Commit 43d96681 authored by Gregory Ashton's avatar Gregory Ashton

Improvements to the PP plots

- Add legend/colors for each parameter
parent 7d720596
......@@ -3,6 +3,7 @@ from __future__ import division
import os
from distutils.version import LooseVersion
from collections import OrderedDict, namedtuple
from itertools import product
import numpy as np
import pandas as pd
......@@ -1051,20 +1052,28 @@ class Result(object):
= priors[key].prob(self.posterior[key].values)
def get_all_injection_credible_levels(self):
def get_all_injection_credible_levels(self, keys=None):
Get credible levels for all parameters in self.injection_parameters
Get credible levels for all parameters
keys: list, optional
A list of keys for which return the credible levels, if None,
defaults to search_parameter_keys
credible_levels: dict
The credible levels at which the injected parameters are found.
if keys is None:
keys = self.search_parameter_keys
if self.injection_parameters is None:
raise(TypeError, "Result object has no 'injection_parameters'. "
"Cannot copmute credible levels.")
"Cannot compute credible levels.")
credible_levels = {key: self.get_injection_credible_level(key)
for key in self.search_parameter_keys
for key in keys
if isinstance(self.injection_parameters[key], float)}
return credible_levels
......@@ -1259,7 +1268,8 @@ def plot_multiple(results, filename=None, labels=None, colours=None,
return fig
def make_pp_plot(results, filename=None, save=True, **kwargs):
def make_pp_plot(results, filename=None, save=True, confidence_interval=0.9,
lines=None, legend_fontsize=9, keys=None, **kwargs):
Make a P-P plot for a set of runs with injected signals.
......@@ -1271,6 +1281,15 @@ def make_pp_plot(results, filename=None, save=True, **kwargs):
The name of the file to save, the default is "outdir/pp.png"
save: bool, optional
Whether to save the file, default=True
confidence_interval: float, optional
The confidence interval to be plotted, defaulting to 0.9 (90%)
lines: list
If given, a list of matplotlib line formats to use, must be greater
than the number of parameters.
legend_fontsize: float
The font size for the legend
keys: list
A list of keys to use, if None defaults to search_parameter_keys
Additional kwargs to pass to matplotlib.pyplot.plot
......@@ -1279,25 +1298,47 @@ def make_pp_plot(results, filename=None, save=True, **kwargs):
matplotlib figure
fig = plt.figure()
credible_levels = pd.DataFrame()
for result in results:
credible_levels = credible_levels.append(
result.get_all_injection_credible_levels(), ignore_index=True)
n_parameters = len(credible_levels.keys())
x_values = np.linspace(0, 1, 101)
for key in credible_levels:
plt.plot(x_values, [sum(credible_levels[key].values < xx) /
len(credible_levels) for xx in x_values],
color='k', alpha=min([1, 4 / n_parameters]), **kwargs)
plt.plot([0, 1], [0, 1], linestyle='--', color='r')
plt.xlim(0, 1)
plt.ylim(0, 1)
result.get_all_injection_credible_levels(keys), ignore_index=True)
if lines is None:
colors = ["C{}".format(i) for i in range(8)]
linestyles = ["-", "--", ":"]
lines = ["{}{}".format(a, b) for a, b in product(linestyles, colors)]
if len(lines) < len(credible_levels.keys()):
raise ValueError("Larger number of parameters than unique linestyles")
x_values = np.linspace(0, 1, 1001)
# Putting in the confidence bands
N = len(credible_levels)
edge_of_bound = (1. - confidence_interval) / 2.
lower = scipy.stats.binom.ppf(1 - edge_of_bound, N, x_values) / N
upper = scipy.stats.binom.ppf(edge_of_bound, N, x_values) / N
# The binomial point percent function doesn't always return 0 @ 0,
# so set those bounds explicitly to be sure
lower[0] = 0
upper[0] = 0
fig, ax = plt.subplots()
ax.fill_between(x_values, lower, upper, alpha=0.2, color='k')
for ii, key in enumerate(credible_levels):
pp = np.array([sum(credible_levels[key].values < xx) /
len(credible_levels) for xx in x_values])
plt.plot(x_values, pp, lines[ii], label=key, **kwargs)
ax.set_xlim(0, 1)
ax.set_ylim(0, 1)
if save:
if filename is None:
filename = 'outdir/pp.png'
fig.savefig(filename, dpi=500)
return fig
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment