def _SP(xdata, mx, ydata, my):
"""SP = sum of product of deviations.
Helper function for calculating covariance directly.
"""
if mx is None:
# Two pass algorithm.
xdata = as_sequence(xdata)
mx = stats.mean(xdata)
if my is None:
# Two pass algorithm.
ydata = as_sequence(ydata)
my = stats.mean(ydata)
return _generalised_sum(zip(xdata, ydata), lambda t: (t[0]-mx)*(t[1]-my))
def split(xdata, ydata=None):
"""Helper function which splits xydata into (xdata, ydata)."""
# The two-argument case is easy -- just pass them unchanged.
if ydata is not None:
xdata = as_sequence(xdata)
ydata = as_sequence(ydata)
if len(xdata) < len(ydata):
ydata = ydata[:len(xdata)]
elif len(xdata) > len(ydata):
xdata = xdata[:len(ydata)]
assert len(xdata) == len(ydata)
return (xdata, ydata)
# The single argument case could be either [x0, x1, x2, ...] or
# [(x0, y0), (x1, y1), (x2, y2), ...]. We decide which it is by
# looking at the first item, and treating it as canonical.
it = iter(xdata)
try:
first = next(it)
except StopIteration:
# If the iterable is empty, return two empty lists.
return ([], [])
# If we get here, we know we have a single iterable argument with at
# least one item. Does it look like a sequence of (x,y) values, or
# like a sequence of x values?
try:
n = len(first)
except TypeError:
# Looks like we're dealing with the case [x0, x1, x2, ...]
# This isn't exactly *multivariate*, but we support it anyway.
# We leave it up to the caller to decide what to do with the
# fake y values.
xdata = [first]
xdata.extend(it)
return (xdata, [None]*len(xdata))
# Looks like [(x0, y0), (x1, y1), (x2, y2), ...]
# Here we expect that each point has two items, and fail if not.
if n != 2:
raise TypeError('expecting 2-tuple (x, y) but got %d-tuple' % n)
xlist = [first[0]]
ylist = [first[1]]
for x,y in it:
xlist.append(x)
ylist.append(y)
assert len(xlist) == len(ylist)
return (xlist, ylist)
