def sbar(data, title, xlabel, ylabel, xticks, group_names, legend=True, fontsize=16, ylim=()):
"""
plot a stacked bar graph (N bars stacked on top of eachother)
data an Nx or Nx * Nn array or ndarray
xticks labels for each x value
group_names labels for each n value
"""
legend = bool(int(legend))
data = np.array(data)
group_names = list(group_names)
n_labels = group_names
percent = 0
# mpl.rcParams['axes.linewidth'] = 1
# rcParams['axes.linewidth'] = 7 # set the value globally
# if data is 1d, make it a 2d array
if len(data.shape) == 1:
# print "1D array: Reshaping"
data = data.reshape(len(data), -1)
else:
assert data.shape[1] == len(group_names), "data.shape[1] = %d, len(group_names) = %d" % (
data.shape[1],
len(group_names),
)
# print data.shape
assert len(data.shape) == 2
assert data.shape[0] == len(xticks), "data.shape[0] = %d, len(xticks) = %d" % (data.shape[0], len(xticks))
rot = 0
max_x_label = max([len(str(s)) for s in xticks])
max_text_width = 100 / len(xticks)
print "label:", max_x_label, "max:", max_text_width
if max_x_label > max_text_width:
rot = 30
# mpl.rc('font',**{'family':'sans-serif', 'sans-serif':['Helvetica'], 'size':fontsize})
mpl.rc("font", **{"size": fontsize})
legend_outside = 0
n = data.shape[1]
show_legend = legend and (n > 1)
print "Show Legend =", show_legend
# left bottom width height
box = [0.1, 0.08, 0.85, 0.82]
figsize = [8, 6] # WxH
bbanchor = [0.5, 1] # x,y
legend_loc = "lower center"
if legend_outside and show_legend:
print "Putting legend outside plot"
box[2] /= 1.25
bbanchor = (1, 1)
legend_loc = "upper left"
figsize[0] *= 1.25
if rot:
box[1] += 0.08
box[3] -= 0.08
if ylabel == "":
box[0] -= 0.05
if title != "":
box[3] -= 0.05
# Create Figure
fig = plt.figure(figsize=figsize)
# fig, ax = plt.subplots()
# left bottom width height
ax = fig.add_axes(box)
plt.title(title)
plt.xlabel(xlabel)
plt.ylabel(ylabel)
plt.autoscale(enable=True, axis=u"both", tight=False)
# gridelines
ax.yaxis.grid(True, zorder=0, linestyle="solid", color="0.9", linewidth=1)
# dont use offset at the top of the figure
y_formatter = mpl.ticker.ScalarFormatter(useOffset=False)
ax.yaxis.set_major_formatter(y_formatter)
if len(group_names) == 1:
width = 0.4
else:
width = 0.8 / len(group_names)
ind = np.arange(data.shape[0])
n = data.shape[1]
if len(xticks) > 20:
xticks = [""]
ax.set_xticklabels(xticks, rotation=rot)
xtp = ind # +width*n/2
ax.set_xticks(xtp)
# ax.spines['left'].set_visible(False)
# ax.spines['bottom'].set_visible(False)
# convert to percent
if percent:
data *= 100
fmt = "%.0f%%" # Format you want the ticks, e.g. '40%'
yticks = mtick.FormatStrFormatter(fmt)
ax.yaxis.set_major_formatter(yticks)
dmin = np.min(data.sum(1))
dmax = np.max(data.sum(1))
diff = dmax - dmin
if len(ylim) == 0:
ymin = dmin - 0.1 * diff
ymax = dmax + 0.1 * diff
elif len(ylim) == 1:
ymin = 0
ymax = ylim[0]
elif ylim[0] == "":
ymin = dmin - 0.1 * diff
ymax = ylim[1]
elif ylim[1] == "":
ymin = ylim[0]
ymax = dmax + 0.1 * diff
else:
ymin = ylim[0]
ymax = ylim[1]
if type(xticks[0]) is np.string_ or type(xticks[0]) is str:
xmin = 0
xmax = len(xticks)
else:
xmin = int(xticks[0])
xmax = int(xticks[-1])
xmin -= 0.5
xmax += -0.5
# plt.style.use('ggplot')
# plt.xticks(np.arange(0, len(xticks), 1))
# plt.autoscale(enable=True, axis=u'both', tight=True)
axes = plt.gca()
axes.set_xlim([xmin, xmax])
axes.set_ylim([ymin, ymax])
# axes.set_yticks(np.arange(ymin,ymax,0.1))
# axes.spines['left'].linewidth = 2
# axes.spines['bottom'].linewidth = 5
# Colors / Patterns
color = myplot.cmap(n)
