Text extraction (Special math elements)

[Prasaderp]

Hi,
I was working around extracting text from pdf's. PyMuPDF is very good at extracting complex elements related to math. While I was working around it, I found out that it is not able to extract elements like fractions from the source PDF.

I have attached 2 screenshots of the input PDF and extracted text in a PPT using python PPTX.

If you could let me know, is it the limitation of the PyMupDF or python-pptx for this rendering issue

[JorjMcKie]

Mathematical formulas usually are a conglomerate of text (written in a mixture of standard and specialized math fonts) plus vector graphics.
Fraction lines, extended length radical lines and other elements are represented as vector graphics - which hence do not appear as part of the text extraction. You can for sure extract them (page.get_drawings()), but this will be a separate output, so you will need to fiddle around for associating a horizontal line with the right text portion.

[Prasaderp]

Thanks for the Information!

[JorjMcKie]

For example if you look at this formula (defined in Word):

and convert this to PDF using LibreOffice, then you will get these text and vector graphics elements:

The colors indicate:

• red: standard font
• blue: a math font
• green: vector graphics

When using Word for the PDF conversion, the analysis is this:

IAW Word conversion exclusively uses their special math font for the text part in formulas.