- Test multiple tool calls without explicit indices receive sequential indices
- Verify Delta class assigns indices 0, 1, 2... instead of defaulting all to 0
- Add comprehensive assertions for tool call details preservation
- Cover provider-agnostic streaming response scenarios
- Test external ID parameter propagation through authentication chain
- Cover both standard Bedrock and Converse API authentication flows
- Verify assume_role STS calls include ExternalId when provided
- Ensure backward compatibility when external ID not specified
- Add specific test for BedrockConverseLLM parameter extraction
- Extend existing dynamic parameter tests to include aws_external_id
- Test ARN format inference profile model detection
- Test model_id parameter filtering to prevent API errors
- Test cross-region inference profile format support
- Ensure backward compatibility with regular Nova Canvas models
- Cover both request body generation and parameter transformation
- Keep CompactifAI provider detection logic
- Include new OVHCloud provider from main branch
- Both providers now work correctly with model prefix detection
- Update all model references from llama-2-7b-compressed to cai-llama-3-1-8b-slim
- Move CompactifAI tests from tests/llm_translation to tests/test_litellm/llms/compactifai/
- Update documentation examples to use the new model name
- Remove integration test inheritance to make tests pure mock tests
This addresses review feedback to use mock tests and updated model naming.
- Test basic and streaming completions with proper mocking
- Cover authentication, parameter handling, and error scenarios
- Test provider detection and async functionality
- Verify request headers and response transformation
- Follow LiteLLM testing patterns with respx/httpx mocking
- Ensure full compatibility with OpenAI-style responses
* Use _PROXY_MaxParallelRequestsHandler_v3 by default (#14352)
(cherry picked from commit f3fa45cf8fbd5f5cce2f45a7312776d5005fb08e)
(cherry picked from commit 5b680bb4a3)
* Use random api_key for parallel requests test
* Fix off-by-one error in parallel request rate limit
The rate limiter was incorrectly rejecting requests when the limit was met, but not exceeded. The check in `is_cache_list_over_limit` was `int(counter_value) + 1 > current_limit`, which caused the first request to be rejected if the limit was 1.
This commit removes the `+ 1`, changing the logic to `int(counter_value) > current_limit`. The check now correctly allows requests up to the specified parallel limit.
* Test actual parallel requests
* Ensure rate limiting works correctly for multiple users
* Add sequential rate-limit test
* Revert random key usage
