CSV Anonymizer v1.0.19

• Improve name pseudonymization so generic name columns with single-token values use name-like replacements instead of generic alphanumeric strings.
• Make full-name replacements compose consistent first/last token pseudonyms and avoid preserving original full-name tokens where the local replacement pool allows it.
• Add a representative people-name CSV fixture and regression tests for name detection, preview output, token consistency, and replacement quality.