Login

Snippets by djypsy

Snippet List

Revisiting Pygments and Markdown

A variation on a theme, inspired by [snippet 39][39] and [snippet 119][119]. The intent is to provide a more generic and simple mechanism for combining [Markdown][markdown] with [Pygments][pygments]. Common scenarios could include blogging or commenting. Snippet 119 seemed too specific and perhaps not as efficient, needing to process the HTML twice to accomplish it's ends. The one snag in the implementation is the need to use a tag other than `code` as a wrapper. See the comments for details. You will need the [BeautifulSoup][soup] module installed. Sample usage: from django.db import models class Blog(models.Model): '''Bare bones blogging model''' title = models.CharField(maxlength=255) slug = models.SlugField(maxlength=255, prepopulate_from=('title',)) pub_date = models.DateTimeField() # the cooked view, cached for quick retrieval blog = models.TextField() # the raw markdown-encoded text, saved for subsequent edits markdown = models.TextField() def save(self): from datetime import datetime if not self.id and not self.pub_date: self.pub_date = datetime.now() self.blog = pygmented_markdown(self.markdown) super(Blog, self).save() [39]: http://www.djangosnippets.org/snippets/39/ [119]: http://www.djangosnippets.org/snippets/119/ [soup]: http://www.crummy.com/software/BeautifulSoup/ [markdown]: http://www.freewisdom.org/projects/python-markdown/Installation [pygments]: http://pygments.org/

  • pygments
  • beautifulsoup
  • markdown
Read More

Using descriptors for lazy attribute caching

Python's [descriptor][1] protocol can seem a bit esoteric at first; however, it can be invaluable in handling everyday idioms and patterns - something that the Django framework authors have taken advantage of in numerous occasions (e.g.: [auth middleware][2]). One such idiom I see and use often and would like to generalize is the attribute-existence check-or-set routine illustrated here: def get_foo(self): if not hasattr(self, '_foo'): self._foo = init_foo() return self._foo Rather than coding this up multiple times (either for a given class or across many unrelated classes), I would prefer to delegate this repetitive work to a descriptor and remain [DRY][3]. The means to this end is implemented as a variation on the Python `property` construct, and is intentionally over simplistic (I leave the details of the heavy lifting up to the reader). The basic premise shown in source here is simply straight-forward Python, a quick and dirty example of how it could be utilized within a Django context is shown here: from django.db import models from cacheprop import CacheProperty2 ARTIFACT_TYPES = ( ('F', _('File')), ('D', _('Directory')), ('A', _('Alias')), ) class Artifact(models.Model): # model fields name = models.CharField(maxlength=64) type = models.CharField(maxlength=1, choices=ARTIFACT_TYPES) file_metadata = CacheProperty2( lambda self: self.filemetadata_set.get(artifact__id=self.id) ) class FileMetadata(models.Model): byte_size = models.IntegerField() artifact = models.ForeignKey(Artifact, unique=True) [1]: http://users.rcn.com/python/download/Descriptor.htm [2]: http://code.djangoproject.com/browser/django/trunk/django/contrib/auth/middleware.py [3]: http://c2.com/cgi/wiki?DontRepeatYourself

  • python
  • descriptors
Read More

djypsy has posted 2 snippets.