Incomplete AnalyzePage method in MuPDF.NET4LLM/helpers/Utils.cs

[rsegerink]

Hi,

I would like to use the AnalyzePage method from

MuPDF.NET/MuPDF.NET4LLM/helpers/Utils.cs

Line 561 in cdba4e0

public static Dictionary<string, object> AnalyzePage(Page page, List<Block> blocks = null)

just like we have this in the python version. But it is incomplete, especially the ocrSpans check, it is missing the CharFlags check because this property is not jet made available on the Span class, do you have a timeframe when this will be available?

thanks,

Rob

[JorjMcKie]

This check is not the most important one to answer "Need OCR?". For the time being, just check for font name "GlyphLessFont" and you are at least covered for the case "has been OCR-ed by Tesseract".
This should indeed be the most frequent reason for a better OCR version.

In addition, I have been working on a new version of that function with a significantly improved detection rate and at the same time a much better performance.
It will become part of PyMuPDF4LLM's next version.