#!/usr/bin/env python
# coding: utf-8
# In[1]:
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from mpl_toolkits import mplot3d
import matplotlib.cm as cm
from scipy.interpolate import griddata
import os
import sys
import concurrent
from pyqstrat.pq_utils import set_defaults, has_display
from pyqstrat.plot import *
set_defaults()
[docs]class Experiment:
'''An Experiment stores a suggestion and its result
Attributes:
suggestion: A dictionary of variable name -> value
cost: A float representing output of the function we are testing with this suggestion as input.
other_costs: A dictionary of other results we want to store and look at later.
'''
def __init__(self, suggestion, cost, other_costs):
self.suggestion = suggestion
self.cost = cost
self.other_costs = other_costs
[docs] def valid(self):
'''
Returns True if all suggestions and costs are finite, i.e not NaN or +/- Infinity
'''
if not all(np.isfinite(list(self.suggestion.values()))): return False
if not np.isfinite(self.cost): return False
if not all(np.isfinite(list(self.other_costs.values()))): return False
return True
def __repr__(self):
return f'suggestion: {self.suggestion} cost: {self.cost} other costs: {self.other_costs}'
[docs]class Optimizer:
'''Optimizer is used to optimize parameters for a strategy.'''
[docs] def __init__(self, name, generator, cost_func, max_processes = None):
'''
Args:
name: string used to display title in plotting, etc.
generator: A generator (see Python Generators) that takes no inputs and yields a list of dictionaries with parameter name -> parameter value.
cost_func: A function that takes a dictionary of parameter name -> parameter value as input and outputs cost for that set of parameters.
max_processes: If not set, the Optimizer will look at the number of CPU cores on your machine to figure out how many processes to run.
'''
self.name = name
self.generator = generator
self.cost_func = cost_func
import sys
if sys.platform in ['win32', 'cygwin']:
if max_processes is not None and max_processes != 1:
raise Exception("max_processes must be 1 on Microsoft Windows")
max_processes = 1
self.max_processes = max_processes
self.experiments = []
def _run_single_process(self):
try:
for suggestion in self.generator:
if suggestion is None: continue
cost, other_costs = self.cost_func(suggestion)
self.generator.send(cost)
self.experiments.append(Experiment(suggestion, cost, other_costs))
except StopIteration:
# Exhausted generator
return
#TODO: Needs to be rewritten to send costs back to generator when we do parallel gradient descent, etc.
def _run_multi_process(self, raise_on_error):
fut_map = {}
with concurrent.futures.ProcessPoolExecutor(self.max_processes) as executor:
for suggestion in self.generator:
if suggestion is None: continue
future = executor.submit(self.cost_func, suggestion)
fut_map[future] = suggestion
for future in concurrent.futures.as_completed(fut_map):
try:
cost, other_costs = future.result()
except Exception as e:
new_exc = type(e)(f'Exception: {str(e)} with suggestion: {suggestion}').with_traceback(sys.exc_info()[2])
if raise_on_error: raise new_exc
else: print(str(new_exc))
continue
suggestion = fut_map[future]
self.experiments.append(Experiment(suggestion, cost, other_costs))
[docs] def run(self, raise_on_error = False):
'''Run the optimizer.
Args:
raise_on_error: If set to True, even if we are running a multiprocess optimization, any Exceptions will bubble up and stop the Optimizer.
This can be useful for debugging to see stack traces for Exceptions.
'''
if self.max_processes == 1: self._run_single_process()
else: self._run_multi_process(raise_on_error)
[docs] def experiment_list(self, sort_order = 'lowest_cost'):
'''Returns the list of experiments we have run
Args:
sort_order: Can be set to lowest_cost, highest_cost or sequence.
If set to sequence, experiments are returned in the sequence in which they were run
'''
if sort_order == 'lowest_cost':
experiments = sorted(self.experiments, key = lambda x : x.cost, reverse = True)
elif sort_order == 'highest_cost':
experiments = sorted(self.experiments, key = lambda x : x.cost, reverse = False)
elif sort_order == 'sequence': # in order in which experiment was run
experiments = self.experiments
else:
raise Exception(f'invalid sort order: {sort}')
return experiments
[docs] def df_experiments(self, sort_column = 'cost', ascending = True):
'''
Returns a dataframe containing experiment data, sorted by sort_column (default "cost")
'''
if len(self.experiments) == 0: return None
pc_keys = list(self.experiments[0].other_costs.keys())
sugg_keys = list(self.experiments[0].suggestion.keys())
records = [[exp.suggestion[k] for k in sugg_keys] + [exp.cost] + [exp.other_costs[k] for k in pc_keys] for exp in self.experiments if exp.valid()]
df = pd.DataFrame.from_records(records, columns = sugg_keys + ['cost'] + pc_keys)
df.sort_values(by = [sort_column], ascending = ascending, inplace = True)
return df
[docs] def plot_3d(self, x, y, z = 'all', plot_type = 'surface', figsize = (15,15), interpolation = 'linear',
cmap = 'viridis', marker = 'X', marker_size = 50, marker_color = 'r', xlim = None, ylim = None, hspace = None):
"""Creates a 3D plot of the optimization output for plotting 2 parameters and costs.
