Feature: Table-Specific Chunking Strategy

[mubashir-oss]

Summary

Implement a new chunking strategy that preserves tables as cohesive units during document processing.

Benefits

• Maintains semantic integrity of tabular data
• Enhances utility for data-heavy documents (e.g., financial reports, scientific papers)
• Prevents splitting of logically grouped table rows or columns

Implementation Notes

• Detect and isolate tables as distinct chunks
• Ensure table boundaries are preserved during chunking
• Optionally tag chunks as type: table for downstream processing

Impact

Improves accuracy and relevance of extracted content from structured documents, and enables better downstream use in RAG pipelines or data extraction workflows.