[AHMETHAKANBEZIR1]

lziv_complexity dispatches unsigned integer arrays through the string branch instead of the documented numeric-symbol branch. Decimal representations are then concatenated, changing the sequence and its normalization. For example, [10, 11, 10] as uint8 gives raw complexity 4 and normalized complexity about 1.7233, while the same integer symbols as int64 give 3 and about 1.5850.

Include NumPy's unsigned-integer dtype kind alongside booleans, floats and signed integers in the existing conversion to uint32. This preserves the existing numeric casting policy and string-input behavior; it does not change the LZ parsing algorithm or the handling of values outside uint32's range.

Validation on Windows CPU, Python 3.12.14:

• 32 regression cases cover uint8/16/32/64, raw/normalized output, binary and relabeled multi-digit alphabets, and two independently known sequence complexities. On unmodified master 6843920, 12 fail and 20 controls pass; with the fix all 32 pass. Inputs are checked for mutation.
• Full suite: 66 passed with NumPy 1.26.4, and 66 passed with NumPy 2.5.4 (Numba 0.68.0, SciPy 1.17.1, scikit-learn 1.9.1).
• Repository Ruff lint/format checks and whitespace check passed.

Other operating systems, doctests and performance benchmarks were not run.

AI disclosure: OpenAI Codex autonomously investigated and implemented the fix/tests, wrote this description and ran the reported checks. No independent human pre-submission review has been performed.

[raphaelvallat]

Thank you for your contribution!

AntroPy now has guidelines for pull requests, including specific steps for AI-assisted PRs: #61

Before I review this PR, could you please go through those steps and update the PR accordingly? In particular: state the user-impact rating (P1-P3) with a short justification, start the description with a summary in your own words, report the exact AI model used (if any), and run an AI review and post a summary of it as a comment. Thanks!