Skip to main content

Generate randomized strings of characters using a template.

Project description

Generate randomized strings of characters using a template.

This is a Python module that enables a user to generate test data, unique ids, passwords, vouchers or other randomized data very quickly using a template language. The template language is superficially similar to regular expressions but instead of defining how to find or capture strings, it defines how to generate randomized strings.

An example template for generating a strong password:


will generate something like the following:



from strgen import StringGenerator


from strgen import StringGenerator


from strgen import StringGenerator

The latter produces a list of 10 strings that are unique within the list.


>>> from strgen import StringGenerator
>>> StringGenerator('[\l\d]{4:18}&[\d]&[\p]').render()

The template is a string that is a sequence of one or more of the following:

  • Literal text

  • Character class

  • Group

In more detail:

Literal: <any string>

Any literal string.



Special characters need to be escaped with backslash \\.

Character class: [<class specification>]

Much like in regular expressions, it uses strings of characters and hyphen for defining a class of characters.



Randomly choose characters from the set of lower case letters, digits and the underscore.

Character Set Codes

  • \W: whitespace + punctuation

  • \a: ascii_letters

  • \c: lowercase

  • \d: digits

  • \h: hexdigits

  • \l: letters

  • \o: octdigits

  • \p: punctuation

  • \r: printable

  • \s: whitespace

  • \u: uppercase

  • \w: _ + letters + digits

Quantifier: {x:y}

Where x is lower bound and y is upper bound. This construct must always follow immediately a class with no intervening whitespace. It is possible to write {:y} as a shorthand for {0:y} or {y} to indicate a fixed length.



Generates a string from zero to 8 in length composed of lower case alphabetic characters


Generates a string with either four lower case alphabetic characters or a string that is 9 digits in length.

Using a character class and no quantifier will result in a quantifier of 1. Thus:


will result always in either a, b, or c.

Group: (<group specification>)

A group specification is a collection of literals, character classes or other groups divided by the OR operator | or the permutation operator &.

OR Operator

The binary | operator can be used in a group to cause one of the operands to be returned and the other to be ignored with an even chance.

Permutation Operator

The binary & operator causes its operands to be combined and permuted. This addresses the use case for many password requirements, such as, “at least 6 characters where 2 or more are digits”. For instance:


If a literal or a group is an operand of the permutation operator, it will have its string value permuted with the other operand.


will produce strings like:


Concatenation and Operators

Classes, literals and groups in sequence are concatenated in the order they occur. Use of the | or & operators always bind the operands immediately to the left and right:


produces something like:


In otherwords, the digits occur first in sequence as expected. This is equivalent to this:


Special Characters, Escaping and Errors

There are fewer special characters than regular expressions:


They can be used as literals by escaping with backslash. All other characters are treated as literals. The hyphen is only special in a character class, when it appears within square brackets. The template parser tries to raise exceptions when syntax errors are made, but not every error will be caught, like having space between a class and quantifier.

Character Classes and Quantifiers

Use a colon in the curly braces to indicate a range. There are sensible defaults:

[\w]       # randomly choose a single word character
[\w]{0:8}  # generate word characters from 0-8 in length
[\w]{:8}   # a synonym for the above
[\w]{8}    # generate word characters of exactly 8 in length
[a-z0-9]   # generate a-z and digits, just one as there is no quantifier
[a-z0-9_!@]  # you can combine ranges with individual characters

Here’s an example of generating a syntactically valid but, hopefully, spurious email address:


The first name will be exactly 10 lower case characters; the last name will be 5-10 characters of lower case letters, each separated by either a dot or underscore. The domain name without domain class will be 3 - 12 lower case characters and the domain type will be one of ‘.com’,’.net’,’.org’.

Using multiple character set codes repeatedly will increase the probability of a character from that set occuring in the result string:


This will provide a string that is three times more likely to contain a digit than if this were used:



When using the unique=True flag, it’s possible the generator cannot possibly produce the required number of unique strings. For instance:


This will generate an exception but not before attempting to generate the strings.

The number of times the generator needs to render new strings to satisfy the list length and uniqueness is not determined at parse time. However, the maximum number of times it will try is by default n x 10 where n is the requested length of the list. Therefore, taking the above example, the generator will attempt to generate the unique list of 0’s and 1’s 100 x 10 = 1000 times before giving up.


Unicode is supported for both the template and output.

Character Sets and Locale

Character sets used for backslashed character codes are exactly the Python character sets from the string package. Some of these are dependent on the locale settings.

Randomness Methods

The generator tries to use random.SystemRandom() for randint, choice, etc. It falls back to random.randint and associated methods if it can’t use SystemRandom.

Rationale and Design Goals

In Python, the need to generate random strings comes up very frequently and is accomplished usually (though not always) via something like the following code snippet:

import random
import string
''.join(random.choice(string.ascii_uppercase + string.digits) for x in range(10))

This generates a string that is 10 characters made of uppercase letters and digits. Unfortunately, this solution becomes cumbersome when real-world requirements are added. Take for example, the typical requirement to generate a password: “a password shall have 6 - 20 characters of which at least one must be a digit and at least one must be a special character”. The above solution then becomes much more complicated and changing the requirements is an error-prone and unnecessarily complex task.

The equivalent using the strgen package is the following:

from strgen import StringGenerator as sg

strgen is far more compact, flexible and logical than using the standard solution:

  • It tries to use a better entropy mechanism and falls back gracefully if this is not available.

  • The user can easily modify the specification (template) with minimal effort and maximum precision.

  • Modifications to the template are simpler and far less error prone than writing all the code necessary to implement changes in a random string specification

  • It covers a broader set of use cases: unique ids, persistent unique filenames, test data, etc.

  • The template syntax is very easy to learn for anyone familiar with regular expressions while being much simpler.

  • It supports unicode.

  • It proposes a standard way of expressing common requirements, like “a password shall have 6 - 20 characters of which at least one must be a digit and at least one must be a special character”:


This package is designed with the following goals in mind:

  • Provide an abstract template language that does not depend on a specific implementation language.

  • For Python, reduce dependencies on other packages.

  • Keep syntax as simple as possible while being pragmatic.

  • Provide an implementation design with associated behaviour that strikes the right balance between ease-of-implementation and ease-of-use.

  • Superficially similar to regular expressions to enable developers to quickly pick up the template syntax.

  • Support non-ASCII languages (unicode).


Released under the BSD license.

Original Author:

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

StringGenerator-0.1.tar.gz (4.9 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page