django dynamic number of named groups in url - regex

I would like to develop a web service with dynamic urls.
A common url pattern with named groups would be something like this:
(r'^articles/(?P<year>\d{4})/(?P<month>\d{2})/$', 'news.views.month_archive')
For my purpose I need a little more dynamic, because I don't know the number of groups from the beginning.
E.g. possible urls could be:
articles/post/comments
articles/post/author
article/post/
...
I could use something like this:
(r'^(?P<cat1>)/(?P<cat2>)/$', 'module.views.function')
(r'^(?P<cat1>)/(?P<cat2>)/(?P<cat3>)/$', 'module.views.function')
(r'^(?P<cat1>)/(?P<cat2>)/(?P<cat3>)/(?P<cat4>)/$', 'module.views.function')
..
and so on. Is there a smarter way to do this? Thanks in advance!

Wait if you only have 1 article then you want it to be article/post/, otherwise you want it to be articles/post/?
Since URLs are just regular expressions you could do something like
(r'^articles?/post/$', 'some_view')
the ? tests either 0 or 1 of s.
But fundamentally your URLs don't seem to make sense, when you write articles/post/comments, it sounds like you want to post comments on all articles? Same with article/post/, what does that mean? Which article are you trying to post (you only have one and at most one)?
Typically URLs look like articles/5/post/comments, which would mean "I want to post comments on article with a pk of 5".
Sorry if my answer didn't help. I guess I didn't really understand your question.

Related

Hierarchical URLs in Django

Is there a way to implement hierarchical query pattern in Django? As far as I know, the framework only allows to route to views by parsing URLs of a specific format, like:
/customers/{order} -> customer.views.show_orders(order)
But what if I need something like this:
/book1/chapter1/section1/paragraph1/note5 -> notes.view.show(note_id)
where note_id is the id of the last part of the URL, but the URL could have different number of components:
/book1/chapter1
/book1/chapter1/section1
etc.
Each time, it would point to the relevant part of the book depth depending on the depth. Is this doable?
I know there is this: https://github.com/MrKesn/django-mptt-urls, but I am wondering if there is another solution. This isn't ideal for me.
Django URLs are just regular expressions, so the simplest way would be to just ignore everything prior to the "note" section of the URL. For example:
url(r'^.*/note(?P<note_id>[0-9]+)$', 'notes.view.show'),
However, this would ignore the book, chapter, paragraph components. Which would mean your notes would need unique ids across the system, not just within the book. If you needed to capture any number of the interim parts it would be more complicated.
I can't confirm this will work right now, but using non-capture groups in regular expressions, you should be able to capture an optional book and chapter like so:
url(r'^(?:book(?P<book_id>[0-9]+)/)?(?:chapter(?P<chapter_id>[0-9]+)/)?note(?P<note_id>[0-9]+)$', 'notes.view.show'),
Use named groups to accomplish this: https://docs.djangoproject.com/en/dev/topics/http/urls/#named-groups
url(r'^book(?P<book_id>\d+)/chapter(?P<chapter_id>\d+)/section(?P<section_id>\d+)/paragraph(?P<paragraph_id>\d+)/note(?P<note_id>\d+)$', notes.view.show(book_id, chapter_id, section_id, paragraph_id, note_id)
For those who really need a variable-depth URL structure and need the URL to consist strictly of slugs, not IDs, knowing all the components of the URL is critical to retrieve the correct record from the database. Then, the only solution I can think of is using:
url(r'^.*/$', notes.views.show, name='show')
and then parsing the content of the URL to get the individual components after retrieving the URL in the view using the request.path call. This doesn't sound ideal, but it is a way to accomplish it.

Django documentation, part3 understanding problems

I read the documentation of Django but now I am at a point where I need some explanation. It is on this site and I understand the views but I really don't get how the urls work. It looks pretty cryptic and confusing to me. Can anybody explain to me how the urls work and what their purpose is?
Your urls.py file is virtual. They do it this way so you don't need to worry about a static url to http://yoursite.com/polls/34. By using this number as a regular expression /(d+) you can keep it dynamic so one url with this regular expression can be millions of different polls.
when the url is requested that regular expression number (whether it's 1 or 13352) is sent to the view which then says, I need to query the database for a Poll that has a PrimaryKey (PK) of whatever this number is. If it's found the Poll object is sent to the template by the view. The template then displays all the data in the poll object.
The bottom line is using something like this you can have one line for a url which is essentially millions of different urls. I use this same format for a movies website I'm creating www.noobmovies.com. I follow the same structure for Stars, Movies and blogs. Essentially three lines of code has created urls for 10,000 pages or so.
There is a dedicated Django documentation page for that: https://docs.djangoproject.com/en/1.6/topics/http/urls/
Maybe it will help you?

