses', 'profit_abs_wins', 'profit_abs_loss', + 'profit_tot', 'wl_ratio_pct', 'avg_win', 'avg_loss'] + + sortcols = ['total_num_buys'] + + _print_table(new, sortcols, show_index=True) + if "1" in glist: + new = bigdf.groupby(['enter_reason']) \ + .agg({'profit_abs': ['count', 'sum', 'median', 'mean'], + 'profit_ratio': ['sum', 'median', 'mean']} + ).reset_index() + new.columns = ['enter_reason', 'num_buys', 'profit_abs_sum', 'profit_abs_median', + 'profit_abs_mean', 'median_profit_pct', 'mean_profit_pct', + 'total_profit_pct'] + sortcols = ['profit_abs_sum', 'enter_reason'] + + new['median_profit_pct'] = new['median_profit_pct'] * 100 + new['mean_profit_pct'] = new['mean_profit_pct'] * 100 + new['total_profit_pct'] = new['total_profit_pct'] * 100 + + _print_table(new, sortcols) + if "2" in glist: + new = bigdf.groupby(['enter_reason', 'exit_reason']) \ + .agg({'profit_abs': ['count', 'sum', 'median', 'mean'], + 'profit_ratio': ['sum', 'median', 'mean']} + ).reset_index() + new.columns = ['enter_reason', 'exit_reason', 'num_buys', 'profit_abs_sum', + 'profit_abs_median', 'profit_abs_mean', 'median_profit_pct', + 'mean_profit_pct', 'total_profit_pct'] + sortcols = ['profit_abs_sum', 'enter_reason'] + + new['median_profit_pct'] = new['median_profit_pct'] * 100 + new['mean_profit_pct'] = new['mean_profit_pct'] * 100 + new['total_profit_pct'] = new['total_profit_pct'] * 100 + + _print_table(new, sortcols) + if "3" in glist: + new = bigdf.groupby(['pair', 'enter_reason']) \ + .agg({'profit_abs': ['count', 'sum', 'median', 'mean'], + 'profit_ratio': ['sum', 'median', 'mean']} + ).reset_index() + new.columns = ['pair', 'enter_reason', 'num_buys', 'profit_abs_sum', + 'profit_abs_median', 'profit_abs_mean', 'median_profit_pct', + 'mean_profit_pct', 'total_profit_pct'] + sortcols = ['profit_abs_sum', 'enter_reason'] + + new['median_profit_pct'] = new['median_profit_pct'] * 100 + new['mean_profit_pct'] = new['mean_profit_pct'] * 100 + new['total_profit_pct'] = new['total_profit_pct'] * 100 + + _print_table(new, sortcols) + if "4" in glist: + new = bigdf.groupby(['pair', 'enter_reason', 'exit_reason']) \ + .agg({'profit_abs': ['count', 'sum', 'median', 'mean'], + 'profit_ratio': ['sum', 'median', 'mean']} + ).reset_index() + new.columns = ['pair', 'enter_reason', 'exit_reason', 'num_buys', 'profit_abs_sum', + 'profit_abs_median', 'profit_abs_mean', 'median_profit_pct', + 'mean_profit_pct', 'total_profit_pct'] + sortcols = ['profit_abs_sum', 'enter_reason'] + + new['median_profit_pct'] = new['median_profit_pct'] * 100 + new['mean_profit_pct'] = new['mean_profit_pct'] * 100 + new['total_profit_pct'] = new['total_profit_pct'] * 100 + + _print_table(new, sortcols) + + +def _print_results(analysed_trades, stratname, group, + 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'] + + bigdf = pd.DataFrame() + for pair, trades in analysed_trades[stratname].items(): + bigdf = pd.concat([bigdf, trades], ignore_index=True) + + if bigdf.shape[0] > 0 and ('enter_reason' in bigdf.columns): + if group is not None: + glist = group.split(",") + _do_group_table_output(bigdf, glist) + + if enter_reason_list is not None and not enter_reason_list == "all": + enter_reason_list = enter_reason_list.split(",") + bigdf = bigdf.loc[(bigdf['enter_reason'].isin(enter_reason_list))] + + if exit_reason_list is not None and not exit_reason_list == "all": + exit_reason_list = exit_reason_list.split(",") + bigdf = bigdf.loc[(bigdf['exit_reason'].isin(exit_reason_list))] + + if indicator_list is not None: + if indicator_list == "all": + print(bigdf) + else: + available_inds = [] + for ind in indicator_list.split(","): + if ind in bigdf: + available_inds.append(ind) + ilist = ["pair", "enter_reason", "exit_reason"] + available_inds + print(tabulate(bigdf[ilist].sort_values(['exit_reason']), + headers='keys', tablefmt='psql', showindex=False)) + else: + print(tabulate(bigdf[columns].sort_values(['pair']), + headers='keys', tablefmt='psql', showindex=False)) + else: + print("\\_ No trades to show") + + +def _print_table(df, sortcols=None, show_index=False): + if (sortcols is not None): + data = df.sort_values(sortcols) + else: + data = df + + print( + tabulate( + data, + headers='keys', + tablefmt='psql', + showindex=show_index + ) + ) + + +def process_entry_exit_reasons(backtest_dir: Path, + pairlist: List[str], + strategy_name: str, + analysis_groups: Optional[str] = "0,1,2", + enter_reason_list: Optional[str] = "all", + exit_reason_list: Optional[str] = "all", + indicator_list: Optional[str] = None): + + try: + trades = load_backtest_data(backtest_dir, strategy_name) + except ValueError as e: + raise OperationalException(e) from e + 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)