%xx-escapes

Have you ever come across those strange looking %xx-escapes, that are used in URLs. Yes, I don't like them either ...

I can imagine a class of use cases where such kind of escape characters may need to be handled: say, you'd like to annotate - in a local or remote document - a number of URLs, with a given substring (of the URL itself).

Let's look at the case where the cardinality of our URLs set is unity:

# Python 2.7.12
In [1]: url
Out[1]: 'https://www.a-random-ml-website.com/na%C3%AFve-bayes'

Since I don't suppose any of you would be particularly keen on annotating this URL with na\xc3\xafve-bayes, we'll have to resort to utility functions provided by urllib2¹:

In [2]: import urllib2

In [3]: urllib2.unquote(url)
Out[3]: 'https://www.a-random-ml-website.com/na\xc3\xafve-bayes'

Ugh. We've suddenly time-travelled back to a few days ago! Luckily enough, our memories of the previous timeline branch have been preserved intact - we are empowered with the knowledge of the unicodedata module:

In [4]: url = urllib2.unquote(url).decode('utf8')

In [5]: url
Out[5]: u'https://www.a-random-ml-website.com/na\xefve-bayes'

In [6]: import unicodedata

In [7]: better_looking_url = unicodedata.normalize('NFKD', url).encode('ascii', 'ignore')

In [8]: better_looking_url
Out[8]: 'https://www.a-random-ml-website.com/naive-bayes'

Now, you can slice and dice better_looking_url as you please, for your annotation purposes.