hermes-agent/gateway/relay/adapter.py
Ben Barclay 64a507da44
feat(relay): handle passthrough_forward over the WS (Phase 5 §5.1, gateway half) (#50702)
The connector half (gateway-gateway) moves the passthrough plane's post-ACK
forward off the HTTP gatewayEndpoint onto the gateway's outbound /relay WS via
a new passthrough_forward frame. This is the gateway side: the relay adapter
now RECEIVES and handles that frame, so a hosted gateway (no public IP) can
process forwarded Class-2/3 traffic (Discord interactions, Twilio) over the
socket it already holds — closing the "passthrough inbound doesn't work for
hosted gateways" gap.

- ws_transport.py: decode the passthrough_forward frame; PassthroughForward
  dataclass + _passthrough_from_wire (base64 body -> exact bytes, byte parity
  with the connector's toPassthroughForward); set_passthrough_handler mirrors
  set_interrupt_inbound_handler.
- transport.py: PassthroughHandler type + set_passthrough_handler on the
  RelayTransport protocol.
- adapter.py: connect() wires the passthrough handler; _on_passthrough decodes
  the (already-sanitized, token-free) forward and, for a Discord interaction,
  converts it to a MessageEvent routed through the normal agent path
  (handle_message) — the reply egresses over the outbound / token-less
  follow_up path, so the gateway never holds the interaction credential. Never
  raises (a bad forward can't kill the read loop). Non-discord forwards (Twilio)
  are logged + dropped for now.
- docs/relay-connector-contract.md: document the passthrough_forward frame +
  PassthroughForward shape + §3.1.

The interaction -> MessageEvent CONVERSION semantics (slash-command vs button
UX, option rendering) are the open sub-design flagged in the spec; the TRANSPORT
+ receive mechanism (this) is settled per Ben's Gate-2 decision: "the relay
adapter handles receiving these events over the WS."

Tests (tests/gateway/relay/test_relay_passthrough.py): byte-preservation
round-trip (+ malformed-body tolerance), connect() wiring, application-command
and message-component interactions route through handle_message with correct
session source + scope capture, malformed/non-discord forwards dropped cleanly.
100 relay tests green. Pairs with the connector PR (gateway-gateway).
2026-06-22 20:10:57 +10:00

