Pseudonymize data like text, email addresses or license tags.
- Running tests
- Developing gocept.pseudonymize
- Change log
This package provides helper functions to pseudonymize data like text, integers, email addresses or license tags.
It uses the crypt.crypt() function for pseudonymization, which means, longer text blocks require multiple crypt.crypt() calls.
gocept.pseudonymize provides single functions for pseudonymization of various data types. Each function takes the value, which should be pseudonymized, and a secret, which is passed as a salt to the crypt module. If secret and value do not change the pseudonymize function returns the exact same result when called again:
>>> import gocept.pseudonymize >>> gocept.pseudonymize.text('Here is my little text', 'secret') 'u7YJWz RqdYkfNUFgZii2Y' >>> gocept.pseudonymize.text('Here is my little text', 'secret') 'u7YJWz RqdYkfNUFgZii2Y'
The result has always the same string length as the input. But there is no guaranty that it is still valid in the domain of the input value. For example the checksum of the pseudonymized IBAN is not correct.
This package is tested to be compatible with Python version 2.7 and 3.3.
There are different pseudonymization function because it is not always possible to guess the correct one by looking at the input data.
For a name use the name function:
>>> gocept.pseudonymize.name('Vimladil', 'secret') 'R5lprkud'
For an address consisting of street and house number use the street function:
>>> gocept.pseudonymize.street('Testweg 34a', 'secret') 'Kui1xre 723'
For an integer value use the integer function:
>>> gocept.pseudonymize.integer(4711, 'secret') 2111
For a decimal value use the decimal function:
>>> from decimal import Decimal >>> gocept.pseudonymize.decimal(Decimal('-123.45'), 'secret') Decimal('-8772.11')
For an email address use the email function:
>>> gocept.pseudonymize.email('email@example.com', 'secret') 'firstname.lastname@example.org'
For an IBAN account number use the iban function:
>>> gocept.pseudonymize.iban('US00123456787650047623', 'secret') 'DE10312010975100119998'
For a BIC (Business Identifier Code) use the bic function:
>>> gocept.pseudonymize.bic('PBNKDEFFXXX', 'secret') 'GTY1BPG8PE2'
For a license tag of a car use the license_tag function:
>>> gocept.pseudonymize.license_tag('HAL-AB 123', 'secret') 'PUD-AM 117'
For a phone number use the phone function:
>>> gocept.pseudonymize.phone('+49 172 34123142', 'secret') '0104118118111676'
For a date use the date function:
>>> from datetime import date >>> gocept.pseudonymize.date(date(1983, 1, 11), 'secret') datetime.date(3021, 1, 18)
For a date represented as string use the datestring function. It takes a format string and keeps zeros date parts as zero.:
>>> gocept.pseudonymize.datestring('00/03/2003', 'secret', format='DD/MM/YYYY') '00/10/7399'
For a time value use the time function:
>>> from datetime import time >>> gocept.pseudonymize.time(time(23, 59, 59), 'secret') datetime.time(13, 11, 49)
There are some additional pseudonymizer functions and helper functions in this package.
|Current change log:|
- Fix phone() so it does not break if the input is only one character long. (https://bitbucket.org/gocept/gocept.pseudonymize/issues/1)
- A value pseudonymized by text() no longer contains full stops, they are converted to spaces. Thus the pseudonymized values may change since version 1.1. (string() now has the former behavior of text(), see below.)
- email() now returns its result in all lower case.
- Add string() pseudonymizer returning a string containing numbers, digits and full stops. (This is what text() formerly did.)
- Fix all pseudonymizers: if called with a value which evaluates to
Falsethe value is returned. But integer() still pseudonymizes
- Fix email() so it does not break on an input value which does not contain
- Add street() pseudonymizer.
- Add bic() (business identifier code) pseudonymizer.
- Add name() pseudonymizer.
- Claim support for PyPy.
- Officially support Python 3.4, 3.5 and 3.6.
- Bring test coverage to 100 % even for code branches and enforce it for the future.
- Re-license from ZPL to MIT.
- Fix handling of usage of glibc2 supported additional encryption algorithms ( signalled using $<id>$<salt>$ as salt).
- Bugfix: text() pseudonymizer now works as expected for texts longer than 11 bytes. Previously it returned an 11 byte result for longer texts ignoring the part after the 11th byte (default behavior of the used crypt implementation). (#1296)
- Fixed handling of
Extended crypt(signalled by starting the salt with an underscore): Salt is now correctly stripped from result. Caution: This leads to different pseudonymization results when using a secret starting with underscore than in version 0.3.
- Fix tests in documentation + testing documentation now.
- Add new pseudonymizers:
- Caution: Due to changed implementation of the date() function it returns different values than in version 0.2.
- date() does not return pseudonymized years smaller than
1900anymore as datetime.date can not handle years smaller that
- Initial release.