Pytest-bdd scenario- how to stop execution based on previous scenario failure

[idea1002]

Is there a way to execute pytest-bdd scenario only if previous scenario passed and skip execution if previous scenario failed.

For example: I have 2 pytest-bdd scenarios (Scn A and Scn B) with same marker as @Set1 . I execute tests using this tag
I want to stop execution of second scenario Scn B if Scn A failed.

[elchupanebrej]

Please check pytest plugin https://pytest-dependency.readthedocs.io/en/stable/usage.html
In your case, you have to inject two tests by "scenario" and mark result tests as dependent. This has to work, but in case if wouldn't - please provide info

[rjauregui1]

This Stack Overflow post contains a better solution for including the pytest plugin pytest-dependency into pytest-bdd. Although this might work for most implementations, the pytest-dependency plugin relies on passing arguments as kwargs, and one containing a list.

The Stack Overflow answer provided above makes use of ast to parse the marker. I made some updates to resolve issues with pytest-dependency integration:

Hope this helps others! It took me a while to figure it out.

from typing import Callable, cast
import pytest
import ast

def pytest_bdd_apply_tag(tag: str, function) -> Callable:
    tree = ast.parse(tag)
    body = tree.body[0].value
    if isinstance(body, ast.Call):
        arg = {}
        name = body.func.id
        # Obtain all keywords ('name', 'depends') applied to the marker
        for kw in body.keywords:
            if isinstance(kw.value, ast.List):
                kwvlist = []
                # Obtain the list of keyword values (EX: depends=['TC-1', 'TC-2'] would give ['TC-1', 'TC-2'] as result)
                # Store in a python `list`
                for kwvelts in kw.value.elts:
                    kwvlist.append(kwvelts.value)
            else:
                # Value is an ast.Constant, so we get it's value (EX: name='TC-3')
                kwvlist = kw.value.value
            # Create the `dict` we will use to inject `kwargs` into the `dependency` marker.
            arg.update({kw.arg: kwvlist})
        if len(body.args) > 0:
            arg = body.args[0].value
            mark = getattr(pytest.mark, name).with_args(arg)
        else:
            # Expand the `dict` so it provides kwargs and not `args={}`
            mark = getattr(pytest.mark, name).with_args(**arg)
    else:
        mark = getattr(pytest.mark, tag)
    marked = mark(function)
    return cast(Callable, marked)