litellm_fix_mapped_tests_core: fix test isolation and mock injection issues

[shin-bot-litellm]

Problem

Four tests in litellm_mapped_tests_core were failing:

1. test_register_model_with_scientific_notation - KeyError due to test isolation issues
2. test_search_uses_registry_credentials - Mock not being called due to incorrect patch path
3. test_send_email_missing_api_key - Real API calls despite mocking
4. test_stream_transformation_error_sync - Mock not effective, real API called

Solution

test_register_model_with_scientific_notation

• Use unique model name to avoid conflicts with other tests
• Clear LRU caches before test to prevent stale data
• Clean up model_cost entry after test

test_search_uses_registry_credentials

• Use patch.object() on the actual base_llm_http_handler instance
• String-based patching for instance methods can fail; direct object patching is more reliable

test_send_email_missing_api_key

• Directly inject mock HTTP client into logger instance
• This bypasses any caching issues that could cause the fixture mock to be ineffective

test_stream_transformation_error_sync

• Patch litellm.completion directly instead of the handler module litellm reference
• This ensures the mock is effective regardless of import order

Regression

These tests were affected by LRU caching added in #19606 and HTTP client caching.

Failing Tests Link

https://app.circleci.com/pipelines/github/BerriAI/litellm/56737/workflows/9aa8a62f-0d73-4f49-a084-aa5408142a69/jobs/1140414/tests

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
litellm	Building	Preview, Comment	Feb 1, 2026 0:30am

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[shin-bot-litellm]

Closing in favor of PR from properly named branch to trigger CircleCI tests.

[greptile-apps]

Greptile Overview

Greptile Summary

Fixes test isolation and mock injection issues in four failing tests caused by LRU caching and HTTP client caching introduced in #19606.

Key Changes:

• test_register_model_with_scientific_notation: Uses unique model name and properly clears LRU caches before/after test to prevent conflicts
• test_search_uses_registry_credentials: Switches from string-based patching to patch.object() on the actual instance for more reliable mocking
• test_send_email_missing_api_key: Directly injects mock HTTP client into logger instance to bypass caching issues
• test_stream_transformation_error_sync: Simplifies mock path from module-level to direct litellm.completion patching

All changes are test-only with no production code impact. The fixes properly address cache-related test isolation issues.

Confidence Score: 5/5

• This PR is safe to merge - test-only changes with proper isolation fixes
• All changes are limited to test files, addressing specific test isolation and mocking issues. The fixes use established patterns (direct mock injection, unique test names, cache clearing). Only minor style improvements suggested.
• No files require special attention

Important Files Changed

Filename	Overview
tests/test_litellm/enterprise/enterprise_callbacks/send_emails/test_resend_email.py	Improves mock injection by directly assigning mock client to logger instance, preventing cache-related issues
tests/test_litellm/google_genai/test_google_genai_handler.py	Simplifies patch path from module-level to direct litellm.completion patching
tests/test_litellm/test_utils.py	Adds proper test isolation with unique model name, cache clearing, and cleanup
tests/test_litellm/vector_stores/test_vector_store_registry.py	Fixes instance method patching by using patch.object() on imported handler instance

Sequence Diagram

sequenceDiagram
    participant Test as Test Case
    participant Cache as LRU Cache/Client Cache
    participant Mock as Mock Objects
    participant SUT as System Under Test
    
    Note over Test,SUT: Before Fix: Cache Issues
    Test->>Cache: Initialize (cached data exists)
    Test->>Mock: Setup mock (ineffective)
    Cache-->>SUT: Returns cached real client
    SUT->>SUT: Uses real client (test fails)
    
    Note over Test,SUT: After Fix: Proper Isolation
    Test->>Cache: Clear caches (_invalidate_model_cost_lowercase_map)
    Test->>Test: Use unique identifiers (test-model-unique-12345)
    Test->>Mock: Create mock client
    Test->>SUT: Direct injection (logger.async_httpx_client = mock)
    Test->>Mock: Setup patch.object(instance, method)
    SUT->>Mock: Uses injected mock (test passes)
    Test->>Cache: Cleanup (delete entries, clear caches)

Loading

[greptile-apps]

_{4 files reviewed, 2 comments}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

get_model_info and _cached_get_model_info_helper are imported but never used

Suggested change

	from litellm.utils import get_model_info, _cached_get_model_info_helper, _invalidate_model_cost_lowercase_map
	from litellm.utils import _invalidate_model_cost_lowercase_map

Prompt To Fix With AI

This is a comment left during a code review.
Path: tests/test_litellm/test_utils.py
Line: 2293:2293

Comment:
`get_model_info` and `_cached_get_model_info_helper` are imported but never used

```suggestion
    from litellm.utils import _invalidate_model_cost_lowercase_map
```

How can I resolve this? If you propose a fix, please make it concise.

[greptile-apps]

inconsistent with the async test approach at line 215 which patches litellm.google_genai.adapters.handler.litellm. Consider using the same direct patching approach for consistency

Suggested change

	# Patch litellm.completion directly to prevent real API calls
	with patch("litellm.completion", return_value=mock_stream):
	with patch("litellm.google_genai.adapters.handler.litellm") as mock_litellm:
	mock_litellm.completion = MagicMock(return_value=mock_stream)

_{Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!}

Prompt To Fix With AI

This is a comment left during a code review.
Path: tests/test_litellm/google_genai/test_google_genai_handler.py
Line: 186:187

Comment:
inconsistent with the async test approach at line 215 which patches `litellm.google_genai.adapters.handler.litellm`. Consider using the same direct patching approach for consistency

```suggestion
        with patch("litellm.google_genai.adapters.handler.litellm") as mock_litellm:
            mock_litellm.completion = MagicMock(return_value=mock_stream)
```

<sub>Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!</sub>

How can I resolve this? If you propose a fix, please make it concise.