Can Mars convert existing PDF to PDFXML?

Question

Hello, I am trying to find a way to process the text content
of PDF files inside an AIR app I'm building. So far, I haven't had
any luck figuring out how to get the content of an existing PDF so
that I can process it (essentially I want to search the text for
specific regular expressions).
I'm wondering if I could use Mars to convert an existing PDF
into PDFXML -- which I could then process? Is this possible?
Follow-up: can I do this all in-memory without having to
write out the XML to disk temporarily? Any comments on the
performance considerations of doing so? The PDF files vary in size
from under 1MB to perhaps 10-15MB.
Thanks in advance,
Davis

Answer 1

Hi Davis,
Sorry for the delay in responding. I've been thinking a
little about how this could be done, but at the moment, I would
probably say it is not possible. Mars/PDFXML uses UCF for packaging
its contents (UCF being a zip file with some restrictions, but is
still a fully conformant ZIP file) and we use a separate UCF reader
inside of the PDFXML plug-in (this is an Adobe created UCF
library). This library abstracts out how we access the contents of
the UCF package and how it is handled on disk. It also provides
byte serving negotiation for web pages. What it doesn't do today is
have any mechanism to keep everything in memory or to provide
access to an external client to the data we are adding to it. Until
we serialize this out to disk, it is pretty much inaccessible.
If you had a PDFXML file, I think that accessing it through
AIR/Flex would be a lot easier than accessing the contents of a PDF
file. But there is no way to automate this directly from a source
PDF.
Matt