Skip to content

nshafer/django-hashid-field

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

https://github.com/nshafer/django-hashid-field/actions/workflows/tests.yml/badge.svg?branch=master

WARNING! This module is retired!

This library has been retired. It is no longer recommended for new projects. Existing projects do not require changes, but should consider migrating to a different library. See the RETIRED document for more information.

Django Hashid Field

A custom Model Field that uses the Hashids library to obfuscate an IntegerField or AutoField. It can be used in new models or dropped in place of an existing IntegerField, explicit AutoField, or an automatically generated AutoField.

Features

  • Stores IDs as integers in the database
  • Allows lookups and filtering by hashid string or Hashid object and (optionally) integer.
  • Can enable integer lookups globally or per-field
  • Can be used as sort key
  • Allows specifying a salt, min_length and alphabet globally
  • Supports custom salt, min_length, alphabet, prefix and allow_int_lookup settings per field
  • Allows prefixing hashids with custom string, e.g. prefix="user_" for hashids like "user_h6ks82g"
  • Can drop-in replace an existing IntegerField (HashidField) or AutoField (HashidAutoField)
  • Supports "Big" variants for large integers: BigHashidField, BigHashidAutoField
  • Supports Django 3.2 setting DEFAULT_AUTO_FIELD = 'hashid_field.BigHashidAutoField'
  • Supports Django REST Framework Serializers
  • Supports exact ID searches in Django Admin when field is specified in search_fields.
  • Supports common filtering lookups, such as __iexact, __contains, __icontains, though matching is the same as __exact.
  • Supports subquery lookups with field__in=queryset
  • Supports other lookups: isnull, gt, gte, lt and lte.
  • Supports hashing operations so the fields can be used in Dictionaries and Sets.

Requirements

This module is tested and known to work with:

  • Python 3.7, 3.8, 3.9, 3.10, 3.11, 3.12
  • Django 3.2, 4.2, 5.0
  • Hashids 1.3
  • Django REST Framework 3.14

Please Note: Python 2.x is at its end of life and is no longer supported.

Installation

Install the package (preferably in a virtualenv):

$ pip install django-hashid-field

Configure a global SALT for all HashidFields to use by default in your settings.py. (Note: Using a global salt for all fields will result in IDs from different fields/models being the same. If you want to have unique hashid strings for the same id, then also configure per-field salts as described in Field Parameters below.)

HASHID_FIELD_SALT = "a long and secure salt value that is not the same as SECRET_KEY"
# Note: You can generate a secure key with:
#     from django.core.management.utils import get_random_secret_key; print(get_random_secret_key())

Add it to your model

from hashid_field import HashidField

class Book(models.Model):
    reference_id = HashidField()

Migrate your database

$ ./manage.py makemigrations
$ ./manage.py migrate

Basic Usage

Use your field like you would any other, for the most part. You can assign integers:

>>> b = Book()
>>> b.reference_id = 123
>>> b.reference_id
Hashid(123): OwLxW8D

You can assign valid hashids. It's valid only if it can be decoded into an integer based on your settings:

>>> b.reference_id = 'r8636LO'
>>> b.reference_id
Hashid(456): r8636LO

You can access your field with either hashid strings or Hashid objects:

>>> Book.objects.filter(reference_id='OwLxW8D')
<QuerySet [<Book:  (OwLxW8D)>]>
>>> b = Book.objects.get(reference_id='OwLxW8D')
>>> b
<Book:  (OwLxW8D)>
>>> h = b.reference_id
>>> h
Hashid(123): OwLxW8D
>>> Book.objects.filter(reference_id=h)
<Book:  (OwLxW8D)>

You can lookup objects with integers if you set HASHID_FIELD_ALLOW_INT_LOOKUP = True or allow_int_lookup=True as a parameter to the field.

reference_id = HashidField(allow_int_lookup=True)

Now integer lookups are allowed. Useful if migrating an existing AutoField to a HashidAutoField, but you need to allow lookups with older integers.

