Divide by zero in distance.yule

[peterbell10]

Yule dissimilarity of two identical and constant vectors results in a divide by zero.

Reproducing code example:

from scipy.spatial.distance import yule
>>> from scipy.spatial.distance import yule
>>> yule([1, 1, 0], [1, 1, 0])
0.0

>>> yule([1, 1, 1], [1, 1, 1])
/home/peter/git/scipy/build/testenv/lib/python3.7/site-packages/scipy/spatial/distance.py:1299: RuntimeWarning: invalid value encountered in double_scalars
  return float(2.0 * ntf * nft / np.array(ntt * nff + ntf * nft))
nan

Wolfram alpha confirms, the result should be zero:
https://www.wolframalpha.com/input/?i=YuleDissimilarity%5B%7B1%2C+1%2C+1%7D%2C+%7B1%2C+1%2C+1%7D%5D

Scipy/Numpy/Python version information:

1.7.0.dev0+6673cb8 1.20.1 sys.version_info(major=3, minor=7, micro=8, releaselevel='final', serial=0)