Future-proofing the global code

[sergeyf]

Frees us from Python 3.8...

[Copilot]

Pull Request Overview

This PR enhances multiprocessing support by ensuring global variables are correctly initialized in worker processes, standardizing pool creation with the spawn context, and refactoring ternary and comprehension formatting. It also adds tests to verify consistency and fallback behavior when using multiple processes.

• Introduce _init_pool helpers and switch to multiprocessing.get_context("spawn") in both featurizer and data modules
• Refactor multiline ternary expressions and list comprehensions for readability
• Add comprehensive tests to confirm single‐ vs. multi‐process output consistency and fallback logic

Reviewed Changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.

File	Description
tests/test_featurizer.py	New tests for featurization consistency, global dataset init, and fallbacks
tests/test_data.py	New tests for preprocessing consistency, global var init, and fallbacks
s2and/featurizer.py	Use spawn context, add _init_pool, refactor feature‐tuple formatting
s2and/data.py	Use spawn context, add _init_pool, refactor comprehension formatting

Comments suppressed due to low confidence (1)

tests/test_data.py:230

• [nitpick] The test verifies key fields but omits other derived attributes (e.g., title_ngrams_chars, venue_ngrams); consider extending assertions to cover all preprocessing outputs for full consistency checks.

            assert paper_single.title == paper_multi.title, f"Title mismatch for paper {paper_id}"