>>> Book.objects.filter(reference_id=123)
<QuerySet [<Book:  (OwLxW8D)>]>

By default, the objects returned from a HashidField are an instance of the class Hashid (this can be disabled globally or per-field), and allow basic access to the original integer or the hashid:

>>> from hashid_field import Hashid
>>> h = Hashid(123)
>>> h.id
123
>>> h.hashid
'Mj3'
>>> print(h)
Mj3
>>> repr(h)
'Hashid(123): Mj3'

Hashid Auto Field

Along with HashidField there is also a HashidAutoField that works in the same way, but that auto-increments just like an AutoField.

from hashid_field import HashidAutoField

class Book(models.Model):
    serial_id = HashidAutoField(primary_key=True)

The only difference is that if you don't assign a value to it when you save, it will auto-generate a value from your database, just as an AutoField would do. Please note that HashidAutoField inherits from AutoField and there can only be one AutoField on a model at a time.

>>> b = Book()
>>> b.save()
>>> b.serial_id
Hashid(1): AJEM7LK

It can be dropped into an existing model that has an auto-created AutoField (all models do by default) as long as you give it the same name and set primary_key=True. So if you have this model:

class Author(models.Model):
    name = models.CharField(max_length=40)

Then Django has created a field for you called 'id' automatically. We just need to override that by specifying our own field with primary_key set to True.

class Author(models.Model):
    id = HashidAutoField(primary_key=True)
    name = models.CharField(max_length=40)

And now you can use the 'id' or 'pk' attributes on your model instances:

>>> a = Author.objects.create(name="John Doe")
>>> a.id
Hashid(60): N8VNa8z
>>> Author.objects.get(pk='N8VNa8z')
<Author: Author object>

In Django 3.2 a new setting, "DEFAULT_AUTO_FIELD" was added to change all auto-generated AutoFields to a specific class. This is fully supported with django-hashid-field, and can be enabled with:

DEFAULT_AUTO_FIELD = 'hashid_field.HashidAutoField'
DEFAULT_AUTO_FIELD = 'hashid_field.BigHashidAutoField'

Care must be given, as this will alter ALL models in your project. Usually you would only set this in a new project. Also, since this changes the auto-generated field, only global settings will be used for that field. If you desire specific settings for different models, then using this setting is not advised.

Django may configure the per-app default_auto_field when a new app is created. This will prevent the global setting from taking affect. If you are attempting to use the global setting, you should check your app configs in project/app_name/apps.py to ensure they don't interfere.

from django.apps import AppConfig


class ScheduleConfig(AppConfig):
    default_auto_field = 'django.db.models.BigAutoField'
    name = 'schedule'

Global Settings

HASHID_FIELD_SALT

You can optionally set a global Salt to be used by all HashFields and HashidAutoFields in your project. Do not use the same string as your SECRET_KEY, as this could lead to your SECRET_KEY being exposed to an attacker. Please note that changing this value will cause all HashidFields to change their values, and any previously published IDs will become invalid. Can be overridden by the field definition if you desire unique hashid strings for a given field, as described in Field Parameters below.

Type:

string

Default:

""

Note:

The upstream hashids-python library [only considers the first 43 characters of the salt](davidaurelio/hashids-python#43).

Example:
HASHID_FIELD_SALT = "a long and secure salt value that is not the same as SECRET_KEY"

HASHID_FIELD_MIN_LENGTH

Default minimum length for (non-Big) HashidField and AutoHashidField. It is suggested to use 7 for HashidField and HashidAutoField, so that all possible values (up to 2147483647) are the same length.

Type:

integer

Default:

7

Example:
HASHID_FIELD_MIN_LENGTH = 20

HASHID_FIELD_BIG_MIN_LENGTH

Default minimum length for BigHashidField and BigHashidAutoField. It is suggested to use 13 for BigHashidField and BigHashidAutoField, so that all possible values (up to 9223372036854775807) are the same length.

Type:

integer

Default:

13

Example:
HASHID_FIELD_BIG_MIN_LENGTH = 30

HASHID_FIELD_ALPHABET

The default alphabet to use for characters in generated Hashids strings. Must be at least 16 unique characters.