319 lines
15 KiB
Python

"""RelayAdapter — one generic gateway adapter fronted by the connector. EXPERIMENTAL.
A single ``BasePlatformAdapter`` subclass that, at handshake, receives a
``CapabilityDescriptor`` from the connector telling it which platform it is
fronting and which capabilities to advertise to the ``GatewayStreamConsumer``.
It implements the four abstract methods (``connect`` / ``disconnect`` / ``send``
/ ``get_chat_info``) plus the capability surface (``MAX_MESSAGE_LENGTH``,
``message_len_fn``, ``supports_draft_streaming``) by delegating wire I/O to an
injected transport and reading capabilities off the descriptor.
There is NO per-platform gateway code: the connector is the only side that knows
"this chat_id maps to a Discord channel, send it via the Discord websocket."
The gateway sees an ordinary ``MessageEvent`` in and calls ``adapter.send`` out.
EXPERIMENTAL: the transport protocol and descriptor schema may change without a
deprecation cycle until >=2 Class-1 platforms validate them.
"""
from __future__ import annotations
import logging
from typing import Any, Callable, Dict, Optional
from gateway.config import Platform, PlatformConfig
from gateway.platforms.base import BasePlatformAdapter, MessageEvent, SendResult
from gateway.relay.descriptor import CapabilityDescriptor
from gateway.relay.transport import RelayTransport
from gateway.session import SessionSource
logger = logging.getLogger(__name__)
def _utf16_len(text: str) -> int:
"""Count UTF-16 code units (Telegram's length unit)."""
return len(text.encode("utf-16-le")) // 2
# Table-driven length-unit selection from the descriptor's ``len_unit``.
_LEN_FNS: Dict[str, Callable[[str], int]] = {
"chars": len,
"utf16": _utf16_len,
}
class RelayAdapter(BasePlatformAdapter):
"""Generic relay adapter advertising a connector-negotiated capability profile."""
def __init__(
self,
config: PlatformConfig,
descriptor: CapabilityDescriptor,
transport: Optional[RelayTransport] = None,
) -> None:
# The relay adapter fronts many platforms but presents as a single
# logical platform to the runner; Platform.RELAY identifies it.
super().__init__(config, Platform.RELAY)
self.descriptor = descriptor
self._transport = transport
# Capability surface read by stream_consumer (getattr(..., 4096)).
self.MAX_MESSAGE_LENGTH = descriptor.max_message_length
# chat_id -> guild_id (Discord) / workspace scope, learned from inbound
# events. The connector's egress guard resolves the owning tenant from
# the OUTBOUND action's metadata.guild_id; the gateway's generic delivery
# path (run.py _thread_metadata_for_source) only carries thread_id, so we
# re-attach the scope here from what we saw inbound. Keyed by chat_id
# (channel) since that's what send() receives. See routedEgressGuard.ts.
self._scope_by_chat: Dict[str, str] = {}
self.supports_code_blocks = descriptor.markdown_dialect not in ("", "plain")
# ── capability surface (from descriptor) ─────────────────────────────
@property
def message_len_fn(self) -> Callable[[str], int]:
return _LEN_FNS.get(self.descriptor.len_unit, len)
def supports_draft_streaming(
self,
chat_type: Optional[str] = None,
metadata: Optional[Dict[str, Any]] = None,
) -> bool:
return self.descriptor.supports_draft_streaming
# ── abstract methods (delegated to the transport) ────────────────────
async def connect(self) -> bool:
if self._transport is None:
raise RuntimeError("RelayAdapter has no transport configured")
self._transport.set_inbound_handler(self._on_inbound)
# Inbound interrupts (connector -> owning gateway) arrive as
# interrupt_inbound frames over the SAME outbound WS; bridge them to the
# adapter's interrupt path. WS-only: there is no inbound HTTP receiver.
set_interrupt = getattr(self._transport, "set_interrupt_inbound_handler", None)
if callable(set_interrupt):
set_interrupt(self.on_interrupt)
# Passthrough-plane forwards (Discord interactions, Twilio, …) also ride
# the SAME outbound WS (Phase 5 §5.1) — the connector edge-ACKed and
# forwards the real request here, so a hosted gateway needs no public
# inbound port. Bridge them to the adapter's passthrough handler.
set_passthrough = getattr(self._transport, "set_passthrough_handler", None)
if callable(set_passthrough):
set_passthrough(self._on_passthrough)
ok = await self._transport.connect()
if not ok:
return False
# Negotiate the real capability descriptor from the connector and adopt
# it — the placeholder passed at construction is replaced by what the
# connector advertises for the platform this gateway actually fronts.
try:
descriptor = await self._transport.handshake()
except Exception as exc: # noqa: BLE001 - a failed handshake = a failed connect
logger.warning("relay handshake failed: %s", exc)
return False
self._apply_descriptor(descriptor)
# Inbound (messages + interrupts) is delivered over the outbound WS via
# the connector's relay bus — there is NO inbound HTTP endpoint (hosted
# gateways have no public IP). The transport's reader already dispatches
# `inbound` / `interrupt_inbound` frames to the handlers wired above.
return True
def _apply_descriptor(self, descriptor: CapabilityDescriptor) -> None:
"""Adopt a (re)negotiated descriptor into the live capability surface."""
self.descriptor = descriptor
self.MAX_MESSAGE_LENGTH = descriptor.max_message_length
self.supports_code_blocks = descriptor.markdown_dialect not in ("", "plain")
async def _on_inbound(self, event) -> None:
"""Bridge a connector-delivered MessageEvent into the normal adapter path."""
self._capture_scope(event)
await self.handle_message(event)
def _capture_scope(self, event) -> None:
"""Remember chat_id -> guild scope from an inbound event so our outbound
(the agent's reply) can re-assert it for the connector's egress tenant
resolution. Never raises — scope tracking must not break inbound."""
try:
src = getattr(event, "source", None)
scope = getattr(src, "guild_id", None) if src else None
chat = getattr(src, "chat_id", None) if src else None
if scope and chat:
self._scope_by_chat[str(chat)] = str(scope)
except Exception: # noqa: BLE001 - scope tracking must never break inbound
pass
def _with_scope(self, chat_id: str, metadata: Optional[Dict[str, Any]]) -> Dict[str, Any]:
"""Ensure the outbound metadata carries guild_id for the connector's
egress tenant resolution. The connector resolves the owning tenant from
metadata.guild_id (Discord); without it egress is declined as
