Split PDF, name odd pages and name even pages in sequence

I recommend PDFNomad for things like this.

Hello

You may try something like the following applescript script, which is a wrapper of rubycocoa script. It will let you choose the source pdf file and split it into odd and even files in the same directory as the original.

set infile to (choose file of type {"com.adobe.pdf"} with prompt "Choose source pdf file.")'s POSIX path

do shell script "/System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/bin/ruby -w <<'EOF' - " & infile's quoted form & "
require 'osx/cocoa'
OSX.require_framework 'PDFKit'
include OSX

raise ArgumentError, \"Usage: #{File.basename($0)} pdf [out_directory]\" unless [1, 2].include? ARGV.length 
infile = File.expand_path(ARGV[0])
outdir = ARGV[1] ? File.expand_path(ARGV[1]) : File.dirname(infile)
raise ArgumentError, \"No such directory: #{outdir}\" unless File.directory?(outdir)

url = NSURL.fileURLWithPath(infile)
doc = PDFDocument.alloc.initWithURL(url)
bname = File.basename(infile).sub(/\\.pdf$/i, '')
front, back = '_Front', '_Back'

(0 .. (doc.pageCount - 1)).each do |i|
    page = doc.pageAtIndex(i)
    doc1 = PDFDocument.alloc.initWithData(page.dataRepresentation)
    doc1.writeToFile(\"#{outdir}/#{bname}#{i % 2 == 0 ? front : back}#{i / 2 + 1}.pdf\")
end
EOF
"

Hope this may help,

H

PS. If copied code has extra spaces in front of every line, which appears to be the case with some browsers including Firefox, please remove them before running the script. Also note that the line holding EOF must be of only three characters EOF without any leading or trailing spaces. Fora software often corrupts the source code in this regard.

Here is an Automator application that prompts for a file, validates that it is a PDF document, and passes the filename to a Bash script that decides even or odd page, and extracts these specific pages into pdfname_front1 ... pdfname_frontn, pdfname_back1 ... pdfname_backn until the entire document is processed. When finished, it pops an AppleScript dialog showing the name of the processed PDF, and the number of front and back extractions completed.

This solution is best used where the PDF file is in a folder, as by default, the extraction process writes the output files back to the original document location.

It requires the Skim application installed in /Applications. This was written and tested on OS X 10.9.4.

Automator Application

The Bash script that replaces the contents of the Run Shell Script window.

#!/bin/bash

#

# Requirement: Install Skim PDF reader

# Location: http://skim-app.sourceforge.net

# Usage: progname.sh sample.pdf

# Author: VikingOSX, August 2014, Apple Support Community

skimPDF="/Applications/Skim.app/Contents/SharedSupport/skimpdf"

pdfFile="$@"

basename="${pdfFile%.*}"

pagecnt=`mdls -name kMDItemNumberOfPages "${pdfFile}" | egrep -io --color=never "([[:digit:]]+)"`

even=1

odd=1

function ASDialog () {

`osascript <<-AppleScript

set pdfFile to "${1}"

set evens to "${2}"

set odds to "${3}"

set msg to ""

set msg to msg & "PDF Processed: " & pdfFile & return

set msg to msg & "Front Pages: " & tab & evens & return

set msg to msg & "Back Pages: " & tab & odds

tell application "System Events"

display dialog msg with title "Processing Complete" giving up after 20

end tell

return quit

AppleScript`

}

for mypdf in "$@"

do

for page in $(seq 1 $pagecnt)

do

if [[ $((page % 2)) == 0 ]]; then

#even extractions

`${skimPDF} extract "${pdfFile}" "${basename}""_Front"$((even)) -page ${page}`

((++even))

else

#odd extractions

`${skimPDF} extract "${pdfFile}" "${basename}""_Back"$((odd)) -page ${page}`

((++odd))

fi

done

done

ASDialog "${pdfFile}" $((--even)) $((--odd))

exit 0

Hello

You're welcome. Glad to hear it helped.

If the script is saved as a compiled script, you may use NSAppleScript's -initWithContentsOfURL:error: method. Something like the following snippet.

    NSDictionary *errdict;
    NSAppleEventDescriptor *desc;

    NSString *path = [[NSBundle mainBundle] pathForResource:@"PDFscript" ofType:@"scpt"];
    NSURL *url = [NSURL fileURLWithPath: path];

    NSAppleScript *scpt = [[NSAppleScript alloc] initWithContentsOfURL: url error: &errdict];
    if (!scpt)
        fprintf(stderr, "Failed to load script: %s\n", [[errdict description] UTF8String]);
    else
    {
        desc = [scpt executeAndReturnError: &errdict];
        [scpt release];
        if (!desc)
            fprintf(stderr, "Runtime error in script: %s\n", [[errdict description] UTF8String]);
    }

However, if you're writing code in objective-c, you can do everything in objective-c without using applescript script which is running rubycocoa script which is invoking cocoa methods in the first place. 🙂

Regards,

H

What version of OS X?

Using the exact same script that I posted, I ran it the following ways. Each time, I removed all extracted pages.

1. Script and PDF in same folder, with just pdf-file.pdf name as command-line argument to script.
2. Same as [1], but pdf-file with no extension. One extraction to _Back1 on 17 page PDF.
3. Script in local bin folder and ~/test/pdf-file.pdf supplied on the command-line.
4. As Automator application on Desktop, choosing the PDF within the test folder location.
5. As [4], but choosing only the containing folder results in no processing and zero dialog results.

