Split data into batches. Args: data: List of data items batch_size: Size of each batch Returns: Lis

Synchronous wrapper for async_batch_request_with_messages. This function ensures all cleanup tasks complete before the event loop closes

Build a structured system prompt from knowledge base entries. Args: knowledge_entries: List of knowledge base en

Get a subclass by its registered name, or its fully qualified class name. Parameters ---------- name : ``str``

Convert openai chat-completions-style messages to a readable string.

description: Copy a file to another location. model: File --- success: { "message": "string", "destination_file_p

description: Delete a directory with its sub-directories and files. model: Directory --- {"message": "string"}

Evaluate a batch of question-answer pairs. Args: questions: List of questions answers1: List of firs

Compare predicted SQL against gold standard by execution results Args: db_id: Target database identifier

Execute SQL query and return results

Extract database ID from question. Args: question: String in format "dbid|question" Returns:

Extract SQL query from text response

Returns a copy API Docs in OpenAI's function calling format.

Generate a response from the LLM using tuning model. Args: prompt: User prompt system_prompt: Option

Get a summary of the current knowledge base.

The answer will be considered correct if: (a) it normalizes to the same string as the ground truth answer OR (b) sympy can simplify t

Check if the file is a Git LFS pointer.

Loads data from a list of file paths.

Returns a copy API Docs in MCP format.