OPT: speed up sliding staypoint generation

[hongyeehh]

No description provided.

[Copilot]

Pull request overview

This PR optimizes the generate_staypoints(..., method="sliding") pipeline by reducing per-user overhead (avoiding joblib when n_jobs==1) and by replacing repeated Shapely operations inside the sliding-window loop with NumPy-based computations.

Changes:

• Add a fast path for n_jobs==1 that iterates per-user groups directly with tqdm progress.
• Speed up per-user sliding staypoint generation by precomputing timestamp integer arrays, trig arrays for distance, and computing centroids from coordinate arrays (matching prior union/centroid behavior).
• Add regression tests for timezone-aware, non-nanosecond timestamp units and for centroid behavior (duplicate points + antimeridian wrapping).

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
trackintel/preprocessing/positionfixes.py	Introduces optimized per-user sliding staypoint generation (faster distance + centroid computation, avoids parallel overhead for n_jobs==1).
tests/preprocessing/test_positionfixes.py	Adds tests to ensure the optimized path preserves behavior for timestamp dtypes and centroid computation edge cases.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.