latent.optimize.ace.data

Data processing utilities for ACE optimization.

Classes

DataProcessor

DataProcessor(field_mapping: dict[str, str] | None = None)

Converts raw dicts to QAPair objects with field mapping.

Args: field_mapping: Dict mapping {"input": source_field, "output": source_field}. Defaults to identity mapping (expects "input" and "output" keys).

QAPair

QAPair()

A single question-answer pair for ACE optimization.

Methods

DataProcessor.process

process(rows: list[dict[str, Any]]) -> list[QAPair]

Convert raw dicts to QAPair list.

Args: rows: Raw data rows.

Returns: List of QAPairs. The output key is always present (defaults to "").