def plot_data(self,
df,
df_err=[],
fname='plot',
title='title',
xlabel='xlabel',
ylabel='ylabel',
plot_min=0,
plot_max=1,
index_marker_shape=['o', 'v'],
columns_marker_color=['m', 'g'],
columns_title='Column_title',
linestyle='--',
scale_plot=True,
x_scale_factor=1.0,
x_scale_addition=0.0,
plot_on_number_line=False):
fig = plt.figure(dpi=100)
ax = plt.gca()
if plot_on_number_line:
xi = list(df.index.values[1:])
else:
xi = list(range(len(df.index.values[1:])))
plt.errorbar(xi,
df.values[1:],
yerr=0,
capsize=5.0,
label=None,
marker='s',
color='k')
ax.axhline(y=1, xmin=0, xmax=1, linestyle='--', color='k')
plt.xticks(xi, df.index.values[1:] * 2)
ax.set_xlabel(xlabel)
ax.set_ylabel(ylabel)
ax.set_title(title)
if scale_plot:
ax.set_ylim(plot_min, plot_max)
x, y = np.array(xi), df.values[1:].squeeze()
slope, intercept, r_value, p_value, std_err = stats.linregress(x, y)
print('slope = %f, t[4] = %f, p = %f' % (slope, intercept, p_value))
plt.plot(x, y, 'o', label='original data')
plt.plot(x, intercept + slope * x, 'r', label='fitted line')
plt.tight_layout()
plot_figure(fig,
name=fname,
caption='Results for ' + self.experiment_name,
experiment_name=self.experiment_name,
dirname=self.dirname)
plt.close(fig)
def plot_data_as_barplot2(self,
df,
fname='plot',
experiment_name='pelli',
exp_name='LargeLetters',
title='title',
xlabel='xlabel',
ylabel='ylabel',
columns_labels=None,
plot_bar_min=0,
plot_bar_max=40):
base_contrast = df.values[0, 0]
from matplotlib import cm
from matplotlib.ticker import LinearLocator, FormatStrFormatter
if columns_labels is None:
columns_labels = [str(x) for x in df.columns.values]
fig = plt.figure(dpi=100)
ax = plt.gca()
df = df[0:].T.groupby(['num_flank']).mean().T.mean()
df = pd.DataFrame(np.matrix(np.hstack((base_contrast, df.values))).T,
index=['0', '1', '5'],
columns=['rrms'])
df.plot.bar(ax=ax,
style=['-', '-', '--'],
color=['#959acd', '#00b99c', '#a2a5a8'],
rot=0,
capsize=4,
ylim=[plot_bar_min, plot_bar_max],
legend=False)
ax.set_xlabel(xlabel)
ax.set_ylabel(ylabel)
ax.set_title(title)
plt.tight_layout()
plot_figure(fig,
name=fname,
caption='Results for ' + experiment_name + ':' + exp_name,
experiment_name=experiment_name,
dirname=self.dirname)
plt.close(fig)
def plot_data_as_barplot(self,
df,
fname='plot',
title='title',
xlabel='xlabel',
ylabel='ylabel',
target_value=1):
df1 = df.copy()
df1.index = np.array(
[chr(x) for x in np.arange(len(df.index.values)) + 97])
fig = plt.figure(dpi=100)
ax = plt.gca()
df1.plot.bar(ax=ax,
color='k',
rot=0,
capsize=4,
legend=False,
yerr=5.0)
ax.axhline(y=target_value, xmin=0, xmax=1, linestyle='--', color='k')
ax.set_xlabel(xlabel)
ax.set_ylabel(ylabel)
ax.set_title(title)
# if scale_plot:
# ax.set_ylim(plot_min, plot_max)
plt.tight_layout()
plot_figure(fig,
name=fname,
caption='Results for ' + self.experiment_name,
experiment_name=self.experiment_name,
dirname=self.dirname)
plt.close(fig)
def plot_decision_func(self,
df,
df_contrast=None,
target_contrast=0.027,
sigma=0.05,
K=5.0,
zoomed=False,
fname='plot',
experiment_name='pelli',
exp_name='LargeLetters',
title='title',
xlabel='xlabel',
ylabel='ylabel',
index_marker_shape=['o', 'v'],
columns_marker_color=['m', 'g'],
columns_title='Column_title',
scale_range=[0, 1],
upper_limit=0.85,
lower_limit=0.0,
x_scale_factor=1.0,
show_target_as_dash=True,
compute_relative_to_chance=False,
chance=0.0,
compute_error=False,
plot_columns_labels=['0', '1', '2', '3', '4'],
show_legend=False,
decision_prob_label='response',
est_max=0,
transpose_df=True,
use_target_contrast=False,
legend_alpha=0.7,
legend_loc=8,
target_value=False):
if compute_error:
df = pd.DataFrame(df['contrast']).join(
pd.DataFrame(1 - df[decision_prob_label]))
if transpose_df:
df_contrast = df['contrast'].unstack().T
df = df[decision_prob_label].unstack().T
else:
df_contrast = df['contrast'].unstack()
df = df[decision_prob_label].unstack()
if not target_value:
target_value = np.nanmax(df.values[:, 0])
if use_target_contrast:
target_value = target_contrast
from matplotlib import cm
from matplotlib.ticker import LinearLocator, FormatStrFormatter
if df.columns.values.shape[0] > 1:
offset_color = columns_marker_color
else:
offset_color = ['k']
marker_shape = index_marker_shape
marker_color = offset_color[::-1]
if plot_columns_labels is not None:
columns_labels = plot_columns_labels
else:
columns_labels = [str(x) for x in df.columns.values]
x = np.linspace(0, np.nanmax(df_contrast.values), 100000)
y = decision_func(x,
target_contrast=target_contrast,
decision_sigma=sigma,
decision_K=K,
upper_limit=upper_limit,
lower_limit=lower_limit)
if compute_relative_to_chance:
y = ((y - chance) / (upper_limit - chance)) * upper_limit
fig = plt.figure(dpi=100)
ax = plt.gca()
plt.plot(x, y, color='k', linestyle='--')
for j, offset in enumerate(df.columns.values):
c = offset_color[int(np.mod(j, len(offset_color)))]
for i, row in enumerate(df.iterrows()):
s = marker_shape[np.mod(i, len(marker_shape))]
plt.plot(df_contrast.values[i, j],
df.values[i, j],
label=None,
marker=s,
linestyle='None',
markerfacecolor=c,
markeredgecolor=c)
ax = plt.gca()
if zoomed:
ax.set_xlim(np.nanmin(df_contrast.values),
np.nanmax(df_contrast.values))
ax.set_ylim(np.nanmin(df.values), np.nanmax(df.values))
else:
if show_target_as_dash:
ax.axhline(y=target_value,
xmin=0,
xmax=1,
linestyle=':',
color='k')
ax.set_xlim(0.9 * target_contrast, np.nanmax(df_contrast.values))
handles = []
for j, offset in enumerate(df.columns.values.tolist()):
c = offset_color[int(np.mod(j, len(offset_color)))]
l = columns_labels[int(np.mod(j, len(columns_labels)))]
patch = mpatches.Patch(color=c, label=l)
handles.append(patch)
if len(columns_labels) > 1:
first_legend = plt.legend(title=columns_title,
handles=handles,
loc=4)
plt.gca().add_artist(first_legend)
handles = []
for j, flank_dist in enumerate(
np.round(
np.array(df.index.values.tolist()) * x_scale_factor, 4)):
s = marker_shape[int(np.mod(j, len(marker_shape)))]
marker = mlines.Line2D([], [],
color='k',
marker=s,
linestyle='None',
markersize=10,
label=str(flank_dist))
handles.append(marker)
if show_legend:
plt.legend(title='Flank Dist.',
handles=handles,
framealpha=legend_alpha,
loc=legend_loc)
ax.set_xlabel(xlabel)
ax.set_ylabel(ylabel)
ax.set_title(title)
plt.tight_layout()
plot_figure(fig,
name=fname,
caption='Results for ' + experiment_name + ':' + exp_name,
experiment_name=experiment_name,
dirname=self.dirname)
plt.close(fig)
def plot_data2(self,
df,
fname='plot',
experiment_name='pelli',
exp_name='LargeLetters',
title='title',
xlabel='xlabel',
ylabel='ylabel',
plot_min=0,
plot_max=100,
index_marker_shape=['o', 'v'],
columns_marker_color=['m', 'g'],
columns_title='Column_title',
linestyle='--',
scale_plot=True,
show_target_as_dash=True,
x_scale_factor=1.0,
x_scale_addition=0.0,
target_value=None,
**params):
if target_value is None:
target_value = df.iloc[0].values[0]
df = df[0:]
from matplotlib import cm
from matplotlib.ticker import LinearLocator, FormatStrFormatter
if df.columns.values.shape[0] > 1:
offset_color = columns_marker_color
else:
offset_color = ['k']
marker_shape = index_marker_shape
marker_color = offset_color[::-1]
if 'plot_columns_labels' in params:
columns_labels = params['plot_columns_labels']
else:
columns_labels = [str(x) for x in df.columns.values]
fig = plt.figure(dpi=100)
ax = plt.gca()
for j, offset in enumerate(df.columns.values):
c = offset_color[int(np.mod(j, len(offset_color)))]
for i, row in enumerate(df.iterrows()):
s = marker_shape[np.mod(i, len(marker_shape))]
plt.plot((df.index[i] + x_scale_addition) * x_scale_factor,
df.values[i, j],
label=None,
marker=s,
linestyle=linestyle,
markerfacecolor=c,
markeredgecolor=c)
plt.plot((df.index.values + x_scale_addition) * x_scale_factor,
df.values[:, j],
label=None,
linestyle=linestyle,
color=c)
if show_target_as_dash:
ax.axhline(y=target_value,
xmin=0,
xmax=1,
linestyle=':',
color='k')
ax.set_xlabel(xlabel)
ax.set_ylabel(ylabel)
ax.set_title(title)
handles = []
for j, offset in enumerate(df.columns.values.tolist()):
c = offset_color[int(np.mod(j, len(offset_color)))]
l = columns_labels[int(np.mod(j, len(columns_labels)))]
patch = mpatches.Patch(color=c, label=l)
handles.append(patch)
if len(columns_labels) > 1:
first_legend = plt.legend(title=columns_title,
handles=handles,
loc=4)
plt.gca().add_artist(first_legend)
handles = []
for j, flank_dist in enumerate(
np.round(
np.array(df.index.values.tolist()) * x_scale_factor, 4)):
s = marker_shape[int(np.mod(j, len(marker_shape)))]
marker = mlines.Line2D([], [],
color='k',
marker=s,
linestyle='None',
markersize=10,
label=str(flank_dist))
handles.append(marker)
plt.legend(title='Flank Dist.', handles=handles, framealpha=0.7, loc=8)
if scale_plot:
ax.set_ylim(plot_min, plot_max)
plt.tight_layout()
plot_figure(fig,
name=fname,
caption='Results for ' + experiment_name + ':' + exp_name,
experiment_name=experiment_name,
dirname=self.dirname)
plt.close(fig)