# color=itertools.cycle(plt.cm.Paired(np.linspace(0,1,6)))
fill = iter(["", "/"])
p = [] # list of bars
b = [0] * len(xticks) # bottom position
# draw a veritcal line at 1
plt.axhline(y=1.0, zorder=3, linewidth=2, color="black")
for i, n in enumerate(n_labels):
d = list(data[:, i])
print n
print d
inds = ind - (len(n_labels) - 1) * width / 2 + width * i
print "inds:", inds
# annotate values to bars that go outside the plot
if d[0] > ymax:
fs = int(width * 30)
ax.text(inds[0], ymax, "%.1f" % d[0], ha="center", va="bottom", fontsize=fs, zorder=5, rotation=45)
# rect.get_x() + rect.get_width()/2., 1.05*height,
# '%d' % int(height),
# ha='center', va='bottom')
p.append(
# ppl.bar(
plt.bar(
# ind + width*i - width/2,
inds,
d,
width,
zorder=4,
color=next(color),
# color=plt.rcParams['axes.color_cycle'][i],
alpha=1,
# bottom=b,
align="center",
# hatch=next(fill)
hatch="",
)
)
b = np.add(b, d)
if show_legend:
# p = p[::-1]
plt.legend(
p,
n_labels,
ncol=n_labels,
fontsize=fontsize,
loc=legend_loc,
bbox_to_anchor=(bbanchor),
borderaxespad=1,
frameon=False,
)
def bar(data, title, xlabel, ylabel, xticks, group_names, legend=True, fontsize=16):
"""
data an Nx or Nx * Nn array or ndarray
xticks labels for each x value
group_names labels for each n value
"""
legend=bool(int(legend))
data = np.array(data)
group_names = list(group_names)
n_labels = group_names
# if data is 1d, make it a 2d array
if ( len(data.shape) == 1 ):
# print "1D array: Reshaping"
data = data.reshape(len(data),-1)
else:
assert data.shape[1] == len(group_names), "data.shape[1] = %d, len(group_names) = %d" % (data.shape[1],len(group_names))
# print data.shape
assert(len(data.shape) == 2)
assert(data.shape[0] == len(xticks)), "data.shape[0] = %d, len(xticks) = %d" % (data.shape[0],len(xticks))
rot = 0
max_x_label = max([len(str(s)) for s in xticks])
max_text_width = 60/len(xticks)
print "label:",max_x_label,"max:",max_text_width
if (max_x_label > max_text_width):
rot = 30
if (max_x_label > 2*max_text_width):
rot = 60
mpl.rc('font', family='Helvetica', size=fontsize)
legend_outside=1
n = data.shape[1]
show_legend = legend and (n > 1);
print "Show Legend =", show_legend
# left bottom width height
box = [0.10,0.12,0.82,0.85]
figsize=[ 8, 6 ]
bbanchor=[ 1, 1 ]
legend_loc='upper right'
if legend_outside and show_legend:
print "Putting legend outside plot"
box[2] /= 1.25
bbanchor=( 1, 1 )
legend_loc='upper left'
figsize[0] *= 1.25
box[0] -= 0.05
if rot:
box[0] += 0.02
box[1] += 0.08
box[3] -= 0.08
if ylabel == '':
box[0] -= 0.05
if title != '':
box[3] -= 0.05
fig = plt.figure()
#fig, ax = plt.subplots()
# left bottom width height
ax = fig.add_axes(box)
plt.title(title)
plt.xlabel(xlabel)
plt.ylabel(ylabel)
plt.autoscale(enable=True, axis=u'both', tight=False)
ax.yaxis.grid(True)
if (len(group_names) == 1):
width=0.5
else:
width=0.9/len(group_names)
ind = np.arange(data.shape[0])
n = data.shape[1]
if (len(xticks) > 20):
xticks = ['']
ax.set_xticklabels(xticks, rotation=rot, ha='right')
xtp = ind+width*n/2
ax.set_xticks(xtp)
ymin = 1.0
ymax = 1.0
ymax = max(data)*1.1
axes = plt.gca()
# axes.set_xlim([xmin,xmax])
axes.set_ylim([ymin,ymax])
#axes.set_yticks(np.arange(ymin,ymax,0.1))
# add bars
color=myplot.cmap(n)
groups = list()
for d in range(0,n):
c=next(color)
grp = ax.bar(ind + width*d, data[:,d], width, color=c)
groups.append(grp)
if (show_legend):
plt.legend(
groups,
n_labels,
fontsize=fontsize,
loc=legend_loc,
bbox_to_anchor=(bbanchor),
borderaxespad=0.5
)