Python URL manipulation revisited

My last blog post about publishing standalone files to Nexus repositories prompted me to revisit URL manipulation in Python. When I did this the last time, I used Python stand library’s urlparse and it did the job. This time around, I needed to do a different kind of manipulation. Given a URL, I had to set credentials on it.

I started at urlparse and soon realized that Python3 moved this module to urllib.parse. That is not too bad, I thought. After playing around with it, it became clear that urllib.parse can’t manipulate credentials in a URL.

In [1]: from urllib import parse

In [2]: url = parse.urlparse('http://example.com')

In [3]: print(url)
ParseResult(scheme='http', netloc='example.com', path='', params='', query='', fragment='')

In [4]: url.username = "username"
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-4-60e52fe02603> in <module>()
----> 1 url.username = "username"

AttributeError: can't set attribute

In [5]: url.password = "password"
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-5-4f64c7192b99> in <module>()
----> 1 url.password = "password"

AttributeError: can't set attribute

In [6]: url.username == None
Out[6]: True

In [7]: url.password == None
Out[7]: True

After spending time with various URL manipulation libraries in Python, furl was the only library that I found to be capable of this seemingly simple and common enough operation.

In [1]: from furl import furl

In [2]: url = furl('http://example.com')

In [3]: url
Out[3]: furl('http://example.com')

In [4]: url.password = "password"

In [5]: url.username = "user"

In [6]: url.tostr()
Out[6]: 'http://user:[email protected]'