def binary_grouped_scatter(M,groups, othergroups, axis=1,components=(0, 1), group_labels=None, labels=("", ""), title="", interactive=True, show=True):
"""
Return a figure of a grouped scatterplot of the matrix M.
M: Matrix to visualise.
groups: List of groups for the points in M.
othergroups: Dictionary of groups to index that can be highlighted.
axis: Axis of data values to plot. Needs to be either 1 or 0. Defaults to 1.
components: Components to make a scatter plot of. Defaults to (0, 1).
title: Title of the Plot to make. Defaults to "".
labels: Labels of the axes to plot.
group_labels: Labels for the groups. If given, need to be sorted by group number.
show: Should the figure be shown immediatly. Defaults to True.
"""
# ensure axis is of the right type.
if axis != 1 and axis != 0:
raise ValueError("Axis must be 0 or 1")
if axis == 0:
return binary_grouped_scatter(M.T,groups,othergroups, axis=1,components=components,group_labels=group_labels, labels=labels,title=title,show=show)
(pgroups, plabels) = make_labels(groups, group_labels)
# make a figure with a title.
fig = pyplot.figure()
fig.suptitle(title)
# extarct Xs and Ys we want to plot
Xs = M[components[0],:]
Ys = M[components[1],:]
color=iter(cm.rainbow(np.linspace(0,1,len(plabels))))
colors = [next(color) for c in plabels]
# texts to show
keys = list(othergroups.keys())
texts = [None for k in keys]
x_lim = None
y_lim = None
def plot_points(ids, x_lim = None, y_lim = None):
# limit to given ids
X2s = Xs[ids]
Y2s = Ys[ids]
p2groups = pgroups[ids]
# clear the figure
pyplot.clf()
pyplot.cla()
# remove the current axis
try:
pyplot.gca().legend_.remove()
except:
pass
for (i, n) in enumerate(plabels):
gids = np.where(p2groups == i)
if len(gids) == 0:
continue
# make a plot
pyplot.scatter(X2s[gids], Y2s[gids], c=colors[i], label=str(n), picker=False)
if x_lim != None:
pyplot.gca().set_xlim(x_lim)
if y_lim != None:
pyplot.gca().set_ylim(y_lim)
# get some ranges
yrange = pyplot.gca().get_ylim()
xmin = pyplot.gca().get_xlim()[0]
# and some keys and their positions
ytexts = np.linspace(yrange[0], yrange[1], len(keys) + 2)[1:-1]
for i, (k, y) in enumerate(zip(keys, ytexts)):
texts[i] = pyplot.text(xmin, y, k, picker=True)
pyplot.legend()
return (pyplot.gca().get_xlim(), pyplot.gca().get_ylim())
selected_key = [None]
def onpick(event):
try:
key = keys[texts.index(event.artist)]
except:
return
if selected_key[0] == key:
plot_points(list(range(len(pgroups))), x_lim = x_lim, y_lim = y_lim)
selected_key[0] = None
else:
plot_points(othergroups[key], x_lim = x_lim, y_lim = y_lim)
selected_key[0] = key
# re-draw
pyplot.draw()
cid = fig.canvas.mpl_connect('pick_event', onpick)
# set a few labels
pyplot.xlabel(labels[0])
pyplot.ylabel(labels[1])
# plot all the points
(x_lim, y_lim) = plot_points(list(range(len(pgroups))))
# show it if we want to.
if show:
pyplot.show()
# return the figure.
return fig
def grouped_scatter(M,groups, axis=1,components=(0, 1), group_labels=None, labels=("", ""), title="", interactive=True, show=True):
"""
Return a figure of a grouped scatterplot of the matrix M.
M: Matrix to visualise.
groups: List of groups for the points in M.
axis: Axis of data values to plot. Needs to be either 1 or 0. Defaults to 1.
components: Components to make a scatter plot of. Defaults to (0, 1).
title: Title of the Plot to make. Defaults to "".
labels: Labels of the axes to plot.
group_labels: Labels for the groups. If given, need to be sorted by group number.
interactive: Should the plot be made interactive.
show: Should the figure be shown immediatly. Defaults to True.
"""
# ensure axis is of the right type.
if axis != 1 and axis != 0:
raise ValueError("Axis must be 0 or 1")
if axis == 0:
return grouped_scatter(M.T,groups,axis=1,components=components,group_labels=group_labels, labels=labels,title=title,show=show)
(pgroups, plabels) = make_labels(groups, group_labels)
# make a figure with a title.
fig = pyplot.figure()
fig.suptitle(title)
# extarct Xs and Ys we want to plot
Xs = M[components[0],:]
Ys = M[components[1],:]
color=iter(cm.rainbow(np.linspace(0,1,len(plabels))))
plots = []
colors = []
for (i, n) in enumerate(plabels):
gids = np.where(pgroups == i)
# generate a new color
colors.append(next(color))
# make a plot
pplot = pyplot.scatter(Xs[gids], Ys[gids], c=colors[-1], label=str(n), picker=5 if interactive else False)
# and append the current color
plots.append(pplot)
# set a few labels
pyplot.xlabel(labels[0])
pyplot.ylabel(labels[1])
pyplot.legend()
selected_index = [-1]
def onpick(event):
# get the current index
index = plots.index(event.artist)
if selected_index[0] == index:
for (i, p) in enumerate(plots):
p.set_facecolors(colors[i])
# reset the selected index
selected_index[0] = -1
else:
# make all the plots black
for p in plots:
p.set_facecolors("k")
# make this plot normal color and mark it as selected
event.artist.set_facecolors(colors[index])
selected_index[0] = index
# re-draw
pyplot.draw()
if interactive:
cid = fig.canvas.mpl_connect('pick_event', onpick)
# show it if we want to.
if show:
pyplot.show()
# return the figure.
return fig