How can I write a regex to match everything but a char?

I've switched in my wordpress blog from urls like this:
/blog/2012/01/01/how-to-build-a-website
To shorter urls like this
/blog/?p=123
Wordpress has a search engine who works like this
/blog/search/?s=how to build a website
And search for the s params.
I'm trying to use .htaccess Redirectmatch to redirect all the old urls to the search url with the title of the post as the s params.
So if the user serf to
/blog/2012/01/01/how-to-build-a-website
should be redirect to
/blog/search/?s=how to build a website
I've coded this
Redirectmatch blog/\d+/\d+/\d+/(.+) http://www.mysite.com/blog/?s=$1
But this regex grap the whole string after the last / within the - symbol inside it.
In this way if a user serf to
/blog/2012/01/01/how-to-build-a-website
Will be redirected to
/blog/search/?s=how-to-build-a-website
while I want the user redireced to
/blog/search/?s=how to build a website
How can I write the regex to do this?
EDIT:
Yes guys, I know that this kind of urls are ugly :) But I just would know how to do it, because behind this there are some technical issues I'm trying to solve..
Please don't do this. I know it can seem tempting to go for short URLs; after all, you get things like TinyURL and such. Isn't it better to have /blog/?p=123 than /blog/2012/01/01/how-to-build-a-website?
No. It's not.
The reason is because when someone posts a link to your blog article, the longer URL means something. It tells the person how old the article is. It gives the title. It helps people find your article; after all, the URL is given a lot of weight by Google when indexing your site.
URLs used to be built for computers. Something like /blog/?p=123 is perfect for computers; it's easy to parse, it doesn't require any extra database lookups. You can write two articles named "How to Build a Website" and the blog engine doesn't have to make sure it adds a -2 on the second one. It maps easily to the actual structure of files on the server, without making up structure in the URL.
But we've realized since that URLs can be built for humans, too. The URL /blog/2012/01/01/how-to-build-a-website has a form that can be easily understood by humans. Sure, it's a bit longer to type, but all the bits you're typing are easier, and most URLs are copy'n'pasted anyway or just clicked on. It's more work for the computer, sure, but it's worth it. It makes the Internet friendlier.
So I'm sorry, but I won't help you. :)

Django URL Aliases

I'm new to Django and I have a BIG problem. I don't like the "url pattern" philosophy of Django.
I don't want my pages to look like
http://domain.com/object/title-of-object
I want
http://domain.com/title-of-object
and of course I will have more than one type of object.
Is there an elegant way to achieve this with Django (not using hard-coded urls)?
Thanks!
Ever wondered that, if what you want to do seems so hard to acheive, you're doing it wrong? What is so wrong with /foo/name-of-foo/ ?
I'm trying to imagine your use-case and wondering if you need 'human' URLs for only a handful of pages. If so, it would work to go with the /foo/slug-for-foo/ approach but then use the django.contrib.redirects app to support hand-written URLs that redirect to the saner, more RESTful ones?
It is possible. You'll have to create one catch-all URL pattern, for which you'll create a view that will search all possible object types, find the matching one, and process and return that. Usually, this is a bad idea.

django internalization in urls? how to make urls like this: "en/articles" and "pt/artigos"...?

Hy!
I would need to make urls based on language.
Example:
If I had the language english and subcategory named "articles" then the url might be like this:
/en/articles/...
If the language were portuguese and subcategory already translated is "artigos" then the url will be:
/pt/artigos/...
How can I do it?
I must use the urls.py or some monkeypatches?
thanks
This features is already existing in the yet-to-be released version 1.4. You can read about it in the release note.
If your really need this feature, and no existing app feets your needs, you can still try to apply the corresponding patch yourself.
Django LocaleURL is a piece of middleware that does exactly this. The documentation can be found in the source, or online.
Edit: I read over the fact that you want to translate the url itself... I'm not aware of any piece of code that provides this. Perhaps you could extend the LocalURL middleware to take care of the translations for you. Say you have a regex match like (?P<articles>\w+), you could in the middleware determine which view you want to use. Something like the following mapping perhaps?
if article_slug in ['articles', 'artigos', 'article']:
article(request) # Call the article view
I've been using transurlvania with great success, it does exactly what you need and more, however i see that in the next Django release django-i18nurls will be included in django core so perhaps it would be better to learn that