Extracting Full URLs from PDFs

[fabiocordeiro]

Hello everyone,

I’m currently using the Docling library in Python to extract text from PDF files. While it works well for retrieving visible text, I’ve noticed an issue with hyperlinks. Specifically, the library only extracts the short display text of the links, but it doesn’t capture the full URLs embedded in the PDF.

This is problematic because I need to extract both the visible text and the actual URLs. For example, if a link in the PDF displays as "Click here" but points to "https://example.com", I can only retrieve "Click here" using Docling.

I was wondering if there is a way to make Docling extract this information or if anyone has encountered a similar issue and found a workaround. If Docling doesn’t support this, are there any recommended libraries or techniques to extract full URLs from PDF files? I’ve read that tools like PyMuPDF or PyPDF2 might help, but I’m looking for the best way to integrate this functionality with my existing Docling-based workflow.

Any advice, examples, or pointers would be greatly appreciated!

Thank you in advance for help!

[gawaliakshata2-tech]

yes facing same issue for hyperlink extraction from pdf. Markdown export does not extract hidden URLs. Is there a way to extract the URLs and export them right after the link that appears in the text?

[Yelinz]

Same as #1953
Would be interested in a solution as well, due to having to extract the text with a seperate solution otherwise.

[jedg75]

Hello, I'm currently extracting hyperlinks separately using PyMuPDF/PyPDF2 after Docling conversion, then mapping them back to chunks by matching page numbers.

Since TextItem.hyperlink already exists in the data model but isn't populated during parsing, would it make sense to integrate link extraction into the PDF pipeline? I'm happy to work on this if it's a good fit for the project.