Transparent obfuscation of numeric IDs
Project description
TurbID
Overview
TurbID is a Python library that provides ID obfuscation and encryption for sequential integer primary keys. With TurbID, your database can store clear, sequential integer primary keys, while your public API and the rest of the world sees opaque and seemingly random long-form string IDs. This approach avoids the database performance downsides of long random IDs and sidesteps the privacy and security risks of clear integer IDs.
Unlike other libraries that merely encode integers with a randomized alphabet, TurbID uses format-preserving encryption for additional security.
TurbID currently supports SQLAlchemy with an optional extension that provides a custom column type, but it can be extended to work with other ORMs or frameworks.
[!WARNING]
TurbID is not intended for protecting sensitive numeric data, such as credit card numbers or PINs. For these use cases, please use standard, secure encryption methods.
Installation
TurbID is compatible with Python 3.8+ and available on PyPI. Install it with pip, or your package manager of choice:
pip install turbid
But you probably want to install with the optional SQLAlchemy extension:
pip install turbid[sqlalchemy]
SQLAlchemy Usage
With SQLAlchemy, just replace your column's Integer
column type with TurbIDType
:
class User(Base):
__tablename__ = "user"
user_id = sa.Column(TurbIDType(key=KEY, tweak="user"), primary_key=True)
name = sa.Column(sa.String(200))
If you have foreign keys, do the same for the ForeignKey
columns, but remember
to use the same key
and tweak
values as the referenced column:
class Post(Base):
__tablename__ = "post"
post_id = sa.Column(TurbIDType(key=KEY, tweak="post"), primary_key=True)
user_id = sa.Column(TurbIDType(key=KEY, tweak="user"), sa.ForeignKey("user.user_id"))
title = sa.Column(sa.String(200))
You can use your columns as usual, in joins, filters, data retrieval, etc. In queries or when updating data you can use either the original integer ID or the obfuscated string ID, but retrievals will always return the obfuscated string ID.
TurbIDCipher Usage
If you don't use SQLAlchemy or you want to encrypt/decrypt IDs at another layer
of your application, like when serializing objects for responses, you can use
the TurbIDCipher
class directly.
>>> from turbid import TurbIDCipher
>>> import secrets
>>>
>>> key = secrets.token_hex()
>>> tweak = "my_table_name"
>>> obscure_id = TurbIDCipher(key, tweak=tweak)
>>>
>>> # Encrypt an integer ID
>>> encrypted_id = obscure_id.encrypt(12345)
>>> print(f"Encrypted ID: {encrypted_id}")
Encrypted ID: VTxLWjgdCWGjLSIiZtCQCMvu
>>>
>>> # Decrypt the ID back to the original integer
>>> original_id = obscure_id.decrypt(encrypted_id)
>>> print(f"Original ID: {original_id}")
Original ID: 12345
Parameters
The required parameters are:
- key:
- A string that will be hashed to generate the encryption key for the AES cipher.
- Never expose the key in version control or share it publicly.
- You can generate a random key suitable for this purpose using the
secrets.token_hex
function.
- tweak:
- A string that will be hashed to generate the 7-byte tweak value for FF3-1 encryption.
- This parameter is not a secret and is used to differentiate the encrypted
values of different instances using the same secret
key
. - The value must be unique per table to avoid ID collisions.
SQLAlchemy extension parameters:
- prefix:
- An optional string that will be prepended to the encrypted ID. This is also
used as the
tweak
value if an explicit one isn't provided. - If you don't want a prefix on your encrypted IDs, you must provide a
tweak
for each table. - You can provide both a
tweak
and aprefix
. In this case, theprefix
will be merely cosmetic and thetweak
will be used to differentiate the encrypted values.
- An optional string that will be prepended to the encrypted ID. This is also
used as the
Optional parameters with tested defaults:
- length=
24
:- The length of the encrypted ID. The default is 24 characters, but the minimum and maximum lengths are determined by the alphabet length and the maximum value you want to encrypt.
- A
ValueError
will be raised if the alphabet length is incompatible with the specified length.
- alphabet=
string.digits + string.ascii_letters
:- The alphabet used to encode the encrypted ID. The default is all 10 digits and all 52 lowercase and uppercase letters.
- You can use a different alphabet, as long as it contains all 10 digits and no repeated characters.
- The alphabet length must be compatible with the specified
length
.
- key_length=
128
:- The length of the AES key in bits. The default is 128 bits, which is the recommended key length for AES encryption.
- Note this refers to the length of the hashed key generated internally, not the length of the string you provide as the key material.
Compatibility and Testing
TurbID is tested with the following values:
- input ids: sampled from
0
to2^63-1
length
:20
to32
, inclusivealphabet
:string.digits + string.ascii_letters
and"0123456789abcdef"
key_length
:128
,194
, and256
It probably works with other values, but you should review the limitations of the FF3-1 algorithm and the ff3 library and implement tests to ensure it works as expected.
License
TurbID is licensed under the MIT License. See the LICENSE file for details.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file turbid-0.1.1.tar.gz
.
File metadata
- Download URL: turbid-0.1.1.tar.gz
- Upload date:
- Size: 6.7 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: poetry/1.8.3 CPython/3.10.9 Linux/5.15.0-117-generic
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 5dbf4dd92f3ceead22067b3698f99f6b0cfee0386d72da6b9c63188d3a3375a6 |
|
MD5 | 7123dd52ff8e027e5d8d8a8ca5882b1d |
|
BLAKE2b-256 | 2803145a14d7543833ddc8ae450921c040e2855e1c5f6ccb8a6ea066745368d8 |
File details
Details for the file turbid-0.1.1-py3-none-any.whl
.
File metadata
- Download URL: turbid-0.1.1-py3-none-any.whl
- Upload date:
- Size: 7.3 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: poetry/1.8.3 CPython/3.10.9 Linux/5.15.0-117-generic
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | bf5c62e41ab8dcd1072c2b274c800a94fd0a964903717a30f6c57253a9e9e4fa |
|
MD5 | 9e8fbb88587e956e0c1a91ce44d593ac |
|
BLAKE2b-256 | 2593c02109d6b361b8bcf87ef7cf3de53e2bc09582b09a57a0e2d0acfde6fab0 |