def plot_bar_graph(series,
series_colors,
series_labels=None,
series_color_emphasis=None,
series_errs=None,
series_err_colors=None,
series_padding=0.0,
series_use_labels=False,
series_style=None,
plot_xlabel=None,
plot_ylabel=None,
plot_yinvert=False,
plot_title=None,
category_labels=None,
category_ticks=True,
category_padding=0.25,
barwidth=0.35,
xpadding=0.,
stacked=False,
fontsize=8,
legend_fontsize=8,
legend_location='best',
savefile=None,
savefile_size=(3.4, 1.5),
horizontal=False,
show_plot=True):
"""
Plot a bar graph
@param series List of data for each series - each of these will be plotted in a different color
Each series should have the same number of elements.
@param series_labels List of labels for each series - same length as series
@param series_colors List of colors for each series - same length as series
@param series_color_emphasis List of booleans, one for each series, indicating whether the series
color should be bold - if None no series is bold
@param series_errs The error values for each series - if None no error bars are plotted
@param series_err_colors The colors for the error bars for each series, if None black is used
@param plot_xlabel The label for the x-axis - if None no label is printed
@param plot_ylabel The label for the y-axis - if None no label is printed
@param plot_title A title for the plot - if None no title is printed
@param category_labels The labels for each particular category in the histogram
@param category_ticks If true, also place a tick at each category
@param category_padding Fraction of barwidth (0 - 1) - distance between categories
@param barwidth The width of each bar
@param xpadding The padding between the first bar and the left axis and the last bar and the right axis
@param stacked If true, stack the data in each category
@param fontsize The size of font for all labels
@param legend_fontsize The size of font for the legend labels
@param legend_location The location of the legend, if None no legend is included
@param savefile The path to save the plot to, if None plot is not saved
@param savefile_size The size of the saved plot
@param horizontal Plot bars horizontally instead of vertically
@param show_plot If True, display the plot on the screen via a call to plt.show()
@return fig, ax The figure and axis the plot was created in
"""
# Validate
if series is None or len(series) == 0:
raise ValueError('No data series')
num_series = len(series)
if len(series_colors) != num_series:
raise ValueError('You must define a color for every series')
if series_labels is None:
series_labels = [None for l in range(num_series)]
if len(series_labels) != num_series:
raise ValueError('You must define a label for every series')
if series_errs is None:
series_errs = [None for _ in series]
if len(series_errs) != num_series:
raise ValueError(
'series_errs is not None. Must provide error value for every series.'
)
if series_err_colors is None:
series_err_colors = ['black' for _ in series]
if len(series_err_colors) != num_series:
raise ValueError('Must provide an error bar color for every series')
if series_color_emphasis is None:
series_color_emphasis = [False for _ in series]
if len(series_color_emphasis) != num_series:
raise ValueError(
'The emphasis list must be the same length as the series_colors list'
)
if series_use_labels and len(series[0]) > 1:
raise ValueError('Only series containing one category may be labeled.')
if series_style is None:
series_style = [dict() for _ in series]
fig, ax = plt.subplots()
plot_utils.configure_fonts(fontsize=fontsize,
legend_fontsize=legend_fontsize)
if plot_yinvert:
ax.invert_yaxis()
spacing = category_padding * barwidth
num_categories = len(series[0])
category_width = spacing + num_series * barwidth
if stacked:
category_width = spacing + barwidth
index = numpy.array(range(num_categories)) * category_width + xpadding
for idx in range(num_series):
offset = idx * (barwidth + series_padding)
if stacked:
offset = 0.
