feat: pydantic models and inter-process messaging

Moved `InternalMessage` into schemas and created a `BeliefMessage`
model. Also added the ability for agents to communicate via ZMQ to
agents on another process.

ref: N25B-316

src/control_backend/agents/bdi/bdi_core_agent/bdi_core_agent.py

@@ -1,16 +1,17 @@
 import asyncio
 import copy
-import json
 import time
 from collections.abc import Iterable
 
 import agentspeak
 import agentspeak.runtime
 import agentspeak.stdlib
+from pydantic import ValidationError
 
 from control_backend.agents.base import BaseAgent
 from control_backend.core.agent_system import InternalMessage
 from control_backend.core.config import settings
+from control_backend.schemas.belief_message import BeliefMessage
 from control_backend.schemas.ri_message import SpeechCommand
 
 
@@ -58,16 +59,19 @@ class BDICoreAgent(BaseAgent):
             maybe_more_work = True
             while maybe_more_work:
                 maybe_more_work = False
+                self.logger.debug("Stepping BDI.")
                 if self.bdi_agent.step():
                     maybe_more_work = True
 
                 if not maybe_more_work:
                     deadline = self.bdi_agent.shortest_deadline()
                     if deadline:
+                        self.logger.debug("Sleeping until %s", deadline)
                         await asyncio.sleep(deadline - time.time())
                         maybe_more_work = True
                     else:
                         self._wake_bdi_loop.clear()
+                        self.logger.debug("No more deadlines. Halting BDI loop.")
 
     async def handle_message(self, msg: InternalMessage):
         """
@@ -80,10 +84,10 @@ class BDICoreAgent(BaseAgent):
                 self.logger.debug("Processing message from belief collector.")
                 try:
                     if msg.thread == "beliefs":
-                        beliefs = json.loads(msg.body)
+                        beliefs = BeliefMessage.model_validate_json(msg.body).beliefs
                         self._add_beliefs(beliefs)
-                except Exception as e:
-                    self.logger.error(f"Error processing belief: {e}")
+                except ValidationError:
+                    self.logger.exception("Error processing belief.")
             case settings.agent_settings.llm_name:
                 content = msg.body
                 self.logger.info("Received LLM response: %s", content)

src/control_backend/agents/bdi/bdi_core_agent/rules.asl

@@ -1,3 +1,3 @@
-+user_said(NewMessage) <-
-  -user_said(NewMessage);
-  .reply(NewMessage).
++user_said(Message) <-
+  -user_said(Message);
+  .reply(Message).

src/control_backend/agents/bdi/belief_collector_agent.py

@@ -3,6 +3,7 @@ import json
 from control_backend.agents.base import BaseAgent
 from control_backend.core.agent_system import InternalMessage
 from control_backend.core.config import settings
+from control_backend.schemas.belief_message import BeliefMessage
 
 
 class BDIBeliefCollectorAgent(BaseAgent):
@@ -80,7 +81,7 @@ class BDIBeliefCollectorAgent(BaseAgent):
         msg = InternalMessage(
             to=settings.agent_settings.bdi_core_name,
             sender=self.name,
-            body=json.dumps(beliefs),
+            body=BeliefMessage(beliefs=beliefs).model_dump_json(),
             thread="beliefs",
         )