Author rhettinger
Recipients mark.dickinson, rhettinger, skrah, steven.daprano, tim.peters
Date 2018-03-16.18:50:58
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1521226258.78.0.467229070634.issue33089@psf.upfronthosting.co.za>
In-reply-to
Content
A need for a distance-between-two-points function arises frequently enough to warrant consideration for inclusion in the math module.   It shows-up throughout mathematics -- everywhere from simple homework problems for kids to machine learning and computer vision.

In the latter cases, the function is called frequently and would benefit from a fast C implementation that includes good error checking and is algorithmically smart about numerical issues such as overflow and loss-of-precision.

A simple implementation would be something like this:

    def dist(p, q):
        'Multi-dimensional Euclidean distance'
        # XXX needs error checking:  len(p) == len(q)
        return sqrt(sum((x0 - x1) ** 2 for x0, x1 in zip(p, q)))

The implementation could also include value added features such as hypot() style scaling to mitigate overflow during the squaring step:

    def dist2(p, q):
        # https://en.wikipedia.org/wiki/Hypot#Implementation
        diffs = [x0 - x1 for x0, x1 in zip(p, q)]
        scale = max(diffs, key=abs)
        return abs(scale) * sqrt(fsum((d/scale) ** 2 for d in diffs))
History
Date User Action Args
2018-03-16 18:50:58rhettingersetrecipients: + rhettinger, tim.peters, mark.dickinson, steven.daprano, skrah
2018-03-16 18:50:58rhettingersetmessageid: <1521226258.78.0.467229070634.issue33089@psf.upfronthosting.co.za>
2018-03-16 18:50:58rhettingerlinkissue33089 messages
2018-03-16 18:50:58rhettingercreate