style = dict(
label=series_labels[idx],
color=colors.get_plot_color(series_colors[idx],
emphasis=series_color_emphasis[idx]),
ecolor=colors.get_plot_color(series_err_colors[idx]),
linewidth=0,
)
style.update(series_style[idx])
if horizontal:
r = ax.barh(
bottom=index + offset,
#left = index + offset,
height=barwidth, #series[idx],
width=series[idx], # barwidth,
xerr=series_errs[idx],
**style)
else:
r = ax.bar(left=index + offset,
height=series[idx],
width=barwidth,
yerr=series_errs[idx],
**style)
# Label the plot
if plot_ylabel is not None:
ax.set_ylabel(plot_ylabel)
if plot_xlabel is not None:
ax.set_xlabel(plot_xlabel)
if plot_title is not None:
ax.set_title(plot_title)
# Add category tick marks
if series_use_labels:
indices = numpy.arange(num_series)
ticks = xpadding + indices * (barwidth +
series_padding) + 0.5 * barwidth
if horizontal:
ax.set_yticks(ticks)
else:
ax.set_xticks(ticks)
if series_labels is not None:
if horizontal:
ax.set_yticklabels(series_labels)
else:
ax.set_xticklabels(series_labels)
elif category_ticks:
if not stacked:
ticks = index + (num_series / 2.) * barwidth
else:
ticks = index + .5 * barwidth
if horizontal:
ax.set_yticks(ticks)
else:
ax.set_xticks(ticks)
if category_labels is not None:
if horizontal:
ax.set_yticklabels(category_labels)
else:
ax.set_xticklabels(category_labels)
else:
if horizontal:
ax.set_yticks([])
else:
ax.set_xticks([])
# Set the x-axis limits
lims = [
0, 2 * xpadding + num_categories *
(category_width + (num_series - 1) * series_padding) - spacing
]
if horizontal:
ax.set_ylim(lims)
else:
ax.set_xlim(lims)
# Legend
if legend_location is not None:
ax.legend(loc=legend_location, frameon=False)
# Make the axis pretty
plot_utils.simplify_axis(ax)
# Save the file
if savefile is not None:
plot_utils.output(fig,
savefile,
savefile_size,
fontsize=fontsize,
legend_fontsize=legend_fontsize)
# Show
if show_plot:
plt.show()
return fig, ax
def plot(series,
series_colors,
series_color_emphasis=None,
series_labels=None,
series_errs=None,
series_err_colors=None,
fill_error=True,
plot_xlabel=None,
plot_xlim=None,
plot_ylabel=None,
plot_ylim=None,
plot_title=None,
fontsize=8,
legend_fontsize=8,
linewidth=2,
legend_location='best',
savefile=None,
savefile_size=(3.4, 1.5),
show_plot=True,
jitter_x=-1.,
jitter_y=-1.,
jitter_alpha=1.,
marker_size=5,
x_scale='linear',
y_scale='linear',
plot_markers=None,
line_styles=None,
mark_every=1):
"""
Plot yvals as a function of xvals
@param series List of (xvals, yvals) for each series- each of these will be plotted in a different color
@param series_labels List of labels for each series - same length as series
@param series_colors List of colors for each series - same length as series
@param series_color_emphasis List of booleans, one for each series, indicating whether the series
color should be bold - if None no series is bold
@param series_errs The error values for each series - if None no error bars are plotted
@param series_err_colors The colors for the error bars for each series, if None black is used
- @param fill_error If true, shade the area to show error bars, if not, draw actual error bars
@param plot_xlabel The label for the x-axis - if None no label is printed
@param plot_xlim The limits for the x-axis - if None these limits are selected automatically
@param plot_ylabel The label for the y-axis - if None no label is printed
@param plot_ylim The limits for the y-axis - if None these limits are selected automatically
@param plot_title A title for the plot - if None no title is printed
@param fontsize The size of font for all labels
@param legend_fontsize The size of font for the legend labels
@param linewidth The width of the lines in the plot
@param legend_location The location of the legend, if None no legend is included
@param savefile The path to save the plot to, if None plot is not saved
@param savefile_size The size of the saved plot
@param show_plot If True, display the plot on the screen via a call to plt.show()
@param x_scale set to log or linear for the x axis
@param y_scale set to log or linear for the y axis
@param plot_markers if not None, a list of line marker symbols
@param line_styles if not None, a list of line style symbols
@param marker_size
@param jitter_x the scale of the normal distribution for jitter along x direction, if -1., then no jitter along x
@param jitter_y the scale of the normal distribution for jitter along y direction, if -1., then no jitter along y
@param jitter_alpha the alpha value for jittered scatter points
@return fig, ax The figure and axis the plot was created in
"""
# Validate
if series is None or len(series) == 0:
raise ValueError('No data series')
num_series = len(series)
if len(series_colors) != num_series:
raise ValueError('You must define a color for every series')
if series_labels is None:
series_labels = [None for s in series]
if len(series_labels) != num_series:
raise ValueError('You must define a label for every series')
if series_errs is None:
series_errs = [None for s in series]
if len(series_errs) != num_series:
raise ValueError(
'series_errs is not None. Must provide error value for every series.'
