def plot_model_fits(y, yhat, fitparams, err=None, palettes=[gpal, bpal], save=False, cdf=True, bw=.01):
""" main plotting function for displaying model fit predictions over data
"""
sns.set(style='darkgrid', rc={'figure.facecolor':'white'}, font_scale=1.5)
# extract model and fit info from fitparams
nlevels = fitparams['nlevels']
ssd, nssd, nss, nss_per, ssd_ix = fitparams.ssd_info
quantiles = fitparams.quantiles
# make colors and labels
palettes = colors.get_cpals(aslist=True)[:nlevels]
clrs = [pal(2) for pal in palettes]
lbls = get_plot_labels(fitparams)
f, axes = plt.subplots(nlevels, 3, figsize=(14, 5*nlevels), sharey=True)
# make two dimensional for iterating 1st dim
axes = axes.reshape(nlevels, 3)
plot_acc = scurves
if nssd==1:
plot_acc = plot_accuracy
y_dat = unpack_vector(y, fitparams, bw=bw)
y_kde = unpack_vector(y, fitparams, bw=bw, kde=True)
yhat_dat = unpack_vector(yhat, fitparams, bw=bw)
if err is not None:
y_err = unpack_vector(err, fitparams, bw=bw)
sc_err = [e[0] for e in y_err]
qp_err = [[e[1], e[2]] for e in y_err]
else:
sc_err, qp_err = [[None]*nlevels]*2
for i, (ax1, ax2, ax3) in enumerate(axes):
accdata = [y_dat[i][0], yhat_dat[i][0]]
qpdata = [y_dat[i], yhat_dat[i]]
plot_acc(accdata, err=sc_err[i], ssd=ssd[i], colors=clrs[i], labels=lbls[i], ax=ax1)
plot_quantiles(qpdata, err=qp_err[i], quantiles=quantiles, colors=clrs[i], axes=[ax2,ax3], kde=y_kde[i], bw=bw)
axes = format_axes(axes)
if save:
plt.savefig(fitparams['model_id']+'.png', dpi=600)
if fitparams['fit_on']=='subjects' and save:
plt.close('all')
def plot_traces(self, style='HL', ax=None, label_x=True, save=False):
cpals = get_cpals()
sns.set(style='white', font_scale=1.5)
if ax is None:
f, ax = plt.subplots(1, figsize=(5, 5))
tr = self.onset
titl = describe_model(self.depends_on)
if style not in ['HL', 'HML']:
gmu = [ggt.mean(axis=1) for ggt in self.go_traces]
nmu = [ngt.mean(axis=1) for ngt in self.ng_traces]
else:
go_counts = [self.go_traces[i].shape[1]
for i in range(len(self.go_traces))]
ng_counts = [self.ng_traces[i].shape[1]
for i in range(len(self.ng_traces))]
go_lo_ri = randint(0, high=go_counts[0])
go_hi_ri = randint(0, high=go_counts[-1])
ng_lo_ri = randint(0, high=ng_counts[0])
ng_hi_ri = randint(0, high=ng_counts[-1])
ng_hi_ri = np.argmin(self.ng_traces[-1].iloc[-1, :])
glow = self.go_traces[0].iloc[:, go_lo_ri].dropna().values
#pd.concat(self.go_traces[1:3], axis=1).iloc[:,0]
ghi = self.go_traces[-1].iloc[:, go_hi_ri].dropna().values
nglow = self.ng_traces[0].iloc[:, ng_lo_ri].dropna().values
nghi = self.ng_traces[-1].iloc[:, ng_hi_ri].dropna().values
# return nglow, nghi
gmu = [glow, ghi]
nmu = [nglow, nghi]
tr = [tr[0], tr[-1]]
gc = ["#40ac5b", '#10ac1d']
nc = ['#dc3c3c', '#d61b1b']
#gc = cpals['gpal'](len(gmu))
#nc = cpals['rpal'](len(nmu))
gx = [tr[i] + np.arange(len(gmu[i])) * self.dt for i in range(len(gmu))]
nx = [tr[i] + np.arange(len(nmu[i])) * self.dt for i in range(len(nmu))]
ls = ['-', '-']
for i in range(len(gmu)):
ax.plot(gx[i], gmu[i], linestyle=ls[i], lw=1, color=gc[i])
ax.fill_between(gx[i], gmu[i], y2=0, lw=1, color=gc[i], alpha=.25)
for i in range(len(nmu)):
ax.plot(nx[i], nmu[i], linestyle=ls[i], lw=1, color=nc[i])
ax.fill_between(nx[i], nmu[i], y2=0, lw=1, color=nc[i], alpha=.25)
ax.set_ylim(0, self.p['a'].max() * 1.01)
ax.set_xlim(gx[-1].min() * .98, nx[0].max() * 1.05)
if label_x:
ax.set_xlabel('Time', fontsize=26)
ax.set_xticklabels([])
ax.set_yticklabels([])
ax.set_ylabel('$\\theta_{G}$', fontsize=30)
sns.despine()
plt.tight_layout()
if save:
plt.savefig('_'.join([titl, self.decay, 'traces.png']), dpi=300)
plt.savefig('_'.join([titl, self.decay, 'traces.svg']), format='svg', rasterized=True)
return ax
#!/usr/local/bin/env python
from __future__ import division
import sys
from copy import deepcopy
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from radd.tools import colors, analyze
from IPython.display import display, Latex
import warnings
warnings.simplefilter('ignore', np.RankWarning)
warnings.filterwarnings("ignore", module="matplotlib")
cdict = colors.get_cpals('all')
rpal = cdict['rpal']
bpal = cdict['bpal']
gpal = cdict['gpal']
ppal = cdict['ppal']
heat = cdict['heat']
cool = cdict['cool']
slate = cdict['slate']
sns.set(style='darkgrid', rc={'figure.facecolor':'white'}, font_scale=1.2)
def plot_model_fits(y, yhat, fitparams, err=None, palettes=[gpal, bpal], save=False, cdf=True, bw=.01):
""" main plotting function for displaying model fit predictions over data
"""
sns.set(style='darkgrid', rc={'figure.facecolor':'white'}, font_scale=1.5)
# extract model and fit info from fitparams
nlevels = fitparams['nlevels']
ssd, nssd, nss, nss_per, ssd_ix = fitparams.ssd_info
