FortyDN Dev
v3.0.0: fully autonomous engine — remove MillionVerifier, add DoH DNS worker, AI risk scorer (HF PRO, circuit breaker), i18n, normalized verdicts, remote SMTP worker
de745bf | """Internationalization primitives for the autonomous engine. | |
| Covers the two i18n hard requirements from the architecture spec: | |
| 7.1 Punycode / IDN domains — a Unicode domain (e.g. bücher.example) must be | |
| normalized to Punycode (RFC 3492, "xn--…") before MX resolution and SMTP | |
| probing. The `email-validator` library already does this at Layer 1, but | |
| every other layer (DoH resolver, DNS cache keys, SMTP dialogue) must ALSO | |
| receive the ASCII form, so the conversion lives here as the single source. | |
| 7.2 EAI / SMTPUTF8 local parts — a non-ASCII local part (utf8@bücher.example) | |
| can only be delivered when the receiving MX advertises SMTPUTF8. Without | |
| an SMTP dialogue we cannot observe the advertisement, so the engine must | |
| NOT claim a definitive verdict for these addresses: it refuses the probe | |
| deterministically and reports `unsupported_eai` (honest unknown, never a | |
| fake invalid). | |
| """ | |
| from __future__ import annotations | |
| import idna | |
| EAI_REASON = "unsupported_eai" | |
| def to_punycode(domain: str) -> str: | |
| """Return the ASCII (Punycode) form of a domain, best-effort. | |
| Already-ASCII domains pass through unchanged (fast path). A domain that | |
| cannot be encoded (malformed Unicode, over-long label) is returned as-is — | |
| DNS will then fail naturally and the verdict degrades honestly. | |
| """ | |
| if not domain: | |
| return domain | |
| try: | |
| domain.encode("ascii") | |
| return domain # pure ASCII — nothing to do | |
| except UnicodeEncodeError: | |
| pass | |
| try: | |
| return idna.encode(domain, uts46=True).decode("ascii") | |
| except Exception: | |
| # Last resort: Python's builtin IDNA (RFC 3490 without uts46 mapping) | |
| try: | |
| return domain.encode("idna").decode("ascii") | |
| except Exception: | |
| return domain | |
| def domain_is_unicode(domain: str) -> bool: | |
| """True when the domain needs Punycode (contains non-ASCII labels).""" | |
| try: | |
| domain.encode("ascii") | |
| return False | |
| except UnicodeEncodeError: | |
| return True | |
| def local_part_is_ascii(local_part: str) -> bool: | |
| """True when the local part is plain ASCII (no EAI/SMTPUTF8 needed).""" | |
| try: | |
| (local_part or "").encode("ascii") | |
| return True | |
| except UnicodeEncodeError: | |
| return False | |
| def eai_gate(local_part: str) -> str | None: | |
| """Deterministic EAI gate for the mailbox-probe layers. | |
| Returns None when the local part is ASCII and probing may proceed. | |
| Returns the machine reason ("unsupported_eai") when the address requires | |
| SMTPUTF8 and this deployment cannot honestly verify it. | |
| """ | |
| if local_part_is_ascii(local_part): | |
| return None | |
| return EAI_REASON | |