Hi,
Do you still have the code that worked?
I'm not the text extraction specialist here, but what I did was to look
in the uncompressed source of the PDF. The stream has code like this:
0 0 0 rg
0 0.5019 0 rg
1 0 0 rg
The first line sets to black, the second to green, the third to red. And
from what I saw, it can't work at all, because the "rg" operator isn't
processed when extracting text, because PDFTextStripper.properties
doesn't contain the "rg" operator. (The operator is in another list,
which is used when rendering)
So that is what puzzles me. I think it can't work at all. But you said
it did work at a time.
Tilman
Am 27.07.2014 07:43, schrieb Tilman Hausherr:
Hi,
Please upload the PDF somewhere and post the URL, PDF files are
removed from the mailing list.
Tilman
Am 27.07.2014 02:35, schrieb -A:
Hello again. I've been trying to figure out this issue that has come
up for me and in my research I found someone posting on StackOverflow
(http://stackoverflow.com/questions/10844271/how-to-get-font-color-using-pdfbox)
a similar issue where they could not read any colors from a PDF. The
user posted the code and someone else took it, ran it, and reported
that it worked. The users approach was different than mine, but alas.
I'm not sure at this point what is going on. I have stepped through
each individual character and checked the PDGraphicsState object, and
even when I am looking at an open file with visibly red text
(attached) the debugger only reports DeviceGray. If I print out the
ColorSpace name from the PDGraphicsState this is what is printed -
for every character.
I would appreciate if someone could perhaps run the attached text
stripper with the attached PDF file and report back if it actually
prints trueinstead of false, as it does for me. Since I saw this
occurrence elsewhere I'd like to rule that out - in case an IDE
setting of some sort may be causing this?
It should be noted that I began using PDFBox with 1.8.5 and had this
code working fine. Still with 1.8.5 yesterday it was failing.
Upgrading to 1.8.6 yielded the same results.
If this is an actual issue I do not mind attempting to solve it if
someone may have a general idea where to point me as to prevent
needless meddling with graphics state objects. Or, if this should be
reported I can do that as well.
Thanks!
-Aaron
*Previous Message:*
*
*
*
*
I've attached an updated stripper file with the only addition being a
main function to test the class specifically.
When ran with the PDF I have also attached it indeed does not
recognize the red text.
At this point it seems that this issue is solely dependent on PDFBox.
I'll stay tuned for some insight hopefully. If any other information
is needed, let me know!