)
if series_err_colors is None:
series_err_colors = ['black' for s in series]
if len(series_err_colors) != num_series:
raise ValueError('Must provide an error bar color for every series')
if series_color_emphasis is None:
series_color_emphasis = [False for s in series]
if len(series_color_emphasis) != num_series:
raise ValueError(
'The emphasis list must be the same length as the series_colors list'
)
if plot_markers is None:
plot_markers = ['None' for s in series]
if len(plot_markers) != num_series:
raise ValueError('The marker list must contain all series')
if line_styles is None:
line_styles = ['-' for s in series]
if len(line_styles) != num_series:
raise ValueError('The line style list must contain all series')
fig, ax = plt.subplots()
plot_utils.configure_fonts(fontsize=fontsize,
legend_fontsize=legend_fontsize)
num_series = len(series)
for idx in range(num_series):
xvals = series[idx][0]
yvals = numpy.array(series[idx][1])
# by Shen Li
if jitter_x != -1. and jitter_y == -1.:
# http://nbviewer.jupyter.org/gist/fonnesbeck/5850463
# Add some random "jitter" to the x-axis
xvals = numpy.random.normal(xvals, jitter_x, size=len(yvals))
elif jitter_x == -1. and jitter_y != -1.:
# http://nbviewer.jupyter.org/gist/fonnesbeck/5850463
# Add some random "jitter" to the x-axis
yvals = numpy.random.normal(yvals, jitter_y, size=len(xvals))
elif jitter_x != -1. and jitter_y != -1.:
xs = list(numpy.random.normal(xvals, jitter_x, size=len(yvals)))
ys = list(numpy.random.normal(yvals, jitter_y, size=len(xvals)))
xvals = xs
yvals = ys
r = ax.plot(xvals,
yvals,
label=series_labels[idx],
color=colors.get_plot_color(
series_colors[idx],
emphasis=series_color_emphasis[idx]),
marker=plot_markers[idx],
linestyle=line_styles[idx],
lw=linewidth,
markersize=marker_size,
markevery=mark_every,
alpha=jitter_alpha)
errs = series_errs[idx]
if errs is not None:
shade_color = colors.get_plot_color(color=series_err_colors[idx])
if fill_error:
plot_utils.shaded_error(ax,
xvals,
yvals,
numpy.array(errs),
color=shade_color)
else:
ax.errorbar(xvals,
yvals,
yerr=errs,
linestyle='None',
ecolor=shade_color)
# Label the plot
if plot_ylabel is not None:
ax.set_ylabel(plot_ylabel)
if plot_xlabel is not None:
ax.set_xlabel(plot_xlabel)
if plot_title is not None:
ax.set_title(plot_title)
if plot_xlim is not None:
ax.set_xlim(plot_xlim)
if plot_ylim is not None:
ax.set_ylim(plot_ylim)
ax.set_yscale(y_scale)
ax.set_xscale(x_scale)
# Legend
if legend_location is not None:
ax.legend(loc=legend_location, frameon=False)
# Make the axis pretty
plot_utils.simplify_axis(ax)
# Save the file
if savefile is not None:
plot_utils.output(fig,
savefile,
savefile_size,
fontsize=fontsize,
legend_fontsize=legend_fontsize)
# Show
if show_plot:
plt.show()
return fig, ax
def plot_hexbin(series,
series_labels = None,
plot_xlabel = None,
plot_xlim = None,
plot_ylabel = None,
plot_ylim = None,
plot_title = None,
fontsize=8,
vmax = None,
vmin = None,
linewidths = None,
savefile = None,
savefile_size = (3.4, 1.5),
show_plot = True,
gridsize = 20,
series_C = None,
reduce_C_function = numpy.mean,
color_map = cm.jet,
x_scale = 'linear',
y_scale = 'linear',
mincnt = None,
series_edgecolors = None,
marginals = False):
"""
Plot yvals as a function of xvals
@param series List of (xvals, yvals) for each series- each of these will be plotted in a different color
@param plot_xlabel The label for the x-axis - if None no label is printed
@param plot_xlim The limits for the x-axis - if None these limits are selected automatically
@param plot_ylabel The label for the y-axis - if None no label is printed
@param plot_ylim The limits for the y-axis - if None these limits are selected automatically
