OS X Printing Workflow/Automator action for use with Calibre

[DaveNB]

I was thinking the other day I should automate some of the command line steps that I use repeatedly to create LRF documents for my Sony Reader. Specifically I wanted to be able to "print" directly to a LRF file formatted for the Sony Reader.

Sure you could set up custom paper sizes and a print profile for generating .PDF files formatted for the Sony Reader, but the results are inconsistent and often look terrible (not to mention are slow to load/turn pages and have reduced contrast due to the way PDF's are rendered on the Sony Reader).

Based on the fact that OS X renders all it's print jobs in PDF format before sending them to the printer and also based on the fact that Calibre does a pretty darned good job of reflowing PDF's, I came up with the following Automator script that executes the following command line utility (assuming you've already installed Calibre):

cd $HOME/Desktop    # This is where the LRF file will be saved
while ( $# )
	pdf2lrf --enable-autorotation --left-margin=2 --right-margin=2 --top-margin=2 --bottom-margin=2 --page-break-before-tag='\$' --blank-after-para "$1"
	shift
end

Just drag a copy of this Automator script (included in the .zip file attached to this post, unzip it first or it won't work) into the "PDF Services" folder in your "Library" folder (located in your Home directory). I'd suggest you first open it up in Automator to have a look at it's code to make sure that it does what you want it to (ie. you may want to change the destination for the LRF file creation or some of it's margin settings, etc) and more importantly: to make sure it isn't doing something malicious!! Remember, all Automator scripts/applications have the potential to harm to your local user account so due diligence is recommended.

Then, from any application that supports printing, you can select the "Create Sony Reader LRF" option in the drop down menu triggered by the "PDF" button that appears in every print dialogue box (when you select Print from the File menu or when you type <command>+<p>). The resulting LRF file will show up on your Desktop.

I'm not sure if anyone else has already figured out how to do this, but it works great for me. No more saving to HTML or printing to a PDF file, then manually typing the command line html2lrf/pdf2lrf commands. Now it'll all happen quickly and painlessly from the Print dialogue box. You can also use this Automator action as a folder action or create a self contained Applescript Application/Bundle that'll covert any PDF dropped into it.

I use it to create LRF's from my daily RSS news feed for reading on the road and for creating quick little LRF's from web pages that I'm viewing (web pages that have a printer friendly format come out nicely).

Feel free to modify and change this little Automator script to your liking. I'm going to try to figure how to add some more intelligence to it (ie. lets you specify the Author, Title, etc).

Not all of the formatting is perfect on some pages.
Things to fix:
1) There are some pages that render with missing line breaks.
2) I'd like to remove all page breaks (placed in the PDF's as a part of the formatting to fit on your set paper size) which the "--page-break-before-tag='\$'" is supposed to do (but isn't doing consistently).

The next version may do the conversion in 3 steps:
1) PDF -> HTML conversion
2) Strip and modify the HTML to fix problems listed above using sed/grep and other text processing tools
3) Create the LRF

It still need some work, but as a quick and dirty way to make LRF's, it's pretty good.

Enjoy,

Dave

[sammykrupa]

Thanks, this might even allow the Reader to work with my mother!

[kf6gpe]

Aces, thank you!! I was just sitting down to figure out something like this. I'll give this a try and see how it works. Cheers!

[akirk]

Thanks DaveNB! Just found out about this.

I have added something to the code to transfer the generated lrf directly to the device (in my case a prs500, not tested with others):

cd $HOME/Desktop    # This is where the LRF file will be saved
while ( $# )
	pdf2lrf --enable-autorotation --left-margin=2 --right-margin=2 --top-margin=2 --bottom-margin=2 --page-break-before-tag='\$' --blank-after-para "$1"
	set n=`basename "$1" .pdf`
	prs500 cp "$n.lrf" prs500:/Data/media/books/
	shift
end

So now, if the device is connected, the file is auto-transfered to it. Makes it even better for me

[pthwaite]

I assume (as I have a Cybook) to run the same script for Mobi output you would just change the pdf2lrf to pdf2mobi and change the lrf to mobi and Cybook elsewhere?
Then, as you say it would save a lot of messing around.
Howard

[akirk]

Originally Posted by pthwaite

I assume (as I have a Cybook) to run the same script for Mobi output you would just change the pdf2lrf to pdf2mobi and change the lrf to mobi and Cybook elsewhere?

Well, maybe just try it prs500 is actually a command line program, so I am not sure if it can write to Cybook.

$ prs500 cp
Usage: prs500.py cp [options] source destination
Copy files to/from the device

One of source or destination must be a path on the device. 

Device paths have the form
prs500:mountpoint/my/path
where mountpoint is one of / or card:/

Check the output of that command, maybe it says something another device string for the mountpoint.

[pthwaite]

Ho hum...
I try the original script and it goes away and appears to do something, but no output appears on the desktop?

Now, I don't have calibre in the Applications directory, I run it from an external drive (Mac mini with small drive). Could it be that it cannot find Calibre? If so where do I edit it to point to the Calibre installation.

Sorry to appear dim.

[akirk]

Originally Posted by pthwaite

I try the original script and it goes away and appears to do something, but no output appears on the desktop?

Start up a Terminal and type

pdf2lrf

What does it say? For me that program resides in

/usr/bin/pdf2lrf

You can type

ls /usr/bin/pdf2lrf

to check if it is there. If not, you might have a problem with your installation. If it shows up, you can try to open up the workflow in Automator and change the pdf2lrf call to include the path.

[pthwaite]

Well mine is in:
/usr/bin
as well and it lists it, after all Calibre can find it.

However I cannot find any output generated when I run it. ie printing this web page using it suggests it has printed 4 pages, but they don't appear anywhere. Then again it doesn't ask for a file name output.

[akirk]

Originally Posted by pthwaite

Well mine is in:
/usr/bin
as well and it lists it, after all Calibre can find it.

However I cannot find any output generated when I run it. ie printing this web page using it suggests it has printed 4 pages, but they don't appear anywhere. Then again it doesn't ask for a file name output.

Well, the 4 pages you see are for the PDF that is generated and handed to the script. Try opening up the .workflow file from DaveNB by double clicking it and change the reference to pdf2lrf to /usr/bin/pdf2lrf and save it. Then try again.

[pthwaite]

Nothing, running Console gives:
03/03/2009 21:51:07 Automator Runner[2879] Error while processing arguments

Oh well. I'll have to dig deeper.

You don't need Xcode installed do you?

regards

[akirk]

Originally Posted by pthwaite

Nothing, running Console gives:
03/03/2009 21:51:07 Automator Runner[2879] Error while processing arguments

I had that a lot myself when trying to enhance the script, too

Originally Posted by pthwaite

You don't need Xcode installed do you?

I do have it installed, but I doubt you need it.

Just briefly, how I did the debugging, was this:
Put the contents from the script in the .workflow file into a .tcsh file, insert

#!/bin/tcsh

as the first line in the script, make it executable

chmod 755 test.tcsh

and then run it with a PDF as parameter

./test.tcsh "my pdf file.pdf"

This will give you more useful output than through Automator.

This is getting pretty detailed, maybe we should to continue in direct msging?