Scribble on PDF can be exported to PDF - linux system

[cadmus]

After many months of looking and bothering you all about how to export highlights, notes or scribbles to a file that is useful (the RTFs that the sony reader software exports was not useful) I found a glimmer of hope:
http://hoheinzollern.wordpress.com/2011/11/27/sony-reader-notes-to-svg-or-pdf/

thanks to Alessandro Bruni for inventing this gem.

I can put python on my PC. But i am guessing this will not work on a windows machine. I figured by posting this some linux users would be able to use it. I am hopeing someone knows how to make this work on my windows xp machine.

Any help is greatly appreciated
Pete

[opitzs]

You should be able to do this on windows using cygwin.
This uses "sed", which is part of cygwin, so you shouldn't have to many problems, you just have to replace the "sed" with "c:\cygwin\bin\sed" or something to that effect (depending on where you installed it).

Thank you very much for the link.

[cadmus]

Ohh. I should have used a different thread title. He made a windows version for me.

[idle]

Thanks for the link, I just tried it and it looks really useful.

[multiprose]

how are you liking it cadmus? i haven't had time to try it but i'll take a look soon.

[cadmus]

Originally Posted by multiprose

how are you liking it cadmus? i haven't had time to try it but i'll take a look soon.

I just got the T1 this weekend and I have not tried to retreive the scribbles yet. I am eager to hear how it worked for others.

Looking at calibre it seems to be very python oriented plugins. I wonder if this could be made to work under a Calibre Gui.

Pete

[cadmus]

Originally Posted by idle

Thanks for the link, I just tried it and it looks really useful.

Great. Can you send examples of what the output pdf looks like? (multipage preferably) Is the scribble vector or raster image?

[taos]

After lurking around this great forum for a couple of months and being inspired by a number of great ideas, I decided to share some of my findings regarding the export and embedding of hand-written annotations (so far, I've only played with simple pdf-files).

For my first tests, I decided to stick with Windows as operating system and tried to avoid to install a lot of additional software.
Already on my computer:
LibreOffice (for its "export to pdf" functionality), PDF24 (a tool to merge pdf files, something similar should work too), Inkscape (vector graphics editing)
Additional installs:
SQLiteSpy (reading of Sony's books.db), NoteToSvg

What I tried:
For testing purposes I made a simple pdf file (Scribble Test - original.pdf) with LibreOffice Writer and transferred it to the PRS-T1. I added a few free-hand annotations (see and ) to the file on the reader and then connected the reader again with my computer.
With SQLiteSpy I looked for the id of the free-hand annotations in books.db and exported the respective files in /Sony_Reader/database/cache/books/BOOK_ID/markup with NoteToSVG.
Then I opened both the first page of the pdf file and the svg file with Inkscape, grouped all elements of the svg and copied the illustration to the window with the pdf (for some reason I had to resize the graphic, so it might be a good idea to mark two corners when annotating on the reader to facilitate aligning). I saved the file as pdf and did the same with the second page.
Both pages were merged with PDF24. As you can see, the resulting pdf file (Scribble Test - annotated pages.pdf) doesn't look too bad - simple text is fine but there are some obvious problems with equations (that were not properly imported by Inkscape). However, it's too much time and effort to do this for complete textbooks or even epubs.

[cadmus]

Well done. That is exactly what i was hoping for. Thanks for the great examples and the help.

You said "However, it's too much time and effort to do this for complete textbooks or even epubs." yeah but it is possible and i am guessing if someone can make a script it can be done in bulk.

I will likley have questions for you later. THANKS. well done.

[cadmus]

When you did this was the scribble from page one a different file than page 2? So if one were to do this on say a 20 page PDF, and they only marked some of the pages... it could become difficult figuring out what scribble was for what page, huh?

[taos]

I suppose it's not too bad: For every page with scribbles there are two files, a pseudo svg (input file for NoteToSvg) and a minuscule thumbnail png file with the same name (but different extension) in Sony_Reader/database/cache/books/CONTENT_ID/markup. The file name corresponds to the entries in columns "svg_file" and "thumbnail" of table "freehand" in PRS-T1's books.db (Sony_Reader/database/books.db). Thankfully, there is also a "page" column (cell data shows "0.9999" for page 1 and "1.9999" for page 2). IIRC from playing around with text annotations in epubs, this entry is always x - 0.0001 for page x - so not too difficult to find...
BTW, "content_id" in "freehand" table corresponds to "_id" of "books" table where you can find "title" and "author" of your annotated pdf file.

[idle]

For Linux users, things are much easier with this tool linked in the discussion. When set up properly (it needs python and some libraries), it does all the merging by itself. You just run the script (from the command line - there's no GUI, but it's so simple there's no need) and it lists all the annotated PDFs in the reader, you choose one and it saves the merged PDF where you want it. Neat.

I haven't done any annotating yet, just what I did for the sake of trying this thing out, but I expect I'll have an opportunity soon.

Originally Posted by cadmus

Looking at calibre it seems to be very python oriented plugins. I wonder if this could be made to work under a Calibre Gui.

That would be even neater.

[rubentje1991]

Interesting... this way my annotations become much more useful...

When I find some time, I'll try this on my linux box

[taos]

Originally Posted by idle

For Linux users, things are much easier with this tool linked in the discussion.

That, of course, is true. On my linux boxes python and most (if not all) of the needed libraries are already installed. I played around on Windows as a non-privileged user to mimic the situation with work PCs that are still stuck with Windows XP and where you aren't allowed to install any software (only workaround is to use portable applications - that exist e.g. for LibreOffice or Inkscape). Both NoteToSvg and SQLiteSpy work without installation on these computers. Only problem in this situation might be the print-to-pdf functionality that might require an installed ghostscript version.

A portable version of Calibre supporting the import and embedding of freehand notes would be really convenient.

[multiprose]

check your inbox =]