feat: add cache_salt dimension to LLM response cache

tests/unit/test_governed_llm.py

@@ -23,8 +23,13 @@ class _FakeRedisCache:
     def __init__(self) -> None:
         self._store: dict[str, LLMResponse] = {}
 
-    async def get(self, model: str, messages: list[dict[str, str]]) -> LLMResponse | None:
-        key = f"{model}:{json.dumps(messages, sort_keys=True)}"
+    async def get(
+        self,
+        model: str,
+        messages: list[dict[str, str]],
+        cache_salt: str | None = None,
+    ) -> LLMResponse | None:
+        key = f"{model}:{cache_salt}:{json.dumps(messages, sort_keys=True)}"
         return self._store.get(key)
 
     async def set(
@@ -32,8 +37,9 @@ class _FakeRedisCache:
         model: str,
         messages: list[dict[str, str]],
         response: LLMResponse,
+        cache_salt: str | None = None,
     ) -> None:
-        key = f"{model}:{json.dumps(messages, sort_keys=True)}"
+        key = f"{model}:{cache_salt}:{json.dumps(messages, sort_keys=True)}"
         self._store[key] = response