@param plot_title A title for the plot - if None no title is printed
@param fontsize The size of font for all labels
@param linewidths The width of the lines in the plot
@param savefile The path to save the plot to, if None plot is not saved
@param savefile_size The size of the saved plot
@param show_plot If True, display the plot on the screen via a call to plt.show()
@param x_scale set to log or linear for the x axis
@param y_scale set to log or linear for the y axis
@param cmap the color of hexbin
@param series_C In hexbin, 'C' is optional--it maps values to x-y coordinates;
if 'C' is None (default) then the result is a pure 2D histogram
@param reduce_C_function
Could be numpy.mean, .sum, .count_nonzero, .max
Function of one argument that reduces all the values in a bin to a single number
@param mincnt If not None, only display cells with more than mincnt number of points in the cell
@param gridsize how many hexagons in one line
@return fig, ax The figure and axis the plot was created in
"""
# Validate
if series is None or len(series) == 0:
raise ValueError('No data series')
num_series = len(series)
if series_labels is None:
series_labels = [None for s in series]
if len(series_labels) != num_series:
raise ValueError('You must define a label for every series')
fig, ax = plt.subplots()
plot_utils.configure_fonts(fontsize=fontsize)
xvals = series[0]
yvals = numpy.array(series[1])
if series_C == None:
c = None
else:
c = series_C
if series_edgecolors == None:
edgecolors = None
else:
edgecolors = colors.get_plot_color(series_edgecolors)
#http://stackoverflow.com/questions/2369492/generate-a-heatmap-in-matplotlib-using-a-scatter-data-set/2371812#2371812
# http://matplotlib.org/examples/color/colormaps_reference.html
# if 'bins=None', then color of each hexagon corresponds directly to its count
# 'C' is optional--it maps values to x-y coordinates; if 'C' is None (default) then
# the result is a pure 2D histogram
# hb = plt.hexbin(x, y, C=None, gridsize=gridsize, marginals=True, cmap=cm.jet, bins=None)
# hb = plt.hexbin(x, y, C=None, marginals=True, cmap=plt.get_cmap('YlOrBr'), bins=None)
# r = ax.hexbin(xvals, yvals, C=None, marginals=True, cmap=plt.get_cmap('YlOrBr'), bins=None)
if plot_xlim is None or plot_ylim is None:
plot_xlim = [min(xvals),max(xvals)]
plot_ylim = [min(yvals),max(yvals)]
if c == None:
im = ax.hexbin(xvals, yvals,
C=None,
cmap=color_map,
bins=None,
gridsize = gridsize,
xscale = 'linear',
yscale = 'linear',
norm=None,
vmin=vmin,
vmax=vmax,
alpha=None,
linewidths=linewidths,
edgecolors=edgecolors,
extent=[plot_xlim[0], plot_xlim[1], plot_ylim[0], plot_ylim[1]],
mincnt=None,
marginals=marginals)
else:
im = ax.hexbin(xvals, yvals,
C=c,
cmap=color_map,
bins=None,
gridsize = gridsize,
xscale = 'linear',
yscale = 'linear',
norm=None,
vmin=vmin,
vmax=vmax,
alpha=None,
linewidths=linewidths,
edgecolors=edgecolors,
reduce_C_function = reduce_C_function,
mincnt=None,
marginals=marginals)
cb = fig.colorbar(im,ax=ax)
cb.set_label('count')
# plt.axis([min(x)-1., max(x)+1., min(y)-0.1, max(y)+0.1])
plt.axis([plot_xlim[0], plot_xlim[1], plot_ylim[0], plot_ylim[1]])
if plot_ylabel is not None:
plt.ylabel(plot_ylabel)
if plot_xlabel is not None:
plt.xlabel(plot_xlabel)
if plot_title is not None:
plt.title(plot_title)
# Make the axis pretty
plot_utils.simplify_axis(ax)
# Save the file
if savefile is not None:
plot_utils.output(fig, savefile, savefile_size,
fontsize=fontsize)
# Show
if show_plot:
plt.show()
return im
plot_labels = [f[1] for f in features]
from ss_plotting.make_plots import plot_bar_graph
fig, ax = plot_bar_graph(series,
series_labels=series_labels,
series_errs=series_errs,
series_colors=['grey'],
