Avoid breaking apart testcase name when they contain spaces

[MarcCote]

Reference Issues/PRs

This could be related to what @WuYff observed in #275. In any case, this impacts the current HuggingFace dataset, eg. https://huggingface.co/datasets/princeton-nlp/SWE-bench_Verified/viewer/default/test?row=7

What does this implement/fix? Explain your changes.

The function parse_log_pytest_v2(log: str) produces truncated testcase names when they contain spaces in them, eg.

PASSED astropy/coordinates/tests/test_sky_coord.py::test_passing_inconsistent_coordinates_and_units_raises_helpful_error[kwargs0-Unit 'deg' \\(angle\\) could not be applied to 'distance'. ]

This PR fixes that issue.

Minimal reproducible example

from swebench.harness.log_parsers import parse_log_pytest_v2
result = parse_log_pytest_v2("PASSED astropy/coordinates/tests/test_sky_coord.py::test_passing_inconsistent_coordinates_and_units_raises_helpful_error[kwargs0-Unit 'deg' \\(angle\\) could not be applied to 'distance'. ]")
print(result)

{'astropy/coordinates/tests/test_sky_coord.py::test_passing_inconsistent_coordinates_and_units_raises_helpful_error[kwargs0-Unit': 'PASSED'}

after the fix

{"astropy/coordinates/tests/test_sky_coord.py::test_passing_inconsistent_coordinates_and_units_raises_helpful_error[kwargs0-Unit 'deg' \(angle\) could not be applied to 'distance'. ]": 'PASSED'}