API Reference

Core Methods

Complete API documentation for all ragleaf methods and endpoints.

RAGQualityChecker

Main class for quality checking and verification.

from ragleaf import RAGQualityChecker

checker = RAGQualityChecker(
    api_key: str,
    threshold: float = 0.8,
    enable_visualization: bool = False
)

verify_citations()

Verify that citations in your RAG response match the source documents.

result = checker.verify_citations(
    query: str,
    response: str,
    sources: List[Document]
) -> CitationResult

Parameters:

  • query - The user's original query
  • response - The generated RAG response
  • sources - List of source documents used

Returns: CitationResult object with accuracy score and details

visualize_retrieval()

Generate visualization of your retrieval pipeline performance.

viz = checker.visualize_retrieval(
    query: str,
    retrieved_docs: List[Document],
    format: str = "html"
) -> Visualization

Parameters:

  • query - The search query
  • retrieved_docs - Documents retrieved by your system
  • format - Output format: "html", "json", or "png"

optimize_chunks()

Get recommendations for optimal chunk size and overlap settings.

recommendations = checker.optimize_chunks(
    documents: List[Document],
    current_settings: ChunkSettings
) -> OptimizationResult

Parameters:

  • documents - Your document collection
  • current_settings - Current chunking configuration

analyze_prompt()

Analyze and get suggestions for prompt optimization.

analysis = checker.analyze_prompt(
    prompt: str,
    context: str
) -> PromptAnalysis

REST API Endpoints

All methods are also available via REST API:

POST https://api.ragleaf.dev/v1/verify-citations
POST https://api.ragleaf.dev/v1/visualize-retrieval
POST https://api.ragleaf.dev/v1/optimize-chunks
POST https://api.ragleaf.dev/v1/analyze-prompt

Authentication

Include your API key in the Authorization header:

Authorization: Bearer your_api_key