litellm_fix_mapped_tests_core: clear client cache and fix isinstance checks

## Problem
Tests using mocked HTTP clients were hitting real APIs because:
1. HTTP client cache was returning previously cached real clients
2. isinstance checks failed due to module identity issues from sys.path

### Tests affected:
- test_send_email_missing_api_key
- test_send_email_multiple_recipients (resend & sendgrid)
- test_search_uses_registry_credentials
- test_vector_store_create_with_simple_provider_name
- test_vector_store_create_with_provider_api_type
- test_vector_store_create_with_ragflow_provider
- test_image_edit_merges_headers_and_extra_headers
- test_retrieve_container_basic (container API tests)

## Solution
1. Add clear_client_cache fixture (autouse=True) to clear
   litellm.in_memory_llm_clients_cache before each test
2. Fix isinstance checks to use type name comparison
   (avoids module identity issues from sys.path.insert)

## Why not disable_aiohttp_transport
The default transport is aiohttp, so tests should work with it.
Clearing the cache ensures mocks are used instead of cached real clients.

## Regression
PR #19829 (commit f95572e3ed) added @respx.mock but cached clients
from earlier tests were being reused, bypassing the mocks.

tests/test_litellm/containers/test_container_api.py

@@ -32,6 +32,21 @@ from litellm.types.containers.main import (
 )
 
 
+@pytest.fixture(autouse=True)
+def clear_client_cache():
+    """
+    Clear the HTTP client cache before each test to ensure mocks are used.
+    This prevents cached real clients from being reused across tests.
+    """
+    cache = getattr(litellm, "in_memory_llm_clients_cache", None)
+    if cache is not None:
+        cache.flush_cache()
+    yield
+    # Clear again after test to avoid polluting other tests
+    if cache is not None:
+        cache.flush_cache()
+
+
 class TestContainerAPI:
     """Test suite for container API functionality."""
 

tests/test_litellm/enterprise/enterprise_callbacks/send_emails/test_resend_email.py

@@ -9,6 +9,7 @@ from httpx import Response
 
 sys.path.insert(0, os.path.abspath("../../.."))
 
+import litellm
 from litellm_enterprise.enterprise_callbacks.send_emails.resend_email import (
     ResendEmailLogger,
 )
@@ -16,6 +17,21 @@ from litellm_enterprise.enterprise_callbacks.send_emails.resend_email import (
 # Test file for Resend email integration
 
 
+@pytest.fixture(autouse=True)
+def clear_client_cache():
+    """
+    Clear the HTTP client cache before each test to ensure mocks are used.
+    This prevents cached real clients from being reused across tests.
+    """
+    cache = getattr(litellm, "in_memory_llm_clients_cache", None)
+    if cache is not None:
+        cache.flush_cache()
+    yield
+    # Clear again after test to avoid polluting other tests
+    if cache is not None:
+        cache.flush_cache()
+
+
 @pytest.fixture
 def mock_env_vars():
     with mock.patch.dict(os.environ, {"RESEND_API_KEY": "test_api_key"}):

tests/test_litellm/enterprise/enterprise_callbacks/send_emails/test_sendgrid_email.py

@@ -9,11 +9,26 @@ from httpx import Response
 
 sys.path.insert(0, os.path.abspath("../../.."))
 
+import litellm
 from litellm_enterprise.enterprise_callbacks.send_emails.sendgrid_email import (
     SendGridEmailLogger,
 )
 