category_labels=[f[1] for f in features],
category_rotation=45,
xpadding=0.3,
show_plot=False,
simplify=False)
ax.set_ylim([0.70, 0.85])
xlim = ax.get_xlim()
ax.set_xlim([xlim[0], xlim[1] - 0.5])
from ss_plotting import plot_utils
plot_utils.simplify_axis(ax)
plt.savefig("Feature_comparison.eps", format='eps', dpi=1000)
import IPython
IPython.embed()
# ALl, 3F + RoC, 3F/P + RoC, 3F/T + ROC, Z-force + RoC
# def tapo_plot_confusion_matrix(cm, classes, modality, path, normalize=False, title='Confusion matrix', cmap=pp.cm.Blues):
# """ This function prints and plots the confusion matrix.
# Normalization can be applied by setting `normalize=True`.
# """
# if normalize:
# cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]
# print("Normalized confusion matrix")
# else:
# cm = cm.astype('int')
def plot_stacked_vertical_bar_graph(series,
series_colors,
series_labels=None,
series_color_emphasis=None,
series_errs=None,
series_err_colors=None,
series_padding=0.0,
series_use_labels=False,
series_style=None,
plot_xlabel=None,
plot_ylabel=None,
plot_yinvert=False,
plot_title=None,
category_labels=None,
category_ticks=True,
category_padding=0.25,
barwidth=0.35,
xpadding=0.,
fontsize=8,
legend_fontsize=8,
legend_location='best',
legend_title=None,
legend_reverse_label=False,
legend_bbox_to_anchor=None,
legend_pos_rel_plot=None,
legend_title_pos=None,
legend_labelspacing=None,
savefile=None,
savefile_size=(3.4, 1.5),
show_plot=True):
"""
Plot a bar graph
@param series List of data for each series - each of these will be plotted in a different color
Each series should have the same number of elements.
@param series_labels List of labels for each series - same length as series
@param series_colors List of colors for each series - same length as series
@param series_color_emphasis List of booleans, one for each series, indicating whether the series
color should be bold - if None no series is bold
@param series_errs The error values for each series - if None no error bars are plotted
@param series_err_colors The colors for the error bars for each series, if None black is used
@param plot_xlabel The label for the x-axis - if None no label is printed
@param plot_ylabel The label for the y-axis - if None no label is printed
@param plot_title A title for the plot - if None no title is printed
@param category_labels The labels for each particular category in the histogram
@param category_ticks If true, also place a tick at each category
@param category_padding Fraction of barwidth (0 - 1) - distance between categories
@param barwidth The width of each bar
@param xpadding The padding between the first bar and the left axis and the last bar and the right axis
@param fontsize The size of font for all labels
@param legend_fontsize The size of font for the legend labels
@param legend_location The location of the legend, if None no legend is included
@param savefile The path to save the plot to, if None plot is not saved
@param savefile_size The size of the saved plot
@param show_plot If True, display the plot on the screen via a call to plt.show()
@return fig, ax The figure and axis the plot was created in
"""
# Validate
if series is None or len(series) == 0:
raise ValueError('No data series')
num_series = len(series)
if len(series_colors) != num_series:
raise ValueError('You must define a color for every series')
if series_labels is None:
series_labels = [None for l in range(num_series)]
if len(series_labels) != num_series:
raise ValueError('You must define a label for every series')
if series_errs is None:
series_errs = [None for _ in series]
if len(series_errs) != num_series:
raise ValueError(
'series_errs is not None. Must provide error value for every series.'
