| View previous topic :: View next topic |
| Author |
Message |
jigsaw
Joined: 02 Nov 2006
Posts: 1
Location: Australia
|
 Posted: Thu Nov 02, 2006 9:06 pm Post subject: Split PDF based on text within the PDF |
 |
|
Hi,
I was wondering if the PDFsharp object could be used to find text within a PDF file and retrieve the page number it was on. What I need to do is split a PDF file based upon finding some text.
ie. The top of every page has the text
[Customer:xxxxxxxx]
where xxxxxxxx is the customer name. When the xxxxxxxx changes I need to split the PDF. So a single PDF with 10 pages, 4 which are for Customer X, 3 for Customer Y and 3 for Customer Z would need to produce 3 files,
one for customer X of three pages
one for customer Y of four pages
one for customer Z of three pages
It would also be great if the text search could use a regular expression.
Is this possible with PDFsharp? |
|
| Back to top |
|
 |
Stefan Lange
Joined: 12 Oct 2006
Posts: 47
Location: Cologne, Germany
|
| Posted: Thu Nov 02, 2006 10:04 pm Post subject: |
|
|
Hello,
the content of a PDF page is a sequence of bytes that represents graphical commands. These bytes are called the "content stream" of the page. You can get it uncompressed with this code:
| Code: | | page.Contents.CreateSingleContent().Stream.UnfilteredValue; |
A "Hello, World" page may look like this:
| Code: | 1 0 0 1 0 841.8898 cm
1 0 0 -1 0 0 cm
BT
-100 Tz
/F0 -10 Tf
1 0 0 1 70.8661 80.9199 Tm
-10 TL
(Hello)Tj
/F0 -10 Tf
1 0 0 1 99.4111 80.9199 Tm
(World!)Tj
ET |
You can find and as strings.
You should find | Code: | | ([Customer:xxxxxxxx]) |
in your PDF file. This is easy to parse. Try PdfSharp Explorer to analyse your PDF.
But depending on the PDF producer application you find this:
| Code: | | [(H)42(e)32(l)37(l)33(0)]TJ |
There is kerning information (distance adjustment) between the characters. Adobe Acrobat never creates this if you use a fixed size font like Courier. Unfortunately tools like FreePDF always creates distance information, even if it is superfluous.
We at empira currently have the same problem to identify address information in PDF files and split it into single files. We recommend using Adobe Acrobat as producer and Courier New as font for the information text.
Further I wrote the class PdfSharp.Pdf.Content.ContentReader to convert a content stream into a squence of operation (it is in the current source code). Maybe this reader helps you to find your text.
I will publish our solution if we have one (currently we are working on other things).
Regards
Stefan Lange |
|
| Back to top |
|
|
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