Type:

string

Default:

"abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890"

Example:
HASHID_FIELD_ALPHABET = "0123456789abcdef"

HASHID_FIELD_ALLOW_INT_LOOKUP

Allow lookups or fetches of fields using the underlying integer that's stored in the database. Disabled by default to prevent users from being to do a sequential scan of objects by pulling objects by integers (1, 2, 3) instead of Hashid strings ("Ba9p1AG", "7V9gk9Z", "wro12zm"). Can be overridden by the field definition.

Type:

boolean

Default:

False

Example:
HASHID_FIELD_ALLOW_INT_LOOKUP = True

HASHID_FIELD_LOOKUP_EXCEPTION

By default any invalid hashid strings or integer lookups when integer lookups are turned off will result in an EmptyResultSet being returned. Enable this to instead throw a ValueError exception (similar to the behavior prior to 2.0).

Type:

boolean

Default:

False

Example:
HASHID_FIELD_LOOKUP_EXCEPTION = True

HASHID_FIELD_ENABLE_HASHID_OBJECT

The default behavior is to return an instance of the Hashid object (described below) in each instance of your Model. This makes it possible to get both the integer and hashid version of the field. However, other django modules, serializers, etc may be confused and not know how to handle a Hashid object, so you can turn them off here. Instead, a string of the hashid will be returned, and a new attribute with the suffix _hashid will be created on each instance with the Hashid object. So if you have key = HashidField(...) then key_hashid will be created on each instance. Can be overriden by the field definition.

Type:

boolean

Default:

True

Example:
HASHID_FIELD_ENABLE_HASHID_OBJECT = False

HASHID_FIELD_ENABLE_DESCRIPTOR

By default a Hashid*Field on a model will replace the original value returned from the database with a Descriptor that attempts to convert values that are set on that field of an instance with a new Hashid object (or string if ENABLE_HASHID_OBJECT is False), regardless if you set an integer or a valid hashid. For the most part this is completely invisible and benign, however if you have issues due to this descriptor, you can disable it here, or on the field, and the raw value will not be replaced with the Descriptor. Can be overriden by the field definition.

Type:

boolean

Default:

True

Example:
HASHID_FIELD_ENABLE_DESCRIPTOR = False

Field Parameters

Besides the standard field options, there are settings you can tweak that are specific to HashidField and AutoHashidField.

Please note that changing any of the values for salt, min_length, alphabet or prefix will affect the obfuscation of the integers that are stored in the database, and will change what are considered "valid" hashids. If you have links or URLs that include your HashidField values, then they will stop working after changing any of these values. It's highly advised that you don't change any of these settings once you publish any references to your field.

salt

Local overridable salt for hashids generated specifically for this field. Set this to a unique value for each field if you want the IDs for that field to be different to the same IDs on another field. e.g. so that book.id = Hashid(5): 0Q8Kg9r and author.id = Hashid(5): kp0eq0V. Suggestion: fieldname = HashIdField(salt="modelname_fieldname_" + settings.HASHID_FIELD_SALT) See HASHID_FIELD_SALT above.

Type:

string

Default:

settings.HASHID_FIELD_SALT, ""

Note:

The upstream hashids-python library [only considers the first 43 characters of the salt](davidaurelio/hashids-python#43).

Example:
reference_id = HashidField(salt="Some salt value")

min_length

Generate hashid strings of this minimum length, regardless of the value of the integer that is being encoded. This defaults to 7 for the field since the maximum IntegerField value can be encoded in 7 characters with the default alphabet setting of 62 characters.

Type:

int

Default:

7

Example:
reference_id = HashidField(min_length=15)

alphabet

The set of characters to generate hashids from. Must be at least 16 characters.

Type:

string of characters

Default:

Hashids.ALPHABET, which is "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890"

Example:
# Only use numbers and lower-case letters
reference_id = HashidField(alphabet="0123456789abcdefghijklmnopqrstuvwxyz")

prefix

An optional string prefix that will be prepended to all generated hashids. Also affects validation, so only hashids that have this prefix will be considered correct.

IMPORTANT: Do not end your prefix with an underscore, as this can cause issues with the Django admin. See Known Issues below.

