Module pinkfish.plot
Plotting functions.
Expand source code
"""
Plotting functions.
"""
import matplotlib.pyplot as plt
from pandas.plotting import register_matplotlib_converters
# Register matplotlib converters.
register_matplotlib_converters()
import pinkfish.pfstatistics as pfstatistics
import pinkfish.trade as trade
def plot_equity_curve(strategy, benchmark=None, yscale='linear', fname=None):
"""
Plot Equity Curve: Strategy and (optionally) Benchmark.
Parameters
----------
strategy : pd.DataFrame
Daily balance for the strategy.
benchmark: pd.DataFrame, optional
Daily balance for the benchmark (default is None, which implies
that a benchmark is not being used).
yscale: str, {'linear', 'log', 'symlog', 'logit'}
The axis scale type to apply (default is 'linear')
fname: str or path-like or file-like, optional
Save the current figure to fname (default is None, which
implies to not output the figure to a file).
Returns
-------
None
"""
fig = plt.figure()
axes = fig.add_subplot(111, ylabel='Portfolio value in $')
axes.plot(strategy['close'], label='strategy')
axes.set_yscale(yscale)
if benchmark is not None:
axes.plot(benchmark['close'], label='benchmark')
plt.legend(loc='best')
if fname:
plt.savefig(fname, bbox_inches='tight')
def plot_equity_curves(strategies, labels=None, yscale='linear', fname=None):
"""
Plot one or more equity curves on the same plot.
Parameters
----------
strategies : pd.Series of pd.Dataframe
Container of strategy daily balance for each symbol.
labels : list of str, optional
List of labels for each strategy (default is None, which implies
that `strategy.symbol` is used as the label.
yscale: str, {'linear', 'log', 'symlog', 'logit'}
The axis scale type to apply (default is 'linear')
fname: str or path-like or file-like, optional
Save the current figure to fname (default is None, which
implies to not output the figure to a file).
Returns
-------
None
"""
fig = plt.figure(figsize=(16,12))
axes = fig.add_subplot(111, ylabel='Portfolio value in $')
for i, strategy in enumerate(strategies):
if labels is None:
label = strategy.symbol
else:
label = labels[i]
axes.plot(strategy.dbal['close'], label=label)
axes.set_yscale(yscale)
plt.legend(loc='best')
if fname:
plt.savefig(fname, bbox_inches='tight')
def plot_trades(strategy, benchmark=None, yscale='linear', fname=None):
"""
Plot Trades.
Benchmark is the equity curve that the trades get plotted on.
If not provided, strategy equity curve is used.
Parameters
----------
strategy : pd.DataFrame
Daily balance for the strategy.
benchmark: pd.DataFrame, optional
Daily balance for the benchmark.
yscale: str, {'linear', 'log', 'symlog', 'logit'}
The axis scale type to apply (default is 'linear')
fname: str or path-like or file-like, optional
Save the current figure to fname (default is None, which
implies to not output the figure to a file).
Returns
-------
None
"""
if benchmark is None or strategy is benchmark:
benchmark = strategy
label = 'strategy'
else:
label = 'benchmark'
fig = plt.figure()
axes = fig.add_subplot(111, ylabel='Portfolio value in $')
axes.plot(benchmark.index, benchmark['close'], label=label)
axes.set_yscale(yscale)
# Buy trades.
s = strategy['state'] == trade.TradeState.OPEN
s = s.reindex_like(benchmark)
buy = benchmark[s]
axes.plot(buy.index, buy['close'], '^', markersize=10, color='k')
axes.set_yscale(yscale)
# Sell trades.
s = strategy['state'] == trade.TradeState.CLOSE
s = s.reindex_like(benchmark)
sell = benchmark[s]
axes.plot(sell.index, sell['close'], 'v', markersize=10, color='r')
axes.set_yscale(yscale)
plt.legend(loc='best')
if fname:
plt.savefig(fname, bbox_inches='tight')
default_metrics = (
'annual_return_rate',
'max_closed_out_drawdown',
'annualized_return_over_max_drawdown',
'best_month',
'worst_month',
'sharpe_ratio',
'sortino_ratio',
'monthly_std',
'annual_std')
"""
tuple : Default metrics for plot_bar_graph().
The metrics are:
'annual_return_rate'
'max_closed_out_drawdown'
'annualized_return_over_max_drawdown'
'best_month'
'worst_month'
'sharpe_ratio'
'sortino_ratio'
'monthly_std'
'annual_std'
"""
def plot_bar_graph(stats, benchmark_stats=None, metrics=default_metrics,
extras=None, fname=None):
"""
Plot Bar Graph: Strategy vs Benchmark (optional).
Parameters
----------
stats : pd.Series
Statistics from the strategy.
benchmark_stats : pd.Series, optional
Statistics from the benchmark (default is None, which implies
that a benchmark is not being used).
metrics: tuple, optional
The metrics to be plotted (default is `default_metrics`).
extras: tuple, optional
The additional metrics to be plotted (default is None, which
implies no extra metrics should be added).
fname: str or path-like or file-like, optional
Save the current figure to fname (default is None, which
implies to not output the figure to a file).
