/claim #30

This PR introduces hierarchical (parent-child) chunking and integrates it with the agentic RAG pipeline for advanced semantic document processing.

Key changes:

Added a hierarchical chunking strategy that organizes document content into a multi-level structure: Page → Section → Semantic Chunk. Integrated hierarchical chunking with the agentic RAG pipeline, enabling context-aware retrieval and processing. Exposed the new functionality via FastAPI endpoints for easy API access and testing. Updated documentation and usage instructions to reflect the new features. Benefits:

Enables more accurate, context-rich retrieval for LLMs and agentic workflows. Supports complex document structures (e.g., financial reports, research papers). Makes it easier to test and extend hierarchical chunking via the API. Let me know if you want this tailored further or need a more technical breakdown!