feat: enhance token usage tracking and context management for LLM providers

src/providers/base.py

@@ -134,6 +134,20 @@ class BaseProvider(abc.ABC):
         """
         pass
 
+    @abc.abstractmethod
+    def get_usage(self, response: Any) -> dict[str, int] | None:
+        """
+        Extracts token usage information from a non-streaming response object.
+
+        Args:
+            response: The non-streaming response object.
+
+        Returns:
+            A dictionary containing 'prompt_tokens' and 'completion_tokens',
+            or None if usage information is not available.
+        """
+        pass
+
     # Optional: Add a method for follow-up completions if the provider API
     # requires a specific structure different from just appending messages.
     # def create_follow_up_completion(...) -> Any: