def main(rand_nodes=False,
script_name=__file__.split('/')[-1][0:-3],
from_disk=1):
k = get_kwargs_builder()
dinfo, dn = get_setup(**{'rand_nodes': rand_nodes})
# dinfo, dn = get_setup()
ds = get_storages(script_name, dn.keys(), dinfo)
d = {}
d.update(optimize('opt_rate', dn, [from_disk] * 5, ds, **{'x0': 200.0}))
for key in sorted(dn['hist'].keys()):
net = dn['hist'][key]
set_optimization_val(d['opt_rate'][net.get_name()], [net])
d.update(
run('hist', dn, [from_disk] * 5, ds, 'mean_rates',
**{'t_start': k['start_rec']}))
fig, axs = pl.get_figure(n_rows=1,
n_cols=1,
w=1000.0,
h=800.0,
fontsize=16)
show_opt_hist(d, axs, '$GPe_{+d}^{TA}$')
ds['fig'].save_fig(fig)
if DISPLAY: pylab.show()
def main(rand_nodes=False,
script_name= __file__.split('/')[-1][0:-3],
from_disk=0):
k=get_kwargs_builder()
dinfo, dn = get_setup(**{'rand_nodes':rand_nodes})
dn=modify(dn)
ds = get_storages(script_name, dn.keys(), dinfo)
dstim={}
dstim ['IV']=map(float, range(-300,300,100)) #curr
dstim ['IF']=map(float, range(0,500,100)) #curr
dstim ['FF']=map(float, range(0,1500,100)) #rate
d={}
d.update(run_XX('IV', dn, [from_disk]*4, ds, dstim))
d.update(run_XX('IF', dn, [from_disk]*4, ds, dstim))
d.update(run_XX('FF', dn, [from_disk]*4, ds, dstim))
d.update(optimize('opt_rate', dn, [from_disk]*1, ds, **{ 'x0':200.0}))
set_optimization_val(d['opt_rate']['Net_0'], dn['hist'])
d.update(run('hist', dn, [from_disk]*2, ds, 'mean_rates',
**{'t_start':k['start_rec']}))
fig, axs=pl.get_figure(n_rows=2, n_cols=2, w=1000.0, h=800.0, fontsize=16)
show(dstim, d, axs, NAMES)
ds['fig'].save_fig(fig)
if not os.environ.get('DISPLAY'): pylab.show()
def show_3d(d,attr,**k):
models=['SN']
res=k.get('resolution')
titles=k.get('titles')
n=len(models)
m=len(d.keys())
# attr='mean_rate_slices'
fig, axs=ps.get_figure(n_rows=m, n_cols=1, w=500.0, h=800.0, fontsize=12,
projection='3d')
i=0
for model in models:
alpha=0.8
dcm={'Net_0':'jet',
'Net_1':'coolwarm',}
for key in sorted(d.keys()):
obj0=d[key]['set_0'][model][attr]
obj1=d[key]['set_1'][model][attr]
args=[obj0.x_set, obj1.x_set,
# obj1.y-obj0.y,
numpy.mean(obj1.y_raw_data-obj0.y_raw_data, axis=0),
numpy.std(obj1.y_raw_data-obj0.y_raw_data, axis=0)]
for j, arg in enumerate(args):
arg.shape
args[j]=numpy.reshape(arg, [res,res])
x,y,z, z_std=args
axs[i].plot_surface(x, y, z, cmap='coolwarm', rstride=1, cstride=1,
linewidth=0,
shade=True,
alpha=alpha,
antialiased=False,
vmin=-40,
vmax=40)
axs[i].set_zlim([-40, 40])
# axs[i].set_zlabel('SNr firing rate (Hz)')
# axs[i].set_xlabel('CTX increase A1 (proportion)')
# axs[i].set_ylabel('CTX increase A2 (proportion)')
axs[i].set_title(titles[i])
# axs[i+1].plot_surface(x, y, z_std, cmap='coolwarm', rstride=1, cstride=1,
# linewidth=0,
# shade=True,
# alpha=alpha,
# antialiased=False)
# alpha-=0.3
i+=1
# pylab.show()
# print v
for ax in axs:
ax.view_init(elev=15)
return fig
def show_hist(name, d, models=['M1', 'M2', 'FS', 'GA', 'GI', 'ST', 'SN'], **k):
fig, axs = ps.get_figure(n_rows=len(models),
n_cols=1,
w=1000.0,
h=800.0,
fontsize=10)
labels = k.pop('labels', sorted(d.keys()))
colors = misc.make_N_colors('jet', len(labels))
linestyles = ['solid'] * len(labels)
linewidth = [2.0] * len(labels)
j = 0
for key in sorted(d.keys()):
v = d[key]
# axs[0].serunt_title(k)
for i, model in enumerate(models):
if 'spike_stastistic' in v[model]:
st = v[model]['spike_statistic']
st.rates = {
'mean': round(st.rates['mean'], 2),
'std': round(st.rates['std'], 2),
'CV': round(st.rates['CV'], 2)
}
s = str(st.rates)
else:
s = ''
# print s
# print labels[j]
k.update({
'label': (model + ' ' + labels[j] + ' ' + s),