 
+@pytest.fixture(autouse=True)
+def clear_client_cache():
+    """
+    Clear the HTTP client cache before each test to ensure mocks are used.
+    This prevents cached real clients from being reused across tests.
+    """
+    cache = getattr(litellm, "in_memory_llm_clients_cache", None)
+    if cache is not None:
+        cache.flush_cache()
+    yield
+    if cache is not None:
+        cache.flush_cache()
+
+
 @pytest.fixture
 def mock_env_vars():
     # Store original values

tests/test_litellm/test_main.py

@@ -18,6 +18,20 @@ import litellm
 from litellm import main as litellm_main
 
 
+@pytest.fixture(autouse=True)
+def clear_client_cache():
+    """
+    Clear the HTTP client cache before each test to ensure mocks are used.
+    This prevents cached real clients from being reused across tests.
+    """
+    cache = getattr(litellm, "in_memory_llm_clients_cache", None)
+    if cache is not None:
+        cache.flush_cache()
+    yield
+    if cache is not None:
+        cache.flush_cache()
+
+
 @pytest.fixture(autouse=True)
 def add_api_keys_to_env(monkeypatch):
     monkeypatch.setenv("ANTHROPIC_API_KEY", "sk-ant-api03-1234567890")

tests/test_litellm/vector_stores/test_vector_store_create_provider_logic.py

@@ -51,9 +51,10 @@ def test_vector_store_create_with_simple_provider_name():
     )
     
     assert vector_store_provider_config is not None, "Should return a config for OpenAI"
-    assert isinstance(
-        vector_store_provider_config, OpenAIVectorStoreConfig
-    ), "Should return OpenAIVectorStoreConfig for OpenAI provider"
+    # Use type name check instead of isinstance to avoid module identity issues
+    # caused by sys.path manipulation in test setup
+    assert type(vector_store_provider_config).__name__ == "OpenAIVectorStoreConfig", \
+        f"Should return OpenAIVectorStoreConfig for OpenAI provider, got {type(vector_store_provider_config).__name__}"
     
     print("✅ Test passed: Simple provider name 'openai' handled correctly")
 
@@ -97,9 +98,9 @@ def test_vector_store_create_with_provider_api_type():
     )
     
     assert vector_store_provider_config is not None, "Should return a config for Vertex AI"
-    assert isinstance(
-        vector_store_provider_config, VertexVectorStoreConfig
-    ), "Should return VertexVectorStoreConfig for vertex_ai provider with rag_api"
+    # Use type name check instead of isinstance to avoid module identity issues
+    assert type(vector_store_provider_config).__name__ == "VertexVectorStoreConfig", \
+        f"Should return VertexVectorStoreConfig for vertex_ai provider with rag_api, got {type(vector_store_provider_config).__name__}"
     
     print("✅ Test passed: Provider with api_type 'vertex_ai/rag_api' handled correctly")
 
@@ -134,9 +135,9 @@ def test_vector_store_create_with_ragflow_provider():
     )
     
     assert vector_store_provider_config is not None, "Should return a config for RAGFlow"
-    assert isinstance(
-        vector_store_provider_config, RAGFlowVectorStoreConfig
-    ), "Should return RAGFlowVectorStoreConfig for RAGFlow provider"
+    # Use type name check instead of isinstance to avoid module identity issues
+    assert type(vector_store_provider_config).__name__ == "RAGFlowVectorStoreConfig", \
+        f"Should return RAGFlowVectorStoreConfig for RAGFlow provider, got {type(vector_store_provider_config).__name__}"
     
     print("✅ Test passed: RAGFlow provider handled correctly")
 

tests/test_litellm/vector_stores/test_vector_store_registry.py

@@ -21,6 +21,20 @@ from litellm.vector_stores.main import search
 from litellm.vector_stores.vector_store_registry import VectorStoreRegistry
 
 
+@pytest.fixture(autouse=True)
+def clear_client_cache():
+    """
+    Clear the HTTP client cache before each test to ensure mocks are used.
+    This prevents cached real clients from being reused across tests.
+    """
+    cache = getattr(litellm, "in_memory_llm_clients_cache", None)
+    if cache is not None:
+        cache.flush_cache()
+    yield
+    if cache is not None:
+        cache.flush_cache()
+
+
 def test_get_credentials_for_vector_store():
     """Test that get_credentials_for_vector_store returns correct credentials"""
     # Create test vector stores