pdfbox-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andreas Girgensohn <andreasg...@gmail.com>
Subject Re: Font size and text height in PDFBox 0.8.0
Date Fri, 20 Feb 2009 19:54:26 GMT
Hi Andreas

Thanks for the explanation on the font size. This still leaves the
issue with the zero height. This mostly happens when the font size is
one like in my previous example. I found one example with a sensible
font size where the height still is zero. In this case, the width is
very small so that this could be a scaling problem.


Here is an example for the other font related problem that causes an
exception in COSDictionary.getNameAsString because
COSString{HeadingPaginationFont} is stored in the dictionary.


Please let me know if you would like to have more samples.


On Fri, Feb 20, 2009 at 2:03 AM, info@lehmi.de <info@lehmi.de> wrote:
> Hi Andreas
>> I'm using PDFBox to extract text, bounding boxes, and font information
>> from PDF files from a variety of sources. Mostly in files with Type 3
>> fonts but also in others org.apache.pdfbox.util.TextPosition does not
>> return the correct information. In those cases, getHeight returns 0
>> and getFontSize returns 1 (the latter happens much more frequently).
>> PDFBox 0.8.0 (from the svn truck) addresses the issue for about one
>> third of the documents that had problems in PDFBox 0.7.3.  Here is an
>> example of a document that it especially bad. PDFont also does not
>> have any base font information, maybe because of the Type 3 fonts.
> The problem is the way some pdf-generators produces their documents.
> There is the pdf command Tj to set the font size directly and that is
> the result you see using TextPositon.getFontSize(). But in many cases
> the font size is set to default size 1 and it is scaled to the real size
> through the textmatrix. PDFBox reads and uses both to draw the string
> with the right scaling. So every time the expected result is the same,
> wether the pdf-doucment uses Tf = 12 and Tm = 1 or the other way round
> Tf = 1 and Tm = 12.
> I'll extend the TextPosition to get the size as a combination of the
> fontsize and the scaling.
>> P.S.: For a few documents, I ran into a different font related issue
>> (see stack trace below). I added a print statement to determine the
>> values that cause the problem.
> Can you provide us an example for this issue?
> Andreas Lehmkühler

View raw message