Add output_dir fixture

Adds the output_dir fixture and make the logger fixture use it

Extract from the README

Fixtures

output_dir

A function-scoped fixture that yields a Path to a directory unique to the running test, created under --output-dir (default: ./output).

The path is derived from the test's nodeid, with test_ stripped from file and function names and Test stripped from class names:

nodeid	output path
test_foo.py::test_bar	output/foo/bar
test_foo.py::test_bar[x-y]	output/foo/bar/x-y
test_foo.py::TestSuite::test_bar	output/foo/Suite/bar
sub/test_foo.py::test_bar	output/sub/foo/bar

The directory is created before the test runs. After the test, it is removed only if empty — any files written during the test are preserved.

Options:

• --output-dir — base output directory (default: ./output)
• --output-dir-clean-on-start — wipe the entire output directory at session start

Example:

def test_something(output_dir):
    result = run_thing()
    (output_dir / "result.json").write_text(result)

logger

A function-scoped fixture that yields a Logger bound to the running test, writing to output_dir/test.log.

The logger is named after the test's nodeid (e.g. tests/test_foo.py::test_bar). The log file uses the format %(levelname)s: %(message)s.

Example:

import logging

def test_something(logger):
    logger.info("starting test")
    ...
    logger.info("done")