Type:

String

Default:

""

Example:
# Including the type of id in the id itself:
reference_id = HashidField(prefix="order-")

allow_int_lookup

Local field override for default global on whether or not integer lookups for this field should be allowed. See HASHID_FIELD_ALLOW_INT_LOOKUP above.

Type:

boolean

Default:

settings.HASHID_FIELD_ALLOW_INT_LOOKUP, False

Example:
reference_id = HashidField(allow_int_lookup=True)

enable_hashid_object

Local field override for whether or not to return Hashid objects or plain strings. Can be safely changed without affecting any existing hashids. See HASHID_FIELD_ENABLE_HASHID_OBJECT above.

Type:

boolean

Default:

settings.HASHID_FIELD_ENABLE_HASHID_OBJECT, True

Example:
reference_id = HashidField(enable_hashid_object=False)

enable_descriptor

Local field override for whether or not to use the Descriptor on instances of the field. Can be safely changed without affecting any existing hashids. See HASHID_FIELD_ENABLE_DESCRIPTOR above.

Type:

boolean

Default:

settings.HASHID_FIELD_ENABLE_DESCRIPTOR, True

Example:
reference_id = HashidField(enable_descriptor=False)

Hashid Class

Operations with a HashidField or HashidAutoField return a Hashid object (unless disabled). This simple class does the heavy lifting of converting integers and hashid strings back and forth. There shouldn't be any need to instantiate these manually.

Methods

__init__(value, salt="", min_length=0, alphabet=Hashids.ALPHABET, prefix="", hashids=None):
value:REQUIRED Integer you wish to encode or hashid you wish to decode
salt:Salt to use. Default: "" (empty string)
min_length:Minimum length of encoded hashid string. Default: 0
alphabet:The characters to use in the encoded hashid string. Default: Hashids.ALPHABET
prefix:String prefix prepended to hashid strings. Default: "" (empty string)
hashids:Instance of hashids.Hashids to use for encoding/decoding instead of instantiating another.

Read-Only Properties

id
type:Int
value:The decoded integer
hashid
type:String
value:The encoded hashid string
hashids
type:Hashids()
value:The instance of the Hashids class that is used to encode and decode
prefix
type:String
value:The prefix prepended to hashid strings

Django REST Framework Integration

If you wish to use a HashidField or HashidAutoField with a DRF ModelSerializer, there is one extra step that you must take. Automatic declaration of any Hashid*Fields will result in an ImproperlyConfigured exception being thrown. You must explicitly declare them in your Serializer, as there is no way for the generated field to know how to work with a Hashid*Field, specifically what 'salt', 'min_length' and 'alphabet' to use, and can lead to very difficult errors or behavior to debug, or in the worst case, corruption of your data. Here is an example:

from rest_framework import serializers
from hashid_field.rest import HashidSerializerCharField


class BookSerializer(serializers.ModelSerializer):
    reference_id = HashidSerializerCharField(source_field='library.Book.reference_id')

    class Meta:
        model = Book
        fields = ('id', 'reference_id')


class AuthorSerializer(serializers.ModelSerializer):
    id = HashidSerializerCharField(source_field='library.Author.id', read_only=True)

    class Meta:
        model = Author
        fields = ('id', 'name')

The source_field allows the HashidSerializerCharField to copy the 'salt', 'min_length' and 'alphabet' settings from the given field at app_name.model_name.field_name so that it can be defined in just one place. Explicit settings are also possible:

reference_id = HashidSerializerCharField(salt="a different salt", min_length=10, alphabet="ABCDEFGHIJKLMNOPQRSTUVWXYZ")

If nothing is given, then the field will use the same global settings as a Hashid*Field. It is very important that the options for the serializer field matches the model field, or else strange errors or data corruption can occur.

HashidSerializerCharField will serialize the value into a Hashids string, but will deserialize either a Hashids string or integer and save it into the underlying Hashid*Field properly. There is also a HashidSerializerIntegerField that will serialize the Hashids into an un-encoded integer as well.

