dialogue-tester

The Dialogue class

The Dialogue class helps testing REPLs (interactive interpreters with a Read-Eval-Print-Loop).

Given a multi-line interactive session transcript, a Dialogue instance emulates user interactions by offering a fake_input method which can replace the input built-in function in Python 3.

How to use

You can code your REPL function to accept an input function as an optional argument, or you can monkey patch Python's input built-in. Either way, you replace the standard input with the fake_input method bound to a Dialogue instance. Each time it is invoked, this method will print the next prompt and fake user input from the session transcript, then return the same user input string to the REPL for processing.

The Dialogue.session instance attribute holds a multi-line string with all the prompts, user inputs and the outputs expected from the REPL under test. This is used to assert the test case.

See the dialogue_test.py module for two simple REPLs that exercise the Dialogue class.

Note: the .session attribute is normalized with the dialogue.normalize() function. Use that function to normalize the output captured by your testing framework as well. This normalization removes trailing whitespace from each line, preserving only the newlines separating them, and also removes leading and trailing whitespace from the session as a whole. This avoids many issues with linters and IDEs complaining about trailing whitespace, and also makes it easier to edit the session examples in tests. See the tests in this package for use of normalize in test assertions.