With all the helpful libraries out there, you'd think Python would have an easier way of doing this!
Using urlparse() to get the ParseResult object only allows read-only access to the attributes, making it useless if you want to set the query field.
This solution comes from Ned Batchelder from StackOverflow. I found this to be much more useful than the stuff in urlparse and urllib libraries.
Create a helper class:
| import cgi, urllib, urlparse | |
| class Url(object): | |
| """ | |
| Helper class to help add/remove GET args from an existing URL. | |
| Usage: | |
| url = Url('http://www.koonkii.com/apps/?type=Android&page=5#top') | |
| del url.args['type'] | |
| print url.url | |
| >>> 'http://www.koonkii.com/apps/?page=5#top' | |
| @see http://stackoverflow.com/questions/2873438/is-there-a-better-way-to-write-this-url-manipulation-in-python | |
| """ | |
| def __init__(self, url): | |
| """ | |
| Construct from a string, Url object or Django's HttpRequest. | |
| """ | |
| if isinstance(url, Url): | |
| url = url.url | |
| # Django HTTP request | |
| try: | |
| from django.http import HttpRequest | |
| if isinstance(url, HttpRequest): | |
| url = url.get_full_path() | |
| except ImportError: | |
| pass | |
| # http://, www.koonkii.com, /apps/, ?, {'type':'Android','page':'5'}, top | |
| self.scheme, self.domain, self.path, self.params, self.query, self.fragment = urlparse.urlparse(url) | |
| self.args = dict(cgi.parse_qsl(self.query)) | |
| @property | |
| def url(self): | |
| return str(self) | |
| @property | |
| def query_string(self): | |
| """ | |
| Returns args as a query string | |
| """ | |
| args = dict(self.args) | |
| # Must be encoded in UTF-8 | |
| for k, v in args.items(): | |
| if isinstance(v, unicode): | |
| v = v.encode('utf8') | |
| elif isinstance(v, str): | |
| v.decode('utf8') | |
| args[k] = v | |
| # In case we've modified args | |
| self.query = urllib.urlencode(args) | |
| return '?%s' % self.query if len(self.args) else '' | |
| def __str__(self): | |
| """ | |
| Turn back into a URL and special-case the %UID% | |
| URLs are not UTF-8 compatible and non-ascii characters should be url-encoded. | |
| """ | |
| self.query = urllib.urlencode(self.args) | |
| return "%s" % urlparse.urlunparse((self.scheme, self.domain, self.path, self.params, self.query, self.fragment)) |
To get an idea of what's what:
01.txt = "https://gist.github.com/twig/d42c4ec9ecc5eccf3614#file-python-pil-exif-reader-py-L12"02.url = Url(txt)03.print "original", txt04.print "scheme", url.scheme05.print "domain", url.domain06.print "path", url.path07.print "args", url.args08.print "params", url.params09.print "query", url.query10.print "query_string", url.query_string11.print "fragment", url.fragment12.print "URL", url.url13.print "str", str(url)Prints out:
original: https://gist.github.com/twig/d42c4ec9ecc5eccf3614?q=url&l=python#file-python-pil-exif-reader-py-L12
scheme: https
domain: gist.github.com
path: /twig/d42c4ec9ecc5eccf3614
params:
args: {'q': 'url', 'l': 'python'}
query: q=url&l=python
query_string: ?q=url&l=python
fragment: file-python-pil-exif-reader-py-L12
URL: https://gist.github.com/twig/d42c4ec9ecc5eccf3614?q=url&l=python#file-python-pil-exif-reader-py-L12
str: https://gist.github.com/twig/d42c4ec9ecc5eccf3614?q=url&l=python#file-python-pil-exif-reader-py-L12
To make changes, just use like this:
1.u = Url(url)2.del u.query['page']3.print u.url*update 23/08/2011*
- Removed the use of urllib.urlencode() because it's deprecated. Now uses urllib.
- Added __str__() to accompany __unicode__(). Added build() method.
- Renamed args to be query.
*update 11/11/2014*
- Cleaned up the code a bit
- Added Url.url and Url.query_string attributes
- Updated comments
- Renamed Url.netloc to Url.domain
- Now supports Url and Django's HttpRequest objects in __init__()
