too much

living_agents/character_agent.py

@@ -71,6 +71,33 @@ class CharacterAgent:
 
         return response
 
+    async def chat_with(self, messages: List[LLMMessage]) -> LLMMessage:
+        """Generate reaction based on memory and character"""
+        # Retrieve relevant memories
+        relevant_memories = await self.memory_stream.retrieve_related_memories(situation, k=8)
+        memory_context = "\n".join([f"- {m.description}" for m in relevant_memories])
+
+        context = {
+            'character': self._get_character_prompt(),
+            'character_name': self.character.name,
+            'memory_context': memory_context,
+            'situation': situation}
+        prompt = PromptManager.get_prompt('react_to_situation', context)
+
+        response = await self.llm.chat([{"role": "user", "content": prompt}])
+
+        # create new memories from interaction
+        interaction_context = {
+            'situation': f'I reacted to: \n{situation}',
+            'response': f'My response was: \n{response}',
+        }
+        prompt, schema = PromptManager.get_prompt_with_schema('extract_interaction_memories', interaction_context)
+        memories_response = await self.llm.client.get_structured_response([{"role": "user", "content": prompt}], schema)
+        for new_memory in memories_response['memories']:
+            await self.perceive(new_memory)
+
+        return response
+
     async def plan_day(self) -> List[str]:
         """Generate high-level daily plan"""
         # Retrieve relevant memories about goals, habits, schedule
@@ -255,5 +282,7 @@ Summary:"""
             memory.importance_score = await instance._score_memory_importance(memory)
             await instance._analyze_trait_impact(memory)
 
+        instance.character.template_file = template['yaml_file']
+
         logger.info(f"Character {instance.character.name} created successfully")
         return instance

living_agents/datatypes.py

@@ -2,7 +2,7 @@ from dataclasses import dataclass, field
 from typing import Dict, List, Optional, Literal, TypedDict
 from datetime import datetime
 from uuid import uuid4
-import random
+from pathlib import Path
 
 
 class CharacterTemplate(TypedDict):
@@ -10,7 +10,7 @@ class CharacterTemplate(TypedDict):
     observations: List[str]
     reflections: List[str]
     plans: List[str]
-    yaml_file: str
+    yaml_file: Path
 
 
 @dataclass
@@ -72,6 +72,7 @@ class Character:
     traits: List[CharacterTrait] = field(default_factory=list)
     relationships: Dict[str, str] = field(default_factory=dict)
     goals: List[str] = field(default_factory=list)
+    template_file: Optional[Path] = None
     _id: str = field(default_factory=lambda: str(uuid4())[:8])
 
     def get_trait(self, trait_name, trait_description) -> CharacterTrait:

living_agents/memory_stream.py

@@ -41,7 +41,6 @@ class MemoryStream:
 
         # Track for reflection trigger
         self.recent_importance_sum += memory.importance_score
-        print(f"Recent Importance Sum: {self.recent_importance_sum}")
 
         # Trigger reflection if threshold exceeded
         if self.recent_importance_sum >= self.importance_threshold:

living_agents/prompts/chat_with_agent.md

@@ -0,0 +1,15 @@
+{{character}}
+
+Relevant memories from your past:
+{{memory_context}}
+
+Respond as {{character_name}} describing what you did and how you reacted. Write in first person past tense as if this
+just happened to you.
+
+Examples of the response style:
+
+- "I looked up from my book and smiled nervously..."
+- "I felt my heart race and took a deep breath before I said..."
+- "I hesitated for a moment, then decided to..."
+
+Stay completely in character and be specific about your actions, thoughts, and words.