def _plot(cls,
ax,
y,
style=None,
bw_method=None,
ind=None,
column_num=None,
stacking_id=None,
**kwds):
# 'y' is a Spark DataFrame that selects one column.
# Using RDD is slow so we might have to change it to Dataset based implementation
# once Spark has that implementation.
sample = y.rdd.map(lambda x: float(x[0]))
kd = KernelDensity()
kd.setSample(sample)
assert isinstance(
bw_method,
(int, float)), "'bw_method' must be set as a scalar number."
if bw_method is not None:
# Match the bandwidth with Spark.
kd.setBandwidth(float(bw_method))
y = kd.estimate(list(map(float, ind)))
lines = PandasMPLPlot._plot(ax, ind, y, style=style, **kwds)
return lines
def _plot(cls,
ax,
y,
style=None,
bw_method=None,
ind=None,
column_num=None,
stacking_id=None,
**kwds):
y = KdePlotBase.compute_kde(y, bw_method=bw_method, ind=ind)
lines = PandasMPLPlot._plot(ax, ind, y, style=style, **kwds)
return lines