feat(backend): full remote implementation, untested

backend/src/inference.py

@@ -15,19 +15,22 @@ CONTEXT_WINDOW = 10_000
 
 # TODO: Rename to ChatMessage. Add methods for constructing from instruct/openai/etc formatted objects.
 class LlamaMessage:
-    def __init__(self, role, content):
+    def __init__(self, role, content, reasoning=None):
         if role not in VALID_ROLES:
             raise Exception("message role must be one of: " + ",".join(VALID_ROLES))
         self.role = role
         self.content = content
+        self.reasoning = reasoning
 
     def to_dict(self):
         """Serialize to the dict format expected by llama-cpp / ollama."""
         d = {"role": self.role, "content": self.content}
+        if self.reasoning:
+            d["reasoning"] = self.reasoning
         return d
 
     def __repr__(self):
-        return f"LlamaMessage(role={self.role!r}, content={self.content!r})"
+        return f"LlamaMessage(role={self.role!r}, content={self.content!r}, reasoning={self.reasoning!r})"
 
 class Models:
     llama_8b = "meta/meta-llama-3-8b-instruct"