plexus.reports.blocks.guideline_vetting module

Shared guideline-vetting voting engine for feedback contradiction/alignment analysis.

class plexus.reports.blocks.guideline_vetting.GuidelineVettingService(invoke_bedrock: Callable[[str, bool], Dict[str, Any]] | None = None, invoke_openai: Callable[[str, str], Dict[str, Any]] | None = None)

Bases: object

Reusable per-item voting service used by contradiction and aligned report modes.

__init__(invoke_bedrock: Callable[[str, bool], Dict[str, Any]] | None = None, invoke_openai: Callable[[str, str], Dict[str, Any]] | None = None)

async analyze_items(items: List[Any], guidelines: str | None, max_concurrent: int, score_results_by_item: Dict[str, Any], rubric_memory_contexts_by_item: Dict[str, Dict[str, Any]] | None = None) → List[Dict[str, Any]]

Analyze items via shared voting and return both contradiction and aligned results.

Each returned item includes per-vote traces, confidence, verdict, and strict associated-dataset eligibility (unanimous non-contradiction only).