fix(litellm): patch async embeddings for auto instrumentation (#186)

Trace litellm.aembedding alongside the existing embedding path so patch_litellm() and wrap_litellm() cover async embedding calls.

Add a VCR regression test and cassette to lock down the missing async embedding instrumentation reported in issue 115.

Author: AbhiPrasad

py/src/braintrust/integrations/litellm/__init__.py

@@ -8,8 +8,9 @@ def patch_litellm() -> bool:
     """Patch LiteLLM to add Braintrust tracing.
 
     This wraps litellm.completion, litellm.acompletion, litellm.responses,
-    litellm.aresponses, litellm.embedding, and litellm.moderation to
-    automatically create Braintrust spans with detailed token metrics,
+    litellm.aresponses, litellm.embedding, litellm.aembedding, and
+    litellm.moderation to automatically create Braintrust spans with
+    detailed token metrics,
     timing, and costs.
 
     Returns: