nat.plugins.rag.config

Configuration models and type aliases for NVIDIA RAG integration.

Classes

RAGPipelineConfig

Native nvidia_rag pipeline settings.

Module Contents

class RAGPipelineConfig(/, **data: Any)

Bases: pydantic.BaseModel

Native nvidia_rag pipeline settings.

Groups all RAG-specific settings that control search behavior, query preprocessing, and response quality.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

search_settings: nvidia_rag.utils.configuration.RetrieverConfig = None

ranking: nvidia_rag.utils.configuration.RankingConfig = None

query_rewriter: nvidia_rag.utils.configuration.QueryRewriterConfig | None = None

filter_generator: nvidia_rag.utils.configuration.FilterExpressionGeneratorConfig | None = None

query_decomposition: nvidia_rag.utils.configuration.QueryDecompositionConfig | None = None

reflection: nvidia_rag.utils.configuration.ReflectionConfig | None = None

vlm: nvidia_rag.utils.configuration.VLMConfig | None = None

enable_citations: bool = None

enable_guardrails: bool = None

enable_vlm_inference: bool = None

vlm_to_llm_fallback: bool = None

default_confidence_threshold: float = None