   Fix: Change the Ask for Finder Items Automator action Type: to Files from Files and Folders.

Items 1, 3, and 4 worked as expected on 5-page and 17-page PDF test files. As currently written, the script misbehaves when the .pdf extension is missing from the supplied PDF file. Apparently, mdls will return 0 page count for PDF files without the customary extension. Looking into tweaking the mdls part of the script.

Ok, I have solved some issues I found with the previous script and Automator configuration.

1. The Filter Finder Items action apparently uses kMDItemKind to determine the file type, and if you choose a PDF file without an extension, this action will not pass the file to the Bash script. Result: 0.
2. The mdls command, for the same reason will always fail to get a page count when it encounters a extensionless PDF file.

   Remove the entire mdls/egrep command sequence from the script.

3. Wrote a shell function that runs Python, imports the CoreGraphics module, and returns the page count for PDF files with/out extensions. The function takes as its argument, the input PDF filename, which it passes into Python as a command-line argument ($1). It returns a page count.

   Replace entire Bash script in Automator with the following code.

New Bash Script

#!/bin/bash

#

# Requirement: Install Skim PDF reader

# Location: http://skim-app.sourceforge.net

# Usage: progname.sh sample.pdf

# Author: VikingOSX, August 2014, Apple Support Community

function PDFpagecnt () {

/usr/bin/python -c \

"import sys

from CoreGraphics import *

pdfdoc = sys.argv[1]

provider = CGDataProviderCreateWithFilename(pdfdoc)

pdf = CGPDFDocumentCreateWithProvider(provider)

print('{}'.format(pdf.getNumberOfPages()))" "${1}"

}

function ASDialog () {

`osascript <<-AppleScript

set pdfFile to "${1}"

set evens to "${2}"

set odds to "${3}"

set msg to ""

set msg to msg & "PDF Processed: " & pdfFile & return

set msg to msg & "Front Pages: " & tab & evens & return

set msg to msg & "Back Pages: " & tab & odds

tell application "System Events"

display dialog msg with title "Processing Complete" giving up after 20

end tell

return quit

AppleScript`

}

skimPDF="/Applications/Skim.app/Contents/SharedSupport/skimpdf"

pdfFile="$@"

basename="${pdfFile%.*}"

even=1

odd=1

pagecnt=$(PDFpagecnt "${pdfFile}")

for mypdf in "$@"

do

for page in $(seq 1 $pagecnt)

do

if [[ $((page % 2)) == 0 ]]; then

#even extractions

`${skimPDF} extract "${pdfFile}" "${basename}""_Front"$((even)) -page ${page}`

((++even))

else

#odd extractions

`${skimPDF} extract "${pdfFile}" "${basename}""_Back"$((odd)) -page ${page}`

((++odd))

fi

done

done

ASDialog "${pdfFile}" $((--even)) $((--odd))

exit 0

Won’t be able to help with this. Good luck.

Hello

I presume you're talking about how to specify hard-coded path in AppleScript. In AppleScript, you need to specify the full path. $HOME does not work whilst you can get the POSIX path of current user's home directory by

(path to home folder)'s POSIX path

and its desktop by

(path to desktop)'s POSIX path

and so on. Note that these directory paths have trailing /, which in some cases you may need to remove in shell script.

As for the output directory, I've written the rubycocoa script so that it can accept it as the second argument. So you can call it like this:

set infile to (path to desktop)'s POSIX path & "infile.pdf" -- => "/Users/your-user-name/Desktop/infile.pdf"
set outdir to (path to desktop)'s POSIX path & "notecards" -- => "/Users/your-user-name/Desktop/notecards"

do shell script "/System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/bin/ruby -w <<'EOF' - " & infile's quoted form & " " & outdir's quoted form & "
require 'osx/cocoa'
OSX.require_framework 'PDFKit'
include OSX

raise ArgumentError, \"Usage: #{File.basename($0)} pdf [out_directory]\" unless [1, 2].include? ARGV.length 
infile = File.expand_path(ARGV[0])
outdir = ARGV[1] ? File.expand_path(ARGV[1]) : File.dirname(infile)
raise ArgumentError, \"No such directory: #{outdir}\" unless File.directory?(outdir)

url = NSURL.fileURLWithPath(infile)
doc = PDFDocument.alloc.initWithURL(url)
bname = File.basename(infile).sub(/\\.pdf$/i, '')
front, back = '_Front', '_Back'

(0 .. (doc.pageCount - 1)).each do |i|
    page = doc.pageAtIndex(i)
    doc1 = PDFDocument.alloc.initWithData(page.dataRepresentation)
    doc1.writeToFile(\"#{outdir}/#{bname}#{i % 2 == 0 ? front : back}#{i / 2 + 1}.pdf\")
end
EOF"

Hope this helps,

H

Or possibly are you trying to make the script accept directory of source pdf files instead of single source pdf file?

If so, it is not what your original question says and you need to rewrite the ruby script or applescript script so as to iterate through the pdf files in the given input directory. It is not difficult at all but computer is such a simple-minded thing that you have to tell it exactly what you want it to do. 🙂

H