Is there a workaround for instantiating fixtures when developing within unit tests?

[mikewojnowicz]

When I'm developing, I would like to write test code that looks like this:

def test_my_function(argument1, argument2):

    result = my_function(argument1, argument2) 

    breakpoint()

and then interact with the arguments of test_my_function.

In particular, after the breakpoint() launches the python debugger, I enter ipython, preserving contents of memory, via a command ipy, which is aliased in my .pdbc to

alias ipy from IPython import embed; user_ns=locals(); user_ns.update(globals()); embed(user_ns=user_ns)

This workflow lets me interact with the arguments of test_my_function -- e.g. to check their type and dimensionality, experiment with various transformations, etc.

The problem is that when argument1 and argument2 are pytest fixtures, my workflow is broken (I believe as of pytest v6.0.0). Previously, I could write argument1=argument1() and interact with argument1. Now I cannot.

Some questions:

• Is there some other way to accomplish what I'm trying to do?
• How is it that other people have not run into a similar problem? To me it seems "natural" to develop by writing unit tests and interacting with the arguments of the test in order to develop further. Pytest's blocking of this strikes me as a fundamental problem.

[The-Compiler]

I'd suggest just running the test via pytest instead. If you're using IPython, you might want to use ipytest to do so.