Runtime verification and automated testing for scientific code
Paranoid Scientist is a Python module which allows runtime verification of entry and exit conditions for Python functions using a novel type system which prioritizes the interpretation of types over their representation. It is intended for scientific software. More specifically, it provides the following:
- A novel type system, which emphasizes the meaning of the type instead of the data structure of the type.
- Verification of arbitrary entry and exit conditions, including more complex expressions with universal quantification.
- Automated testing of individual functions to determine, before execution of the program, whether functions conform to their specification.
- A simple and clear function decorator notation
- Python 3.5 or above
- Optional: Numpy (for Numpy types support)
All code is available under the MIT license. See LICENSE.txt for more information.
What types are included by default?
- Numeric types:
- Numeric: Any real number, plus inf/-inf/nan.
- ExtendedReal: Any real number plus inf/-inf.
- Number: Any real number.
- Integer: An integer.
- Natural0: A natural number including 0.
- Natural1: A natural number starting with 1.
- Range(x,y): A number in the interval [x,y].
- RangeClosedOpen(x,y): A number in the interval [x,y).
- RangeOpenClosed(x,y): A number in the interval (x,y].
- RangeOpen(x,y): A number in the range (x,y).
- Positive0: A positive number or zero.
- Positive: A positive number excluding zero.
- NDArray(d=None, t=None): A Numpy NDArray. Optionally, require
it to have
ddimensions. Optionally, require elements of the array to be of type
t(which can be any Paranoid type).
- Unchecked(t=None): Do not check the type for this element.
Optionally, a Paranoid type
tcan be passed as a type to generate during automated testing (to allow testing of functions of unchecked type).
- Constant(v): Can only be the value
- Self: A special type to use inside classes for the
- Nothing: Can only be None.
- Function: Any Python function.
- Boolean: Either True or False.
- And(t1, t2, ...): Any number of Paranoid types may be passed as arguments. Require a value to satisfy all of these types.
- Or(t1, t2, ...): Any number of Paranoid types may be passed as arguments. Require a value to satisfy at least one of these types.
- Not(t): Can be anything other than the Paranoid type
- Unchecked(t=None): Do not check the type for this element. Optionally, a Paranoid type
- Set(els): The argument
elsshould be a list of accepted values. This is equivalent to an enum. For example, Set([0, 1, 2, 3]) accepts only these four numbers.
- List(t): A Python list with elements of Paranoid type
- Dict(k, v): A Python dictionary where keys have Paranoid type
kand values have Paranoid type
- Set(els): The argument
- String: Any string.
- Identifier: Any non-empty string containing only alpha-numeric characters plus underscores and hyphens.
- Alphanumeric: Any non-empty alphanumeric string.
- Latin: Any non-empty string with Latin characters only.
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size & hash SHA256 hash help||File type||Python version||Upload date|
|paranoid_scientist-0.1.4-py3-none-any.whl (23.1 kB) Copy SHA256 hash SHA256||Wheel||py3||Jul 20, 2018|
|paranoid-scientist-0.1.4.tar.gz (19.1 kB) Copy SHA256 hash SHA256||Source||None||Jul 20, 2018|