Returns
-------
pd.DataFrame
Summary metrics.
"""
if extras is None:
extras = ()
metrics += extras
df = pfstatistics.summary(stats, benchmark_stats, metrics)
fig = plt.figure()
axes = fig.add_subplot(111, ylabel='Trading Metrix')
df.plot(kind='bar', ax=axes, color=['g', 'r'])
axes.set_xticklabels(df.index, rotation=60)
plt.legend(loc='best')
if fname:
plt.savefig(fname, bbox_inches='tight')
return df
def optimizer_plot_bar_graph(df, metric):
"""
Plot Bar Graph of a metric for a set of strategies.
This function is designed to be used in analysis of an
optimization of some parameter. First call optimizer_summary()
to generate the dataframe required by this function.
Parameters
----------
df : pd.DataFrame
Summary of strategies vs metrics.
metric : str
The metric to be used in the summary.
"""
df = df.loc[[metric]]
df = df.transpose()
fig = plt.figure()
axes = fig.add_subplot(111, ylabel=metric)
df.plot(kind='bar', ax=axes, legend=False)
axes.set_xticklabels(df.index, rotation=0)Global variables
var default_metrics-
tuple : Default metrics for plot_bar_graph().
The metrics are:
'annual_return_rate' 'max_closed_out_drawdown' 'annualized_return_over_max_drawdown' 'best_month' 'worst_month' 'sharpe_ratio' 'sortino_ratio' 'monthly_std' 'annual_std'
Functions
def optimizer_plot_bar_graph(df, metric)-
Plot Bar Graph of a metric for a set of strategies.
This function is designed to be used in analysis of an optimization of some parameter. First call optimizer_summary() to generate the dataframe required by this function.
Parameters
df:pd.DataFrame- Summary of strategies vs metrics.
metric:str- The metric to be used in the summary.
Expand source code
def optimizer_plot_bar_graph(df, metric): """ Plot Bar Graph of a metric for a set of strategies. This function is designed to be used in analysis of an optimization of some parameter. First call optimizer_summary() to generate the dataframe required by this function. Parameters ---------- df : pd.DataFrame Summary of strategies vs metrics. metric : str The metric to be used in the summary. """ df = df.loc[[metric]] df = df.transpose() fig = plt.figure() axes = fig.add_subplot(111, ylabel=metric) df.plot(kind='bar', ax=axes, legend=False) axes.set_xticklabels(df.index, rotation=0) def plot_bar_graph(stats, benchmark_stats=None, metrics=('annual_return_rate', 'max_closed_out_drawdown', 'annualized_return_over_max_drawdown', 'best_month', 'worst_month', 'sharpe_ratio', 'sortino_ratio', 'monthly_std', 'annual_std'), extras=None, fname=None)-
Plot Bar Graph: Strategy vs Benchmark (optional).
Parameters
stats:pd.Series- Statistics from the strategy.
benchmark_stats:pd.Series, optional- Statistics from the benchmark (default is None, which implies that a benchmark is not being used).
metrics:tuple, optional- The metrics to be plotted (default is
default_metrics). extras:tuple, optional- The additional metrics to be plotted (default is None, which implies no extra metrics should be added).
fname:strorpath-likeorfile-like, optional- Save the current figure to fname (default is None, which implies to not output the figure to a file).
Returns
pd.DataFrame- Summary metrics.
Expand source code
def plot_bar_graph(stats, benchmark_stats=None, metrics=default_metrics, extras=None, fname=None): """ Plot Bar Graph: Strategy vs Benchmark (optional). Parameters ---------- stats : pd.Series Statistics from the strategy. benchmark_stats : pd.Series, optional Statistics from the benchmark (default is None, which implies that a benchmark is not being used). metrics: tuple, optional The metrics to be plotted (default is `default_metrics`). extras: tuple, optional The additional metrics to be plotted (default is None, which implies no extra metrics should be added). fname: str or path-like or file-like, optional Save the current figure to fname (default is None, which implies to not output the figure to a file). Returns ------- pd.DataFrame Summary metrics. """ if extras is None: extras = () metrics += extras df = pfstatistics.summary(stats, benchmark_stats, metrics) fig = plt.figure() axes = fig.add_subplot(111, ylabel='Trading Metrix') df.plot(kind='bar', ax=axes, color=['g', 'r']) axes.set_xticklabels(df.index, rotation=60) plt.legend(loc='best') if fname: plt.savefig(fname, bbox_inches='tight') return df def plot_equity_curve(strategy, benchmark=None, yscale='linear', fname=None)-
Plot Equity Curve: Strategy and (optionally) Benchmark.
