Added localized PII detection (CNP, PESEL, BSN, NINO) + custom redaction tokens

I just updated the document cleaning pipeline on PrivacyPromptAI to support specific national ID formats (like Romanian CNP, Polish PESEL, Dutch BSN, UK NINO) and added a feature to let users customize token naming (e.g., swapping [NAME_1] out for custom strings or {n} numbering).

Everything runs 100% locally in the browser, no document data leaves the client device, and text/DOCX exports remain free/unlimited.