Files
govoplan-mail/src/govoplan_mail/backend/postbox_bridge.py
T

212 lines
7.6 KiB
Python

from __future__ import annotations
import hashlib
from email import policy
from email.message import Message
from email.parser import BytesParser
from email.utils import getaddresses
from sqlalchemy.orm import Session
from govoplan_core.core.mail import (
MailPostboxBridgeProvider,
MailPostboxBridgeRequest,
MailPostboxBridgeResult,
)
from govoplan_core.core.modules import ModuleContext
from govoplan_core.core.postbox import (
PostboxAttachmentRef,
PostboxDeliveryRequest,
PostboxParticipantRef,
PostboxTargetRef,
postbox_delivery_provider,
)
from govoplan_mail.backend.db.models import MailServerProfile
from govoplan_mail.backend.runtime import configure_runtime, get_registry
MAX_BRIDGE_MESSAGE_BYTES = 50 * 1024 * 1024
MAX_POSTBOX_BODY_CHARS = 500_000
class MailPostboxBridgeError(RuntimeError):
pass
class MailPostboxBridge(MailPostboxBridgeProvider):
"""Translate immutable IMAP observations into native Postbox delivery."""
def bridge_message(
self,
session: object,
request: MailPostboxBridgeRequest,
) -> MailPostboxBridgeResult:
if not isinstance(session, Session):
raise TypeError("Mail Postbox bridging requires a SQLAlchemy session.")
if not isinstance(request.target, PostboxTargetRef):
raise MailPostboxBridgeError("A typed Postbox target is required.")
if len(request.raw_message) > MAX_BRIDGE_MESSAGE_BYTES:
raise MailPostboxBridgeError("Mail message exceeds the Postbox bridge limit.")
profile = session.get(MailServerProfile, request.profile_id)
if profile is None or profile.tenant_id != request.tenant_id:
raise MailPostboxBridgeError("Mail profile not found.")
folder = request.folder.strip()
uid = request.uid.strip()
uidvalidity = request.uidvalidity.strip()
if not folder or not uid or not uidvalidity:
raise MailPostboxBridgeError(
"Mail folder, UIDVALIDITY, and immutable UID are required."
)
try:
message = BytesParser(policy=policy.default).parsebytes(request.raw_message)
except Exception as exc:
raise MailPostboxBridgeError("Mail message could not be parsed.") from exc
provider = postbox_delivery_provider(get_registry())
if provider is None:
raise MailPostboxBridgeError("Postbox delivery is not available.")
source_digest = hashlib.sha256(request.raw_message).hexdigest()
source_key = hashlib.sha256(
f"{request.tenant_id}\0{request.profile_id}\0{folder}\0{uidvalidity}\0{uid}".encode("utf-8")
).hexdigest()
result = provider.deliver(
session,
PostboxDeliveryRequest(
tenant_id=request.tenant_id,
target=request.target,
producer_module="mail",
producer_resource_type="imap_message",
producer_resource_id=source_key,
idempotency_key=f"mail-postbox:{source_key}:{source_digest}",
subject=_header(message, "Subject") or "(No subject)",
body_text=_plain_text_body(message),
sender_label=_header(message, "From"),
classification=request.classification,
participants=_participants(message),
attachments=_attachments(
message,
profile_id=request.profile_id,
folder=folder,
uidvalidity=uidvalidity,
uid=uid,
),
metadata={
**dict(request.metadata),
"transport": "mail-imap",
"mail_profile_id": request.profile_id,
"mailbox_folder": folder,
"mailbox_uidvalidity": uidvalidity,
"mailbox_uid": uid,
"rfc_message_id": _header(message, "Message-ID"),
"raw_sha256": source_digest,
},
),
)
return MailPostboxBridgeResult(
postbox_id=result.postbox_id,
message_id=result.message_id,
delivery_id=result.delivery_id,
duplicate=result.duplicate,
source_digest=source_digest,
)
def _header(message: Message, name: str) -> str | None:
value = " ".join(str(message.get(name) or "").split())
return value[:1000] or None
def _plain_text_body(message: Message) -> str | None:
candidates = message.walk() if message.is_multipart() else (message,)
for part in candidates:
if part.get_content_type() != "text/plain":
continue
if part.get_content_disposition() == "attachment":
continue
try:
value = part.get_content()
except Exception:
payload = part.get_payload(decode=True) or b""
value = payload.decode(part.get_content_charset() or "utf-8", errors="replace")
if isinstance(value, bytes):
value = value.decode(part.get_content_charset() or "utf-8", errors="replace")
normalized = str(value).strip()
if normalized:
return normalized[:MAX_POSTBOX_BODY_CHARS]
return None
def _participants(message: Message) -> tuple[PostboxParticipantRef, ...]:
result: list[PostboxParticipantRef] = []
for kind, headers in (
("sender", ("From",)),
("to", ("To",)),
("cc", ("Cc",)),
("bcc", ("Bcc",)),
):
for name, address in getaddresses(
[str(value) for header in headers for value in message.get_all(header, [])]
):
clean_address = address.strip()[:500]
if not clean_address:
continue
result.append(
PostboxParticipantRef(
kind=kind,
reference_type="external_email",
label=name.strip()[:500] or None,
address=clean_address,
)
)
return tuple(result)
def _attachments(
message: Message,
*,
profile_id: str,
folder: str,
uidvalidity: str,
uid: str,
) -> tuple[PostboxAttachmentRef, ...]:
result: list[PostboxAttachmentRef] = []
for index, part in enumerate(message.walk()):
filename = part.get_filename()
if part.get_content_disposition() != "attachment" and not filename:
continue
payload = part.get_payload(decode=True) or b""
digest = hashlib.sha256(payload).hexdigest()
reference_id = hashlib.sha256(
f"{profile_id}\0{folder}\0{uidvalidity}\0{uid}\0{index}\0{digest}".encode("utf-8")
).hexdigest()
result.append(
PostboxAttachmentRef(
reference_type="mail_attachment",
reference_id=reference_id,
name=str(filename or f"attachment-{index + 1}")[:1000],
media_type=part.get_content_type(),
size_bytes=len(payload),
digest=digest,
metadata={
"mail_profile_id": profile_id,
"mailbox_folder": folder,
"mailbox_uidvalidity": uidvalidity,
"mailbox_uid": uid,
"mime_part_index": index,
},
)
)
return tuple(result)
def create_postbox_bridge(context: ModuleContext) -> MailPostboxBridge:
configure_runtime(registry=context.registry, settings=context.settings)
return MailPostboxBridge()
__all__ = [
"MailPostboxBridge",
"MailPostboxBridgeError",
"create_postbox_bridge",
]