Args:
x: Name of the parameter to plot on the x axis, corresponding to the same name in the generator.
y: Name of the parameter to plot on the y axis, corresponding to the same name in the generator.
z: Can be one of:
"cost"
The name of another cost variable corresponding to the output from the cost function
"all", which creates a subplot for cost plus all other costs
plot_type: surface or contour (default surface)
figsize: Figure size
interpolation: Can be ‘linear’, ‘nearest’ or ‘cubic’ for plotting z points between the ones passed in. See scipy.interpolate.griddata for details
cmap: Colormap to use (default viridis). See matplotlib colormap for details
marker: Adds a marker to each point in x, y, z to show the actual data used for interpolation. You can set this to None to turn markers off.
hspace: Vertical space between subplots
"""
if len(self.experiments) == 0:
print('No experiments found')
return
if not has_display(): return
# Get rid of nans since matplotlib does not like them
experiments = [experiment for experiment in self.experiments if experiment.valid()]
if not len(experiments):
print('No valid experiments found')
return
if xlim:
experiments = [experiment for experiment in experiments if experiment.suggestion[x] >= xlim[0] and experiment.suggestion[x] <= xlim[1]]
if ylim:
experiments = [experiment for experiment in experiments if experiment.suggestion[y] >= ylim[0] and experiment.suggestion[y] <= ylim[1]]
xvalues = [experiment.suggestion[x] for experiment in experiments]
yvalues = [experiment.suggestion[y] for experiment in experiments]
zvalues = []
if z == 'all':
zvalues.append(('cost', np.array([experiment.cost for experiment in experiments])))
if len(experiments[0].other_costs):
other_cost_keys = experiments[0].other_costs.keys()
for key in other_cost_keys:
zvalues.append((key, np.array([experiment.other_costs[key] for experiment in experiments])))
elif z == 'cost':
zvalues.append(('cost', np.array([experiment.cost for experiment in experiments])))
else:
zvalues.append((z, np.array([experiment.other_costs[zname] for experiment in experiments])))
subplots = []
for tup in zvalues:
name = tup[0]
zarray = tup[1]
if plot_type == 'contour':
zlabel = None
title = name
else:
zlabel = name
title = None
subplots.append(Subplot(data_list = [
XYZData(name, xvalues, yvalues, zarray, plot_type = plot_type,
marker = marker, marker_size = marker_size, marker_color = marker_color, interpolation = interpolation, cmap = cmap
)], title = title, xlabel = x, ylabel = y, zlabel = zlabel, xlim = xlim, ylim = ylim))
plot = Plot(subplots, figsize = figsize, title = 'Optimizer 2D Test', hspace = hspace)
plot.draw()
[docs] def plot_2d(self, x, y = 'all', plot_type = 'line', figsize = (15,8), marker = 'X', marker_size = 50, marker_color = 'r', xlim = None, hspace = None):
"""Creates a 2D plot of the optimization output for plotting 1 parameter and costs.
Args:
x: Name of the parameter to plot on the x axis, corresponding to the same name in the generator.
y: Can be one of:
"cost"
The name of another cost variable corresponding to the output from the cost function
"all", which creates a subplot for cost plus all other costs
plot_type: line or scatter (default line)
figsize: Figure size
marker: Adds a marker to each point in x, y to show the actual data used for interpolation. You can set this to None to turn markers off.
hspace: Vertical space between subplots
"""
if len(self.experiments) == 0: return
if not has_display(): return
# Get rid of nans since matplotlib does not like them
experiments = [experiment for experiment in self.experiments if experiment.valid()]
if xlim:
experiments = [experiment for experiment in experiments if experiment.suggestion[x] >= xlim[0] and experiment.suggestion[x] <= xlim[1]]
xvalues = [experiment.suggestion[x] for experiment in experiments]
yvalues = []
if y == 'all':
yvalues.append(('cost', np.array([experiment.cost for experiment in experiments])))
other_cost_keys = experiments[0].other_costs.keys()
for key in other_cost_keys:
yvalues.append((key, np.array([experiment.other_costs[key] for experiment in experiments])))
elif y == 'cost':
yvalues.append(('cost', np.array([experiment.cost for experiment in experiments])))
else:
yvalues.append((y, np.array([experiment.other_costs[zname] for experiment in experiments])))
xvalues = np.array(xvalues)
x_sort_indices = np.argsort(xvalues)
xvalues = xvalues[x_sort_indices]
subplots = []
for tup in yvalues:
name = tup[0]
yarray = tup[1]
idx = np.argsort(xvalues)
yarray = yarray[x_sort_indices]
subplots.append(Subplot(data_list = [
XYData(name, xvalues, yarray, plot_type = plot_type,
marker = marker, marker_size = marker_size, marker_color = marker_color
)], xlabel = x, ylabel = name, xlim = xlim))
plot = Plot(subplots, figsize = figsize, title = 'Optimizer 1D Test')
plot.draw()
# Functions used in unit testing
def _generator_1d():
for x in np.arange(0, np.pi * 2, 0.1):
costs = (yield {'x' : x})
def _cost_func_1d(suggestion):
x = suggestion['x']
cost = np.sin(x)
ret = cost, {'std' : -0.1 * cost}
return ret
def _generator_2d():
for x in np.arange(0, np.pi * 2, 0.5):
for y in np.arange(0, np.pi * 2, 0.5):
costs = (yield {'x' : x, 'y' : y})
def _cost_func_2d(suggestion):
x = suggestion['x']
y = suggestion['y']
cost = np.sin(np.sqrt(x**2 + y ** 2))
return cost, {'sharpe' : cost, 'std' : -0.1 * cost}
[docs]def test_optimize():
max_processes = 1 if os.name == 'nt' else 4
optimizer_1d = Optimizer('test', _generator_1d(), _cost_func_1d, max_processes = max_processes)
optimizer_1d.run(raise_on_error = True)
optimizer_1d.plot_2d(x = 'x', plot_type = 'line', marker = 'o', marker_color = 'blue')
optimizer_2d = Optimizer('test', _generator_2d(), _cost_func_2d, max_processes = max_processes)
optimizer_2d.run()
optimizer_2d.plot_3d(x = 'x', y = 'y')
if __name__ == "__main__":
test_optimize()
import doctest
doctest.testmod(optionflags = doctest.NORMALIZE_WHITESPACE)