Primary Key Related Fields

Any models that have a ForeignKey to another model that uses a Hashid*Field as its Primary Key will need to explicitly define how the PrimaryKeyRelatedField should serialize and deserialize the resulting value using the pk_field argument. If you don't you will get an error such as "Hashid(60): N8VNa8z is not JSON serializable". We have to tell DRF how to serialize/deserialize Hashid*Fields.

For the given Author model defined above that has an id = HashidAutoField(primary_key=True) set, your BookSerializer should look like the following.

from rest_framework import serializers
from hashid_field.rest import HashidSerializerCharField


class BookSerializer(serializers.ModelSerializer):
    author = serializers.PrimaryKeyRelatedField(
        pk_field=HashidSerializerCharField(source_field='library.Author.id'),
        read_only=True)

    class Meta:
        model = Book
        fields = ('id', 'author')

Make sure you pass the source field to the HashidSerializer*Field so that it can copy the 'salt', 'min_length' and 'alphabet' as described above.

This example sets read_only=True but you can explicitly define a queryset or override get_queryset(self) to allow read-write behavior.

author = serializers.PrimaryKeyRelatedField(
    pk_field=HashidSerializerCharField(source_field='library.Author.id'),
    queryset=Author.objects.all())

For a ManyToManyField, you must also remember to pass many=True to the PrimaryKeyRelatedField.

HashidSerializerCharField

Serialize a Hashid*Field to a Hashids string, de-serialize either a valid Hashids string or integer into a Hashid*Field (if allow_int_lookup is enabled.)

Parameters

source_field

A 3-field dotted notation of the source field to load matching 'salt', 'min_length' and 'alphabet' settings from. Must be in the format of "app_name.model_name.field_name". Example: "library.Book.reference_id".

salt, min_length, alphabet, prefix, allow_int_lookup

See Field Parameters

HashidSerializerIntegerField

Serialize a Hashid*Field to an integer, de-serialize either a valid Hashids string or integer into a Hashid*Field. See HashidSerializerCharField for parameters.

Please Note: This field will always serialize to an integer and thus will also de-serialize integers into valid objects, regardless of the allow_int_lookup setting.

Known Issues

'Hashid' object is not iterable

With Django 5.0, attempting to filter on a field that is a ForeignKey to another model that uses a Hashid*Field as its primary key will result in an error such as "'Hashid' object is not iterable". The workaround is to specify the exact field of the related model to filter on. e.g. instead of list_filter = ['author'] use list_filter = ['author__name'].

Admin can't find objects with a prefix that ends in an underscore

The Django admin has a quote and unquote function that attempts to encode non-alphanumeric characters in primary keys using the format of _xx where xx is the hex value of the character. This can be a problem if you are using a prefix that ends in an underscore, as there's a chance that your Hashid primary key will be something like, "user_3Ej8Kjm". The Django admin will attempt to unquote that to "user>j8Kjm" by swapping the "_3E" with a ">" character, but that isn't valid and so it can't find the object. It will throw an error like, "Admin Object with id 'prefix<2345678' doesn't exist. Perhaps it was deleted?"

The workaround is to not use a prefix that ends in an underscore. You can end it with a character right after the underscore that can't be used in hexidecimal, though, so it still looks right. e.g. prefix="user_g" so that the above example would actually be "user_g3Ej8Kjm", and "_3E" can't be matched. Or you can use any other character, such as a hyphen, .e.g prefix="user-".

Development

Here are some rough instructions on how to set up a dev environment to develop this module. Modify as needed. The sandbox is a django project that uses django-hashid-id, and is useful for developing features with.

  • git clone https://github.com/nshafer/django-hashid-field.git && cd django-hashid-field
  • mkvirtualenv -a . -p /usr/bin/python3 -r sandbox/requirements.txt django-hashid-field
  • python setup.py develop
  • sandbox/manage.py migrate
  • sandbox/manage.py createsuperuser
  • sandbox/manage.py loaddata authors books editors
  • sandbox/manage.py runserver
  • python runtests.py

For any pull requests, clone the repo and push to it, then create the PR.

To install the latest development version, use:

` pip install git+https://github.com/nshafer/django-hashid-field.git `

LICENSE

MIT License. You may use this in commercial and non-commercial projects with proper attribution. Please see the LICENSE