Fix pytest node ID parsing for literal separators

[bayandin]

Context

Pytest node IDs use :: as structural separators, but literal :: may also appear in parameter IDs and collected filesystem paths. Splitting raw node ID text therefore misclassified parameter IDs and truncated paths, causing incorrect fullName, titlePath, and suite metadata. Separately, stripping everything after [ from every item name truncated non-function items such as bracketed doctest filenames.

This correction intentionally changes fullName for affected tests and therefore changes the derived testCaseId and historyId values. Test-plan selectors that use the previously malformed names must also be updated.

The parser now derives class and function identity from pytest collector objects, takes filesystem identity from the nearest pytest.File ancestor while retaining the existing fallback for custom items, and preserves non-function item names verbatim.

Regression tests cover the parameter ID foo::bar, the file test_module::part.py, and the doctest file spec[foo].txt.

Checklist

• Sign Allure CLA
• Provide unit tests