post_delete.connect(MediaAttachment.objects.delete_attachment)

	def delete_attachment(self, instance, *args, **kwargs):
		ctype = ContentType.objects.get_for_model(instance)
		return self.filter(content_type=ctype, object_id=instance.pk).delete()

Hanpan fucked around with this message at 17:29 on Feb 3, 2011

raymond fucked around with this message at 00:52 on Feb 8, 2011

duck monster fucked around with this message at 03:12 on Feb 15, 2011

epswing fucked around with this message at 06:04 on Feb 15, 2011

MEDIA_ALLOWED_TYPES = ('jpg', 'jpeg', 'gif')

	@property
	def thumbnail(self):
		if hasattr(self, '_thumb'):
			return self._thumb
		else:
			setattr(self, '_thumb', get_thumbnail(self.file))
			return self._thumb

Hanpan fucked around with this message at 11:09 on Feb 17, 2011

• Let's say I were making a blog application, and there were a bunch of ways I wanted to be able to display the data (by date, by author, by tag, etc.). One way I could imagine doing this is with view function called post that takes a bunch of arguments that lead to a specific QuerySet and then pass this all to one template file (post.html), so there's consistency in the way blog posts are displayed. The other way I could think of doing it is to have separate view fucntions (by_date, by_author, by_tag, etc.) that all do their own QuerySet and maybe have their own template, but for the display of the actual post itself, I just include single-post.html or something like that, so again the posts are displayed the same way.
  
  The former seems slightly more DRY to me, but at the expense of having one relatively complicated view function as opposed to bunch of little ones, so I think in that sense the latter approach is superior.
  
  
• I'd really like to have pretty URLs. For example, for an individual blog post, say I had an BlogPost object whose title is "«Je parle le français (un peu)»" written on May 29th, 2011. I'd like the url to be http://www.mysite.com/posts/2011/05/29/je-parle-le-francais-un-peu/.
  
  My solution at the moment feels a little fragile. Let's first assume that I have a function urlify that will transform the string "«Je parle le français (un peu)»" into "je-parle-le-francais-un-peu" (i.e. everything goes to lowercase, spaces are replaced with dashes, punctuation is dropped, accented characters go to ASCII-looking equivalents if available, otherwise are dropped)--basically, it makes strings URL-friendly. Here's a simplified version of how I have things set up (ignore any typos, the real code is actually working right now):
  

  code:

  ### models.py ###
  (usual imports)
  from mysite.scripts.urlify import urlify
  
  class BlogPost(models.Model):
      title = models.CharField(max_length=50)
      timestamp = models.DateTimeField()
      body = models.TextField()
      url = models.CharField(max_length=100, editable=False)
      
      def get_absolute_url(self):
          return r'/posts/%s/%s' % (self.timestamp.strftime(r'%Y/%m/%d'), urlify(self.title))
      
      def save(self, *args, **kwargs):
          # need to leave off front slash due to incompatibilities between urls.py and get_absolute_url() method
          self.url = self.get_absolute_url()[1:]
          is_by_staff = self.author.is_staff
          super(BlogPost, self).save(*args, **kwargs)
  
  ### urls.py ###
  urlpatterns += (r'^(?P<abs_url>posts/\d{4}/\d{2}/\d{2}/[a-z0-9].*)/$', post)
  
  ### views.py ###
  def post(request, abs_url):
      entries = BlogPosts.objects.filter(url = abs_url)
      t = loader.get_template('post.html')
      c = Context({ 'entries': entries })
      return HttpResponse(t.render(c))
  

  So this solution has a couple of things going for it: 1) it's working right now, 2) the URLs are pretty, and 3) the URL field is automatically generated on save, so if I have non-programmy friends/staff members who want to make posts, they don't have to worry about making a nice url.
  
  It kind of sucks, though, because I may want to change get_absolute_url at some point, and that would mean all the BlogPost objects need to be re-saved, so that sucks. (Is there a way I can do a mass re-save?) It's probably bad in some other ways, too.
  
  I'm pretty sure I can't generate a QuerySet from the urlify'd URL that would always find the BlogPost I'm after due to the information loss from the urlify function.
  
  Looking around the web, I see that The Onion's website does something that in Django would look like:
  

  code:

  urlpatterns += (r'^articles/[a-z].*,(?P<article_id>\d+)/$', article_display)
  

  where the [a-z].* is the article title in a format similar to how I want to urlify my BlogPost.titles, but I feel like just displaying BlogPost.id is a little inelegant, and it seems like lots of places don't do that. Do they just have a manually-created BlogPost.url field so then you'd have
  

  code:

  class BlogPost(models.Model):
      ...
      def get_absolute_url(self):
          return 'posts/%s/%s' % (self.timestamp.strftime('%Y/%m/%d'), self.url)
  

  or something like that instead of my version?
  
  
• Here's one that's probably a bit easier: what's the best way to get the title tag text to be all uniform and nice i.e. if I want to quickly change between a pipe separator and an em dash, how would I do go about doing that? I'm thinking this is something that should be confined to views.py somehow, but things are starting to creep into my templates, and that seems very un-DRY.
  
  
• Finally, this one isn't for a blog, but something I want to do in the future. Let's say you could either work in a text box or upload a file. Would it then be reasonable to have
  

  code:

  class Superclass(models.Model):
      common_attribute1 = models.Field()
      common_attribute2 = models.Field()
  
  class Subclass_With_TextField(Superclass):
      body = models.TextField()
  
  class Subclass_With_FileField(Superclass):
      body = models.FileField()
  

  or is there a better way to do this?
  

Yay fucked around with this message at 21:04 on Feb 19, 2011

def content_type(self):
	type_formencoded = "application/x-www-form-urlencoded"
	ctype = self.request.META.get('CONTENT_TYPE', type_formencoded)
	if ctype.strip().lower().find(type_formencoded) >= 0:
		return None
	return ctype

 mod_wsgi (pid=3500): Exception occurred processing WSGI script 'D:/srv/vhosts/titan/mat/public/run.wsgi'.
 Traceback (most recent call last):
  File "C:\\Program Files (x86)\\Python2.7\\lib\\site-packages\\django\\http\\__init__.py", line 408, in next
     chunk = self._iterator.next()
   File "D:\\libs\\piston\\emitters.py", line 298, in stream_render
     yield self.render(request)
   File "D:\\libs\\piston\\emitters.py", line 352, in render
     self._to_xml(xml, self.construct())
   File "D:\\libs\\piston\\emitters.py", line 277, in construct
     return _any(self.data, self.fields)
   File "D:\\libs\\piston\\emitters.py", line 101, in _any
     raise thing
 TypeError: exceptions must be old-style classes or derived from BaseException, not HttpResponse

def _any(thing, fields=()):
            elif isinstance(thing, HttpResponse):
                return thing
            ...

def _any(thing, fields=()):
            elif isinstance(thing, HttpResponse):
                raise HttpStatusCode(thing)
            ...

multigl fucked around with this message at 06:50 on Feb 22, 2011