Quick Start

This webpage is a brief introduction to PyTA, mainly intended for students in first-year computer science students at the University of Toronto.

Installation

In PyCharm:

Open up the Settings dialog (Windows: File -> Settings…, Mac: PyCharm -> Preferences).
Go to Project , and select Project Interpreter.
Make sure you have the right version of Python selected in the “Project Interpreter” dropdown.
Click on the green + icon on the right side.
Search for python-ta, and press “Install Package”.

In the command-line:

On Windows, run python -m pip install python-ta. On macOS or Linux, run python3 -m pip install python-ta (as running just python will likely refer to an older version of Python).

Starting out

The easiest way to run PyTA on a file is very similar to doctest. Include the following lines in the if __name__ == '__main__' block of your file:

if __name__ == '__main__':
    import python_ta
    python_ta.check_all()

Then when you run the file, PyTA will check your code in that file. By default, a page will open in your web browser showing a PyTA report.

Running PyTA in the Python interpreter

You can also run PyTA in the Python interpreter (in PyCharm, this is called the Python Console) and run the following command:

>>> import python_ta

Then you can begin checking Python modules. You can check any Python file in the current directory. For example, if you have a file called ‘hello.py’, you can check it as follows:

>>> python_ta.check_all('hello.py')

If you want to check a file which is in a subdirectory of your current location, simply write the ( relative) path to the file.

On Windows, use double backslashes to separate folders:

>>> python_ta.check_all('subfolder1\\sub2\\a\\my_file.py')

On macOS/Linux, use a forward slash instead:

>>> python_ta.check_all('subfolder1/sub2/a/my_file.py')

Errors vs. warnings

PyTA distinguishes between two types of checks:

logical errors and use of forbidden language features
style errors or violating chosen conventions

The output of python_ta.check_all divides the messages into two sections. Note that the headings will always appear, even if you don’t have any errors. If there aren’t any errors listed, the sections will simply be empty.

If you want to only see the logical errors and forbidden features (useful for debugging purposes), use python_ta.check_errors instead of python_ta.check_all:

>>> python_ta.check_errors('hello.py')

Accessing the documentation

We have a separate PyTA webpage describing the errors that PyTA checks for. If you find yourself wondering what an error message means, you can look up the error by its five-character error code (which is always of the form E0401) and look it up on our website. Or, in the Python interpreter you can call the python_ta.doc function on the code:

>>> python_ta.doc('E0401')

This will open up your web browser to the corresponding entry in our documentation page.

Checking contracts

PyTA supports checking preconditions and representation invariants automatically during the execution of a Python program. These are parsed automatically from function/class docstrings, and turned into executable assertions. Type annotations for functions and class instance attributes are also checked. For examples of both the syntax for specifying contracts and resulting behaviour, please check out https://github.com/pyta-uoft/pyta/blob/master/examples/contracts_demo.py.

This kind of checking is completely separate from python_ta.check_all, and can be enabled by inserting the following statements in your if __name__ == '__main__' block:

if __name__ == '__main__':
    import python_ta.contracts as contracts
    contracts.check_all_contracts()

This will add assertions for every function and class in the current module. If you want to assertions to only a subset of the functions/classes, import and use the decorators check_contracts (for functions) and add_class_invariants (for classes) instead.

If you wish to check contracts for your imported modules, pass the module names as arguments to check_all_contracts.