Jumped the gun--don't want a custom field.
This commit is contained in:
parent
d349b51923
commit
89d859e371
1 changed files with 0 additions and 35 deletions
|
@ -1,35 +0,0 @@
|
|||
import re
|
||||
from unicodedata import normalize
|
||||
|
||||
from django.db import models
|
||||
|
||||
|
||||
class NaturalSortField(models.CharField):
|
||||
"""
|
||||
A custom model field specifically for storing a normalized unicode string,
|
||||
with removed starting articles, for use in natural sorting.
|
||||
|
||||
Code is a modified version of 'django-naturalsortfield' by Nathan Reynolds:
|
||||
https://github.com/nathforge/django-naturalsortfield
|
||||
"""
|
||||
def __init__(self, for_field, **kwargs):
|
||||
self.for_field = for_field
|
||||
kwargs.setdefault('db_index', True)
|
||||
kwargs.setdefault('editable', False)
|
||||
kwargs.setdefault('max_length', 255)
|
||||
super(NaturalSortField, self).__init__(**kwargs)
|
||||
|
||||
def pre_save(self, model_instance, add):
|
||||
return self.naturalize(getattr(model_instance, self.for_field))
|
||||
|
||||
def naturalize(self, string):
|
||||
def naturalize_int_match(match):
|
||||
return '%08d' % (int(match.group(0)),)
|
||||
|
||||
string = normalize('NFKD', string).encode('ascii', 'ignore').decode('ascii')
|
||||
string = string.lower()
|
||||
string = string.strip()
|
||||
string = re.sub(r'^(a|an|the)\s+', '', string)
|
||||
string = re.sub(r'\d+', naturalize_int_match, string)
|
||||
|
||||
return string
|
Loading…
Reference in a new issue