Re: Problem with Search Results of PDFDocument-Class

[Claus Dietrich <claus.dietrich@xxxxxxxxxx>]

Am 01.09.24 um 06:35 schrieb BB:
> I guess you realise that the co-ordinate realms of the DocumentViewer 
> and the pdf RectF are completely different. I guess we need to 
> determine a range based on the topmost-leftmost character and the 
> bottommost-rightmost character position (as the last line may not be 
> as wide) for both realms and work out a conversion from there.

I don't think that this will solve the problem. One of my dirty 
coordinate conversions  ...

Paint.Rectangle(aRectF[i].x + aRectF[i].x / Width * 10, height - 12 - 
aRectF[i].y * (height + 10) / Height, aRectF[i].w, aRectF[i].h * Height 
/ Width * 0.7)

works pretty well on the sample PDF included in the demo app, which is 
based on text only. But then I tried a PDF document out of the real 
world with header, footer, pictures etc. and found that the RectF arrays 
delivered by the PDFDocument-Class are completely unusable. Even the 
number of search hits per page were wrong in one case (5 instead of 6). 
To my regret I can't share the PDF as proof.

> It may also help to understand how DocumentViewer renders a pdf 
> document. Which I dont. 
I suggest to put the focus on the PDFDocument-class first and hope that 
someone can tell us how to use the search results and why the number of 
search hits can be wrong.

Best Regards, Claus