v1.1.0

Changed

• All .prompt() / .get_trustworthiness_score() / .generate() / .score() methods will now catch any errors and return null values alongside a log of why the exception occurred
• try_ methods are deprecated and will share the same functionality as the "non-try" methods

v1.0.22

• Update response_helpfulness default criteria

v1.0.21

Added

• Add new OpenAI models: gpt-4.1, gpt-4.1-mini, gpt-4.1-nano, o3, o4-mini

v1.0.20

Added

• Add new options for similarity_metric: code and discrepancy

Fixed

• Better validation / error message for invalid evals arguments in TrustworthyRAG

v1.0.19

Added

• Add batch processing message

v1.0.18

Added

• Add score_async method to TrustworthyRAG

v1.0.17

Added

• TrustworthyRAG set retries

v1.0.16

Added

• TrustworthyRAG set retries

v1.0.15

Added

• Update context_sufficiency default criteria

v1.0.14

Added

• Add batch support to TrustworthyRAG