File size: 6,888 Bytes
fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 cd07add 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 fd7800b 5db9737 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 | import re
from typing import Optional
from ..models import (
SyntaxResult,
DNSResult,
DisposableResult,
CatchAllResult,
SMTPResult,
ScoringResult,
RiskLevel,
)
def _is_explicit_mailbox_rejection(smtp: SMTPResult) -> bool:
if smtp.smtp_code != 550 or not smtp.smtp_message:
return False
message = smtp.smtp_message.casefold()
enhanced_status = re.search(r"(?<![\d.])5\.1\.1(?![\d.])", message)
mailbox_rejection_phrases = (
"user unknown",
"no such user",
"mailbox does not exist",
"mailbox not found",
"recipient does not exist",
"recipient not found",
"account does not exist",
)
return bool(enhanced_status) or any(
phrase in message for phrase in mailbox_rejection_phrases
)
def calculate_score(
syntax: SyntaxResult,
dns: Optional[DNSResult],
disposable: Optional[DisposableResult],
catch_all: Optional[CatchAllResult],
smtp: Optional[SMTPResult],
*,
mailbox_verification_skipped: bool = False,
) -> ScoringResult:
"""
Layer 7: Comprehensive deliverability scoring.
Score: 0-100 (higher = better deliverability)
"""
score = 100
risk_cap = 100
breakdown = {}
warnings = []
# ---- Syntax (hard fail → score 0) --------------------------------
if not syntax.passed:
breakdown["SYNTAX_INVALID"] = -100
warnings.append("Email syntax is invalid")
return _scoring_result(0, breakdown, warnings)
breakdown["syntax"] = 0
# ---- Domain/MX -------------------------------------------------------
if dns:
if dns.error_code in {"timeout", "resolver_error"}:
breakdown["DNS_TRANSIENT_FAILURE"] = -50
warnings.append("DNS lookup was inconclusive")
return _scoring_result(50, breakdown, warnings)
elif not dns.domain_exists:
breakdown["DNS_MX_MISSING"] = -100
warnings.append("Domain does not exist")
return _scoring_result(0, breakdown, warnings)
elif dns.error_code == "nxdomain" or (
not dns.has_mx_records and not dns.a_fallback
):
breakdown["DNS_MX_MISSING"] = -100
warnings.append("Domain has no valid mail route")
return _scoring_result(0, breakdown, warnings)
elif dns.a_fallback:
risk_cap = min(risk_cap, 74)
breakdown["mx_records"] = -45
score -= 45
warnings.append(
"No MX records — using A/AAAA fallback; mailbox result is limited"
)
else:
breakdown["domain"] = 0
breakdown["mx_records"] = 0
# Authentication bonuses/penalties
if dns.has_spf:
breakdown["spf"] = 3
score += 3
else:
warnings.append("No SPF record — increases spam risk")
if dns.has_dmarc:
breakdown["dmarc"] = 3
score += 3
else:
warnings.append("No DMARC record configured")
if dns.has_dkim:
breakdown["dkim"] = 2
score += 2
# ---- Disposable ---------------------------------------------------
if disposable:
if disposable.is_disposable:
breakdown["DISPOSABLE_DOMAIN"] = -100
warnings.append("Disposable/temporary email domain detected")
return _scoring_result(0, breakdown, warnings)
elif disposable.is_role_based:
risk_cap = min(risk_cap, 74)
breakdown["role_based"] = -10
score -= 10
warnings.append(
"Role-based email (admin@, info@, etc.) — low engagement risk"
)
elif disposable.is_free_provider:
breakdown["free_provider"] = -5
score -= 5
# ---- ML advisory (DistilBERT, unknown domains only) ---------------
# NEVER a hard verdict: a statistical false positive must not destroy a
# legitimate lead. Penalty is small and reversible by promotion gate.
ml_label = getattr(disposable, "ml_label", None)
if ml_label == "DISPOSABLE" and not disposable.is_disposable:
from ..config import settings # local import: scoring stays pure
penalty = max(1, int(getattr(settings, "ml_score_penalty", 5)))
score -= penalty
breakdown["ML_DISPOSABLE_SUSPECT"] = -penalty
confidence = getattr(disposable, "ml_confidence", 0.0) or 0.0
warnings.append(
f"ML model flagged domain as disposable-like ({confidence:.2f})"
)
# ---- Catch-All (-20) ---------------------------------------------
if catch_all and catch_all.is_catch_all:
risk_cap = min(risk_cap, 74)
breakdown["catch_all"] = -20
score -= 20
warnings.append("Catch-all domain — cannot verify specific mailbox")
# ---- SMTP Verification -------------------------------------------
if smtp:
if smtp.verified is True:
breakdown["smtp"] = 10
score += 10
elif smtp.verified is False:
if _is_explicit_mailbox_rejection(smtp):
breakdown["SMTP_MAILBOX_NOT_FOUND"] = -100
warnings.append("SMTP explicitly reported the mailbox as not found")
return _scoring_result(0, breakdown, warnings)
breakdown["smtp"] = -30
score -= 30
warnings.append(
f"SMTP rejected the recipient (code {smtp.smtp_code}); "
"provider policy may make the result inconclusive"
)
else:
risk_cap = min(risk_cap, 70)
breakdown["SMTP_CHECK_INCONCLUSIVE"] = -30
score -= 30
if smtp.error:
warnings.append(f"SMTP check inconclusive: {smtp.error}")
if mailbox_verification_skipped:
risk_cap = min(risk_cap, 70)
breakdown["FAST_MODE_SKIPPED_MAILBOX_CHECK"] = -30
score -= 30
warnings.append("Mailbox verification was not performed")
# Clamp score between 0 and 100
score = max(0, min(100, score))
if score > risk_cap:
breakdown["RISK_CONTEXT_CAP"] = risk_cap - score
score = risk_cap
return _scoring_result(score, breakdown, warnings)
def _scoring_result(
score: int,
breakdown: dict[str, int],
warnings: list[str],
) -> ScoringResult:
# Determine risk level
if score >= 75:
risk = RiskLevel.LOW
elif score >= 40:
risk = RiskLevel.MEDIUM
else:
risk = RiskLevel.HIGH
return ScoringResult(
score=score, risk_level=risk, breakdown=breakdown, warnings=warnings
)
|