ykachala/evals

Prompt and LLM-output regression testing for PHP. Define golden datasets, assert on model output (schema, semantic similarity, LLM-as-judge), and gate CI on quality scores.

dev-main 2026-06-02 08:53 UTC

Requires

• php: >=8.3

Requires (Dev)

• friendsofphp/php-cs-fixer: ^3.0
• phpstan/phpstan: ^2.0
• phpunit/phpunit: ^11.0

Suggests

• ykachala/meter: Track the token cost of an eval run
• ykachala/semantic-cache: Cache judge/embedding calls so eval runs in CI stay cheap

MIT

• testing
• phpunit
• ci
• quality
• regression
• ai
• pest
• llm
• prompt-engineering
• evals
• llm-as-judge