Initialize the FanoutExporter with a list of exporters. :param exporters: List of TracingExporter instances.

Initialize the OtelExporter with the OpenTelemetry endpoint. :param endpoint: The URL of the OpenTelemetry endpoint.

Initialize ToolManager. :param server_configs: List returned by create_server_parameters()

Initialize cache manager Args: cache_dir: Directory to store all cache files

Initialize file cache Args: cache_dir: Directory to store cache files cache_name: Name of the cache

Initialize the pointwise evaluator

Initialize the expert generator. Args: api_client: API client instance cache_file: Cache file path for check

Initialize the domain loader. Args: domains_file: Path to the domain list JSON file

Initialize the task generator. Args: api_client: API client instance cache_file: Cache file path

Initialize the filter. Args: api_client: API client instance cache_file: Cache file path max_wor

Initialize the quality assessor. Args: api_client: API client instance baseline_cache_file: Cache file path

Initialize the API client. Args: model_name: Name of the model to use api_type: API type, supports "openai"

Initialize the pipeline. Args: config: Task generation configuration

Assess answer quality for a single task. Args: task: Task with baseline answer Returns: Task that p

Create configured OpenAI client

Create Anthropic client

Create configured OpenAI client

Create configured OpenAI client

Create configured OpenAI client

Create configured OpenAI client

Send message to OpenAI API. :param system_prompt: System prompt string. :param messages: Message history list. :retur

Send message to Anthropic API. :param system_prompt: System prompt string. :param messages: Message history list. :re

Send message to OpenAI API. :param system_prompt: System prompt string. :param messages: Message history list. :retur

Send message to OpenAI API. :param system_prompt: System prompt string. :param messages: Message history list. :retur

Send message to OpenAI Response API. :param system_prompt: System prompt string. :param messages: Message history list.

Send message to OpenAI API. :param system_prompt: System prompt string. :param messages: Message history list. :retur

Send message to OpenAI API. :param system_prompt: System prompt string. :param messages: Message history list. :retur

Filter a single task. Args: task: Task info Returns: Task that passed filter (with filter_result),

Find servers containing the specified tool name among all servers :param tool_name: Tool name to search for :return: List of

Format response for logging

Generate baseline answer for a single task. Args: task: Task info Returns: Task with baseline_answe

Generate experts for a single domain. Args: domain: Domain name Returns: List of experts for this d

Generate tasks for a single expert. Args: expert: Expert info dictionary Returns: List of tasks for

Install common Python packages in the sandbox. Args: sandbox: The connected sandbox instance sandbox_id: The sandbox ID for error

Score a single dimension of a report with retry mechanism

Update cumulative token usage - implemented by specific implementation class

Answer the question based on the given audio information. Args: audio_path_or_url: The path of the audio file locally or its URL. Pa

Transcribe audio file to text and return the transcription. Args: audio_path_or_url: The path of the audio file locally or its URL. P

Call Gemini vision API.

Call OpenAI vision API.

Clear all caches

Clear all caches

Clear a specific cache Args: cache_name: Name of the cache to clear

Create a linux sandbox and get the `sandbox_id` for safely executing commands and running python code. Note that the `sandbox_id` can only be assigned

Create a new span.

Create a new trace.

Download a file from the internet to the `/home/user` dir of the sandbox. You should use this tool to download files from the internet. Args:

Download a file from the sandbox to local system. Files in sandbox cannot be processed by tools from other servers - only local files and internet URL

End the current sub-agent session

Check if current message_history + summary_prompt would exceed context If it would exceed, remove last assistant-user pair and return

Evaluate all queries (to be implemented by subclasses) Args: *args: Additional positional arguments