'histtype': 'step',
'linestyle': linestyles[j],
'color': colors[j],
'linewidth': linewidth[j]
})
# print k
h = v[model][name].hist(ax=axs[i], **k)
ylim = list(axs[i].get_ylim())
ylim[0] = 0.0
axs[i].set_ylim(ylim)
axs[i].legend_box_to_line()
j += 1
# import pylab
# pylab.show()
return fig, axs
def plot_optimal(size):
fig, axs=ps.get_figure(n_rows=1, n_cols=1, w=400.0*0.8, h=250.0*0.8, fontsize=24)
x,y,z=get_optimal(size)
im=axs[0].pcolor(x, y, z, cmap='coolwarm', vmin=-40, vmax=40)
#
box = axs[0].get_position()
axs[0].set_position([box.x0*1.05, box.y0, box.width*0.85, box.height])
axs[0].set_xlabel('Action 1')
axs[0].set_ylabel('Action 2')
axs[0].my_set_no_ticks(xticks=3,yticks=3, )
return fig
def plot_raw(d, d_keys, attr='mean_coherence'):
for key in sorted(d.keys()):
fig, axs = ps.get_figure(n_rows=1,
n_cols=4,
w=800.0 * 0.65 * 2,
h=300.0 * 0.65 * 2,
fontsize=8,
frame_hight_y=0.5,
frame_hight_x=0.9,
title_fontsize=8)
for i, model in enumerate(models):
ax = axs[i]
ax.plot(d[key]['Net_0'][model][attr], 'b')
ax.plot(d[key]['Net_1'][model][attr], 'r')
ax.set_title(d_keys[key] + ' ' + model)
def plot_raw(d, d_keys, attr='mean_coherence'):
for key in sorted(d.keys()):
fig, axs=ps.get_figure(n_rows=1,
n_cols=4,
w=800.0*0.65*2,
h=300.0*0.65*2,
fontsize=8,
frame_hight_y=0.5,
frame_hight_x=0.9,
title_fontsize=8)
for i, model in enumerate(models):
ax=axs[i]
ax.plot(d[key]['Net_0'][model][attr], 'b')
ax.plot(d[key]['Net_1'][model][attr], 'r')
ax.set_title(d_keys[key]+' '+model)
def show_neuron_numbers(d, models, **k):
attr='firing_rate'
linestyle=['-','--']
labels=k.pop('labels', models)
colors=misc.make_N_colors('jet', max(len(labels), 1))
fig, axs=ps.get_figure(n_rows=1, n_cols=1, w=500.0, h=500.0, fontsize=10)
l_ids=[]
ax=axs[0]
for k, model in enumerate(models):
max_set=0
for j, name in enumerate(sorted([d.keys()[0]])):
v=d[name]
sets=[s for s in sorted(v.keys()) if s[0:3]=='set']
ids=0
for i, _set in enumerate(sets):
if not model in v[_set].keys():
break
obj=v[_set][model][attr]
ids+=len(obj.ids)
if max_set<=i:
max_set+=1
l_ids.append(ids)
obj=Data_bar(**{'y':numpy.array(l_ids)})
obj.bar(ax)
ax.set_xticklabels(labels)
return fig
def show_bulk(d, models, attr, **k):
# attr='mean_rate_slices'
linestyle=['-','--']
labels=k.pop('labels', sorted(d.keys()))
colors=misc.make_N_colors('jet', max(len(labels), 1))
fig, axs=ps.get_figure(n_rows=7, n_cols=1, w=1200.0, h=800.0, fontsize=10)
res=k.pop('res')
i=0
xticks=[]
for j in range(res):
for k in range(res):
if j==k:
xticks.append(i)
i+=1
for k, model in enumerate(models):
ax=axs[k]
# max_set=0
for j, name in enumerate(sorted(d.keys())):
v=d[name]
sets=[s for s in v.keys() if s[0:3]=='set']
for i, _set in enumerate(sets):
if not model in v[_set].keys():
break
obj=v[_set][model][attr]
obj.plot(ax, **{'color':colors[j],
'linestyle':linestyle[i],
'label':labels[j]
})
# if max_set<=i:
# max_set+=1
max_set=len(sets)
ax.set_ylabel(model+' (spikes/s)')
ax.set_xticks(xticks)
#Get artists and labels for legend and chose which ones to display
if k!=0:
continue
handles, _labels = ax.get_legend_handles_labels()
display = range(0,len(d.keys())*max_set, max_set)
#Create custom artists
linetype_labels=[]
linetype_handles=[]
for i in range(max_set):
linetype_handles.append(pylab.Line2D((0,1),(0,0),
color='k',
linestyle=linestyle[i]))
linetype_labels.append('Action '+str(i))
#Create legend from custom artist/label lists
ax.legend([handle for i,handle in enumerate(handles)
if i in display]+linetype_handles,
[label for i,label in enumerate(_labels)
if i in display]+linetype_labels, bbox_to_anchor=(1.22, 1.))
