fix(caching): handle list-based responses, empty-list short-circuit, and message key variations in QdrantSemanticCache

Several issues were identified in the semantic caching logic:
1. When the cache attempted to store or retrieve list-based responses, a `TypeError` occurred during caching: `LiteLLM: ERROR: caching.py:647 - LiteLLM Cache: Exception add_cache: can only concatenate str (not "list") to str`.
2. The use of `str(value)` for serialization generated Python-specific representations (single quotes) incompatible with standard JSON parsing, forcing reliance on fragile `ast.literal_eval` fallbacks.
3. Message retrieval logic used `kwargs.get("messages") or kwargs.get("message")`, which incorrectly treated empty lists `[]` as falsy, causing the cache write/read to be silently skipped when valid empty conversations were provided.
4. Some unit tests contained unnecessary `mock_async_client` patches in synchronous test cases, creating boilerplate/unused mocks.

- Replaced `str(value)` with `json.dumps(value)` in `set_cache` and `async_set_cache` to ensure all cached values are stored as valid, standard JSON.
- Refactored message retrieval in all cache methods to explicitly check for key presence (`if "messages" in kwargs`) rather than relying on falsy evaluation, allowing `[]` to be treated as valid input.
- Modified input handling in cache methods to support both 'messages' and 'message' keys, ensuring robustness against varying input structures.
- Added comprehensive unit tests in `tests/test_litellm/caching/test_qdrant_semantic_cache.py` to validate list-type response caching and confirm the fix.
- Cleaned up test suite by removing unused `mock_async_client` patches in synchronous test cases.

litellm/caching/qdrant_semantic_cache.py

@@ -178,8 +178,8 @@ class QdrantSemanticCache(BaseCache):
         from litellm._uuid import uuid
 
         # get the prompt
-        messages = kwargs.get("messages") or kwargs.get("message")
-        if not messages:
+        messages = kwargs.get("messages") if "messages" in kwargs else kwargs.get("message")
+        if messages is None:
             print_verbose("No messages provided for semantic caching")
             return
         prompt = get_str_from_messages(messages)
@@ -223,8 +223,8 @@ class QdrantSemanticCache(BaseCache):
         print_verbose(f"sync qdrant semantic-cache get_cache, kwargs: {kwargs}")
 
         # get the messages
-        messages = kwargs.get("messages") or kwargs.get("message")
-        if not messages:
+        messages = kwargs.get("messages") if "messages" in kwargs else kwargs.get("message")
+        if messages is None:
             print_verbose("No messages provided for semantic lookup")
             return
         prompt = get_str_from_messages(messages)
@@ -295,8 +295,8 @@ class QdrantSemanticCache(BaseCache):
         print_verbose(f"async qdrant semantic-cache set_cache, kwargs: {kwargs}")
 
         # get the prompt
-        messages = kwargs.get("messages") or kwargs.get("message")
-        if not messages:
+        messages = kwargs.get("messages") if "messages" in kwargs else kwargs.get("message")
+        if messages is None:
             print_verbose("No messages provided for semantic caching")
             return
         prompt = get_str_from_messages(messages)
@@ -357,8 +357,8 @@ class QdrantSemanticCache(BaseCache):
         from litellm.proxy.proxy_server import llm_model_list, llm_router
 
         # get the messages
-        messages = kwargs.get("messages") or kwargs.get("message")
-        if not messages:
+        messages = kwargs.get("messages") if "messages" in kwargs else kwargs.get("message")
+        if messages is None:
             print_verbose("No messages provided for semantic lookup")
             return
         prompt = get_str_from_messages(messages)