Abstract method that should be implemented by subclasses to define the functionality of the tool.

Compare equality of metric values

Return state for pickling

True if self is a _better_ (not necessarily larger) metric value than other

Token counts for prompt, completion, reasoning, and cached. Reasoning tokens are included in completion tokens. Cached tokens

Simulates a standalone Python REPL with an execution time limit. Args: working_dir (Path | str): working directory of th

Return the number of nodes in the journal.

Set state during unpickling

Create detailed prompt to determine next stage configuration

Evaluate whether experiment is ready for next stage

Get structured response from LLM for stage configuration. Args: prompt: The analysis prompt to send to the LLM Returns:

Wrapper function that creates a fresh environment for each process

Save comprehensive stage completion summary

Absorb the result of executing the plotting code from this node.

Return a list of nodes that are considered buggy by the agent.

Create client for vision-language model.

Length of the current debug path - 0 if the node is not a debug node (parent is not buggy) - 1 if the parent is buggy but the

Return a list of nodes representing intial coding drafts

Extract all JSON objects from the text. Caveat: This function cannot handle nested JSON objects.

Generate a textual preview of a directory, including an overview of the directory structure and previews of individual files

Generate a trace of the interactive session in IPython format.

Get multiple responses from vision-language model for the same input. Args: msg: Text message to send image_paths: Path(s) to ima

Get the metric value for a specific dataset

Get the mean value across all datasets (or single value if not multi-dataset)

Return a list of all metric values in the journal.

Return a list of nodes that are not considered buggy by the agent.

Check if the node is a leaf node in the solution tree.

True if the metric value is the worst possible value.

Generate a report from a journal, the report will be in markdown format.

Reads a file that contains lines of the form: "Key": "Value", "Another Key": "Another Value", ... including possibl

Reset all token counts and interactions.

Save experimental notes and summaries to files

Return the stage of the node: - "stage" if the node is an initial solution draft - "debug" if the node is the result of a deb

Get the terminal output of the code execution (after truncating it).