return fig
'''
Created on May 8, 2014
@author: lindahlm
'''
import pylab
import numpy
import pprint
import core.plot_settings as pl
pp = pprint.pprint
_, axs = pl.get_figure(n_rows=1, n_cols=1, w=1000.0, h=800.0, fontsize=16)
ax = axs[0] #pylab.subplot(111)
r = numpy.random.random(100)
ax.hist(r, **{'histtype': 'step', 'label': 'test 1'})
r = numpy.random.random(100)
ax.hist(r, **{'histtype': 'step', 'label': 'test 2', 'linestyle': 'dashed'})
h, labels = ax.get_legend_handles_labels()
artist = []
for hh in h:
color = hh._edgecolor
linestyle = hh._linestyle
obj = pylab.Line2D((0, 1), (0, 0), color=color, linestyle=linestyle)
artist.append(obj)
ax.legend(artist, labels)
'''
Created on Jul 27, 2014
@author: mikael
'''
import numpy
import core.plot_settings as ps
import pylab
y = numpy.array([
[1041, 2422, 2009, 1980],
[1061, 2387, 1968, 1856],
[1014, 2167, 1960, 1788],
])
x = numpy.array([8, 6, 4])
fig, axs = ps.get_figure(n_rows=1,
n_cols=1,
w=500.0,
h=500.0,
fontsize=24,
linewidth=4)
ax = axs[0]
for i in range(4):
ax.plot(x, y[:, i] / float(numpy.max(y)))
ax.my_set_no_ticks(xticks=3)
ax.set_ylabel('Rel performance')
ax.set_xlabel('MSN activated (%)')
pylab.show()
'''
Created on Jul 27, 2014
@author: mikael
'''
import numpy
import core.plot_settings as ps
import pylab
y=numpy.array([[1041, 2422, 2009, 1980],
[1061, 2387, 1968, 1856 ],
[1014, 2167, 1960, 1788],
])
x=numpy.array([8,6,4])
fig, axs=ps.get_figure(n_rows=1, n_cols=1, w=500.0, h=500.0,
fontsize=24, linewidth=4)
ax=axs[0]
for i in range(4):
ax.plot(x, y[:,i]/float(numpy.max(y)))
ax.my_set_no_ticks(xticks=3)
ax.set_ylabel('Rel performance')
ax.set_xlabel('MSN activated (%)')
pylab.show()
'''
Created on May 8, 2014
@author: lindahlm
'''
import pylab
import numpy
import pprint
import core.plot_settings as pl
pp=pprint.pprint
_, axs=pl.get_figure(n_rows=1, n_cols=1, w=1000.0, h=800.0, fontsize=16)
ax=axs[0]#pylab.subplot(111)
r=numpy.random.random(100)
ax.hist(r, **{'histtype':'step', 'label':'test 1'})
r=numpy.random.random(100)
ax.hist(r, **{'histtype':'step', 'label':'test 2', 'linestyle':'dashed'})
h, labels=ax.get_legend_handles_labels()
artist=[]
for hh in h:
color=hh._edgecolor
linestyle=hh._linestyle
obj=pylab.Line2D((0,1),(0,0), color=color, linestyle=linestyle)
artist.append(obj)