'target not routed to an onboarded tenant'. No-op when we have no scope
for this chat (e.g. DMs) or it's already present."""
meta: Dict[str, Any] = dict(metadata or {})
if not meta.get("guild_id"):
scope = self._scope_by_chat.get(str(chat_id))
if scope:
meta["guild_id"] = scope
return meta
async def on_interrupt(self, session_key: str, chat_id: str) -> None:
"""Bridge a connector-delivered /stop into the adapter's interrupt path.
The connector forwards a mid-turn interrupt down the socket owned by
the gateway instance running ``session_key``; this routes it to the
existing per-session interrupt mechanism (sets the
``_active_sessions[session_key]`` Event and clears typing), cancelling
the right turn without touching sibling sessions.
"""
await self.interrupt_session_activity(session_key, chat_id)
async def _on_passthrough(self, forward, buffer_id: Optional[str] = None) -> None:
"""Handle a connector-forwarded passthrough request (Phase 5 §5.1).
The passthrough plane (Discord interactions, Twilio webhooks, …) answers
the provider's latency-critical ACK at the connector EDGE, then forwards
the real, ALREADY-SANITIZED request to this gateway over the outbound WS.
The connector is the trust boundary: it verified the provider signature
at the edge and stripped any shared-identity credential (e.g. a Discord
interaction follow-up token) into its vault — so this body carries no
token, and the agent later acts on it via the token-less ``follow_up``
path (``send_follow_up``), never holding the credential.
For a Discord interaction we decode the (JSON) body and convert it to a
normalized ``MessageEvent`` so it flows through the SAME agent path as a
chat message (``handle_message``); the agent's reply egresses over the
normal outbound/follow_up path. Non-JSON or non-interaction forwards are
logged and dropped for now (Twilio/SMS over the relay is a later unit).
NEVER raises: a malformed forward must not kill the read loop.
NOTE (open semantic sub-design, flagged for review): the interaction ->
MessageEvent mapping below is the v1 default. The exact agent UX for a
slash-command / button interaction (vs. a plain message) — command name
surfacing, option rendering, deferred-vs-immediate response — is the open
piece tracked in the spec; the TRANSPORT + receive mechanism (this whole
path) is settled.
"""
try:
platform = getattr(forward, "platform", "") or ""
if platform == "discord":
event = self._discord_interaction_to_event(forward)
if event is not None:
self._capture_scope(event)
await self.handle_message(event)
return
logger.info(
"relay passthrough_forward dropped (no handler): platform=%s method=%s path=%s",
platform,
getattr(forward, "method", "?"),
getattr(forward, "path", "?"),
)
except Exception: # noqa: BLE001 - a bad forward must never break the reader
logger.warning("relay passthrough_forward handling failed", exc_info=True)
def _discord_interaction_to_event(self, forward):
"""Convert a forwarded Discord interaction body to a MessageEvent, or None.
Builds the session source the same way the connector does for an
interaction (``interactionSessionSource`` on the connector side), so the
agent's session key matches the one the connector bound the follow-up
capability under. Returns None when the body isn't a usable interaction
(e.g. a PING, which the connector already answers at the edge and never
forwards).
"""
import json
from gateway.platforms.base import MessageType
try:
payload = json.loads(bytes(getattr(forward, "body", b"")).decode("utf-8"))
except Exception: # noqa: BLE001
return None
if not isinstance(payload, dict):
return None
# type 1 = PING (answered at the edge, never forwarded); 2 = APPLICATION_COMMAND;
# 3 = MESSAGE_COMPONENT; 5 = MODAL_SUBMIT. Surface a best-effort text.
itype = payload.get("type")
data = payload.get("data") or {}
if itype == 2:
text = str(data.get("name") or "")
elif itype == 3:
text = str(data.get("custom_id") or "")
else:
text = ""
member = payload.get("member") or {}
user = (member.get("user") if isinstance(member, dict) else None) or payload.get("user") or {}
channel_id = str(payload.get("channel_id") or "")
guild_id = payload.get("guild_id")
source = SessionSource(
platform=Platform.RELAY,
chat_id=channel_id,
chat_type="channel" if guild_id else "dm",
user_id=str(user.get("id")) if isinstance(user, dict) and user.get("id") else None,
user_name=str(user.get("username")) if isinstance(user, dict) and user.get("username") else None,
guild_id=str(guild_id) if guild_id else None,
message_id=str(payload.get("id")) if payload.get("id") else None,
)
return MessageEvent(text=text, message_type=MessageType.TEXT, source=source)
async def disconnect(self) -> None:
if self._transport is not None:
await self._transport.disconnect()
async def send(
self,
chat_id: str,
content: str,
reply_to: Optional[str] = None,
metadata: Optional[Dict[str, Any]] = None,
) -> SendResult:
if self._transport is None:
return SendResult(success=False, error="no transport")
result = await self._transport.send_outbound(
{
"op": "send",
"chat_id": chat_id,
"content": content,
"reply_to": reply_to,
"metadata": self._with_scope(chat_id, metadata),
}
)
return SendResult(
success=bool(result.get("success")),
message_id=result.get("message_id"),
error=result.get("error"),
)
async def get_chat_info(self, chat_id: str) -> Dict[str, Any]:
# Proxied to the connector (it owns the platform connection / cache).
if self._transport is None:
return {"name": chat_id, "type": "dm"}
return await self._transport.get_chat_info(chat_id)
async def send_follow_up(
self,
session_key: str,
kind: str,
content: str,
metadata: Optional[Dict[str, Any]] = None,
) -> SendResult:
"""Send via a shared-identity capability bound to a session (A2 outbound).
The gateway never holds the credential: it names the session it is
already in plus the capability ``kind``, and the connector resolves the
real value from its vault and egresses (enforcing the tenant match). Used
e.g. to post a Discord interaction follow-up as the shared bot without
the token ever reaching the gateway. See RelayTransport.send_follow_up.
"""
if self._transport is None:
return SendResult(success=False, error="no transport")
result = await self._transport.send_follow_up(
{
"op": "follow_up",
"session_key": session_key,
"kind": kind,
"content": content,
"metadata": metadata or {},
}
)
return SendResult(
success=bool(result.get("success")),
message_id=result.get("message_id"),
error=result.get("error"),
)