Get page number and bounding box when dealing with docx

[Amin5434]

Hi everyone,

I built a chunking pipeline using Docling HybridChunker, and I need the page number and bounding box informations for my chunks. Unfortunately, I cant' get to collect these for .DOCX.

Do you have a solution ? Or is it not possible for this document type ?

Thank you for your consideration

[shubham-wysa]

please answer 🙏 @vagenas @cau-git

[simonschoe]

@shubham-wysa You will not be able obtain bounding box and page information for .docx files, since internally .docx files do not track these information, they are simply a tree of text elements. Page information result from rendering .docx files through a viewer (e.g., MS Word). If you require these information, you should convert to pdf before ingestion.