Module span 

Span extraction from documents

This module handles extracting meaningful text spans from documents. Spans are the fundamental unit of retrieval in AvocadoDB.

Key Principles

• Spans should be 20-50 lines (roughly 200-500 tokens)
• Respect natural boundaries (paragraphs, code blocks, sections)
• Never split mid-sentence
• Maintain line number accuracy for citations

Functions

extract_spans

Extract spans from document text