Parameters
strategy:pd.DataFrame- Daily balance for the strategy.
benchmark:pd.DataFrame, optional- Daily balance for the benchmark (default is None, which implies that a benchmark is not being used).
yscale:str, {'linear', 'log', 'symlog', 'logit'}- The axis scale type to apply (default is 'linear')
fname:strorpath-likeorfile-like, optional- Save the current figure to fname (default is None, which implies to not output the figure to a file).
Returns
None
Expand source code
def plot_equity_curve(strategy, benchmark=None, yscale='linear', fname=None): """ Plot Equity Curve: Strategy and (optionally) Benchmark. Parameters ---------- strategy : pd.DataFrame Daily balance for the strategy. benchmark: pd.DataFrame, optional Daily balance for the benchmark (default is None, which implies that a benchmark is not being used). yscale: str, {'linear', 'log', 'symlog', 'logit'} The axis scale type to apply (default is 'linear') fname: str or path-like or file-like, optional Save the current figure to fname (default is None, which implies to not output the figure to a file). Returns ------- None """ fig = plt.figure() axes = fig.add_subplot(111, ylabel='Portfolio value in $') axes.plot(strategy['close'], label='strategy') axes.set_yscale(yscale) if benchmark is not None: axes.plot(benchmark['close'], label='benchmark') plt.legend(loc='best') if fname: plt.savefig(fname, bbox_inches='tight') def plot_equity_curves(strategies, labels=None, yscale='linear', fname=None)-
Plot one or more equity curves on the same plot.
Parameters
strategies:pd.Seriesofpd.Dataframe- Container of strategy daily balance for each symbol.
labels:listofstr, optional- List of labels for each strategy (default is None, which implies
that
strategy.symbolis used as the label. yscale:str, {'linear', 'log', 'symlog', 'logit'}- The axis scale type to apply (default is 'linear')
fname:strorpath-likeorfile-like, optional- Save the current figure to fname (default is None, which implies to not output the figure to a file).
Returns
None
Expand source code
def plot_equity_curves(strategies, labels=None, yscale='linear', fname=None): """ Plot one or more equity curves on the same plot. Parameters ---------- strategies : pd.Series of pd.Dataframe Container of strategy daily balance for each symbol. labels : list of str, optional List of labels for each strategy (default is None, which implies that `strategy.symbol` is used as the label. yscale: str, {'linear', 'log', 'symlog', 'logit'} The axis scale type to apply (default is 'linear') fname: str or path-like or file-like, optional Save the current figure to fname (default is None, which implies to not output the figure to a file). Returns ------- None """ fig = plt.figure(figsize=(16,12)) axes = fig.add_subplot(111, ylabel='Portfolio value in $') for i, strategy in enumerate(strategies): if labels is None: label = strategy.symbol else: label = labels[i] axes.plot(strategy.dbal['close'], label=label) axes.set_yscale(yscale) plt.legend(loc='best') if fname: plt.savefig(fname, bbox_inches='tight') def plot_trades(strategy, benchmark=None, yscale='linear', fname=None)-
Plot Trades.
Benchmark is the equity curve that the trades get plotted on. If not provided, strategy equity curve is used.
Parameters
strategy:pd.DataFrame- Daily balance for the strategy.
benchmark:pd.DataFrame, optional- Daily balance for the benchmark.
yscale:str, {'linear', 'log', 'symlog', 'logit'}- The axis scale type to apply (default is 'linear')
fname:strorpath-likeorfile-like, optional- Save the current figure to fname (default is None, which implies to not output the figure to a file).
Returns
None
Expand source code
def plot_trades(strategy, benchmark=None, yscale='linear', fname=None): """ Plot Trades. Benchmark is the equity curve that the trades get plotted on. If not provided, strategy equity curve is used. Parameters ---------- strategy : pd.DataFrame Daily balance for the strategy. benchmark: pd.DataFrame, optional Daily balance for the benchmark. yscale: str, {'linear', 'log', 'symlog', 'logit'} The axis scale type to apply (default is 'linear') fname: str or path-like or file-like, optional Save the current figure to fname (default is None, which implies to not output the figure to a file). Returns ------- None """ if benchmark is None or strategy is benchmark: benchmark = strategy label = 'strategy' else: label = 'benchmark' fig = plt.figure() axes = fig.add_subplot(111, ylabel='Portfolio value in $') axes.plot(benchmark.index, benchmark['close'], label=label) axes.set_yscale(yscale) # Buy trades. s = strategy['state'] == trade.TradeState.OPEN s = s.reindex_like(benchmark) buy = benchmark[s] axes.plot(buy.index, buy['close'], '^', markersize=10, color='k') axes.set_yscale(yscale) # Sell trades. s = strategy['state'] == trade.TradeState.CLOSE s = s.reindex_like(benchmark) sell = benchmark[s] axes.plot(sell.index, sell['close'], 'v', markersize=10, color='r') axes.set_yscale(yscale) plt.legend(loc='best') if fname: plt.savefig(fname, bbox_inches='tight')