- Add get_llm_provider to lazy import registry system
- Remove eager import from __init__.py
- Fix circular import by updating internal modules to import directly from source:
- realtime_api/main.py
- router_utils/pattern_match_deployments.py
- Add type stub for get_llm_provider in TYPE_CHECKING block
- Add tests for lazy loading of LLM provider logic functions
This reduces initial import time and memory usage by deferring the import
of get_llm_provider until it's actually accessed.
- /v2/model/info now returns {"data": []} when llm_router is None or model_list is empty
- /model_group/info now returns {"data": []} when llm_model_list is None or empty
- Fixes UI crash on fresh installs with STORE_MODEL_IN_DB=True
- Added 4 unit tests for empty model list scenarios
Adds log_format parameter supporting json_array (default), ndjson, and single formats. NDJSON format enables webhook integrations like Sumo Logic to parse individual log records at ingest time. Defaults to json_array for backward compatibility.
* fix(cost_calculator): correct gpt-image-1 cost calculation using token-based pricing (#13847)
gpt-image-1 uses token-based pricing (like chat models), not pixel-based pricing
like DALL-E. The old code was calculating incorrect costs by treating it as DALL-E.
Changes:
- Update model pricing JSON with correct token-based costs for gpt-image-1
- Add dedicated cost calculator for OpenAI gpt-image models
- Route gpt-image-1 to token-based calculator in cost router
- Add comprehensive tests for the new calculator
* refactor: simplify gpt-image-1 cost calculator using responses API helper
Reuse _transform_response_api_usage_to_chat_usage and generic_cost_per_token
for gpt-image-1 cost calculation since ImageUsage has the same spec as
ResponseAPIUsage.
* Allow get_nested_value dot notation to support escaping for Kubernetes JWT Support
* Add support for team and org alias fields, add docs, tests
* Fix lint issue with max statements in handle jwt logic
Fix Vertex AI API error: "tools[0].tool_type: one_of 'tool_type' has more
than one initialized field"
The Vertex AI API requires each Tool object to contain exactly one type
of tool (e.g., FunctionDeclaration, GoogleSearch, CodeExecution).
Previously, all tool types were combined into a single Tool object,
causing INVALID_ARGUMENT errors when using multiple tools simultaneously.
This change creates separate Tool objects for each tool type:
- Function declarations in one Tool
- Google Search in its own Tool
- Code Execution in its own Tool
- etc.
Ref: https://cloud.google.com/vertex-ai/generative-ai/docs/reference/rest/v1beta1/Tool🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Fixes#18430
- Pass custom_llm_provider to anthropic_messages_pt instead of hardcoded 'anthropic'
- Add check for vertex_ai provider to force base64 conversion for image URLs
- Add tests to verify behavior for both Vertex AI and regular Anthropic