)
if series_err_colors is None:
series_err_colors = ['black' for _ in series]
if len(series_err_colors) != num_series:
raise ValueError('Must provide an error bar color for every series')
if series_color_emphasis is None:
series_color_emphasis = [False for _ in series]
if len(series_color_emphasis) != num_series:
raise ValueError(
'The emphasis list must be the same length as the series_colors list'
)
if series_use_labels and len(series[0]) > 1:
raise ValueError('Only series containing one category may be labeled.')
if series_style is None:
series_style = [dict() for _ in series]
fig, ax = plt.subplots()
plot_utils.configure_fonts(fontsize=fontsize,
legend_fontsize=legend_fontsize)
if plot_yinvert:
ax.invert_yaxis()
spacing = category_padding * barwidth
num_categories = len(series[0])
category_width = spacing + barwidth
# position for each bar
index = numpy.array(range(num_categories)) * category_width + xpadding
# print index
y_offset = numpy.array([0.0] * num_categories)
for idx in range(num_series):
# x_offset = idx * (barwidth + series_padding)
x_offset = 0.
style = dict(
label=series_labels[idx],
color=colors.get_plot_color(series_colors[idx],
emphasis=series_color_emphasis[idx]),
ecolor=colors.get_plot_color(series_err_colors[idx]),
linewidth=0,
)
style.update(series_style[idx])
r = ax.bar(left=index + x_offset,
height=series[idx],
width=barwidth,
bottom=y_offset,
yerr=series_errs[idx],
**style)
y_offset = y_offset + series[idx]
# Label the plot
if plot_ylabel is not None:
ax.set_ylabel(plot_ylabel)
if plot_xlabel is not None:
ax.set_xlabel(plot_xlabel)
if plot_title is not None:
ax.set_title(plot_title)
# Add category tick marks
if series_use_labels:
indices = numpy.arange(num_series)
ticks = xpadding + indices * (barwidth +
series_padding) + 0.5 * barwidth
ax.set_xticks(ticks)
if series_labels is not None:
ax.set_xticklabels(series_labels)
elif category_ticks:
ticks = index + .5 * barwidth
ax.set_xticks(ticks)
if category_labels is not None:
ax.set_xticklabels(category_labels)
else:
ax.set_xticks([])
# Set the x-axis limits
lims = [
0, 2 * xpadding + num_categories *
(category_width + (num_series - 1) * series_padding) - spacing
]
ax.set_xlim(lims)
# Legend
if legend_location is not None:
if legend_pos_rel_plot == 'right':
# http://stackoverflow.com/questions/4700614/how-to-put-the-legend-out-of-the-plot
# Shrink current axis by 20%
box = ax.get_position()
ax.set_position([box.x0, box.y0, box.width * 0.8, box.height])
legend_location = 'center left'
legend_bbox_to_anchor = (1, 0.5)
# # Put a legend to the right of the current axis
# ax.legend(loc='center left', bbox_to_anchor=(1, 0.5))
# http://matplotlib.org/1.3.1/users/legend_guide.html
l = ax.legend(title=legend_title, labelspacing=legend_labelspacing,\
frameon=False, loc=legend_location, bbox_to_anchor=legend_bbox_to_anchor)
# http://matplotlib.org/1.3.1/users/legend_guide.html
if legend_reverse_label:
handles, labels = ax.get_legend_handles_labels()
l = ax.legend(handles[::-1], labels[::-1],\
title=legend_title, labelspacing=legend_labelspacing,\
frameon=False,loc=legend_location,\
bbox_to_anchor=legend_bbox_to_anchor)
if legend_title_pos is not None:
l.get_title().set_position(legend_title_pos)
# Make the axis pretty
plot_utils.simplify_axis(ax)
# Save the file
if savefile is not None:
plot_utils.output(fig,
savefile,
savefile_size,
fontsize=fontsize,
legend_fontsize=legend_fontsize)
# Show
if show_plot:
plt.show()
return fig, ax