[Old Thread] Export notes from sony PRS700 reader

[ogduartev]

Hi,
just a question... How can I export the notes I have written in the electronic reader?

thanks in advance...

[kovidgoyal]

You can't, since the notes and stored in ebook files, but rather in a separate XML file. A rather silly way of doing things. I might eventually add support for exporting/importing the notes using calibre, but that will take some time.

[mishuxp]

Hello,

Any update on this feature?

thanks!

[itimpi]

Have you raised a Ticket (or checked to see if there is already one raised) for this as a feature request. If not, then I would suggest that you do as otherwise the request can get lost amongst the flood of other items that need looking at.

[mishuxp]

Thanks!

Now I have: http://calibre.kovidgoyal.net/ticket/3415

And what can I do now?

[itimpi]

Wait I guess

At least now you have raised a ticket it will show up whenever these are reviewed to see what work should be done next.

[luqmaninbmore]

The annotations for epub and pdf books are stored in XML files in the annotations folder.

[Elsi]

Originally Posted by luqmaninbmore

The annotations for epub and pdf books are stored in XML files in the annotations folder.

Does that mean that you can easily identify them and transfer them from the reader to a PC/Mac when you've got the reader mounted as a disk drive? Is the XML schema available so that you can parse it?

[kovidgoyal]

Originally Posted by Elsi

Does that mean that you can easily identify them and transfer them from the reader to a PC/Mac when you've got the reader mounted as a disk drive? Is the XML schema available so that you can parse it?

Yeah it's trivial to do, I just haven't implemented it in calibre because it needs some infrastructure in calibre itself.

[mishuxp]

Originally Posted by kovidgoyal

Yeah it's trivial to do, I just haven't implemented it in calibre because it needs some infrastructure in calibre itself.

So we should expect this feature soon?

[kovidgoyal]

Originally Posted by mishuxp

So we should expect this feature soon?

No promises

[mishuxp]

Hello!

Any future prospects on this marvelous feature ?

[jbone1313]

+1 from me.

[thingumybob]

I would LOVE to be able to somehow extract the notes and annotations from my Sony Reader (I have the PRS 600 Touch). It is to me inexplicable that Sony don't permit this with the Reader software (although frankly, that software is cr*p), so I was hoping calibre would permit it by now? If not I will do it myself if somebody can lead me to where the file/s to decrypt are hidden! Thanks.

[Man Eating Duck]

Originally Posted by thingumybob

... I was hoping calibre would permit it by now? If not I will do it myself if somebody can lead me to where the file/s to decrypt are hidden! Thanks.

I actually had a look at this earlier as I'd also like this feature. While I'm not a programmer by trade I have fairly extensive experience with scripting languages, although not so much with Python.

Sony seems to use three different schemes for saving annotations (I have only looked at ebub files, for other formats there might be other syntaxes, after all a PDF does not conform to an xml structure):

* Freehand scribbles: Saved as SVG, plus an entry in an .annot-file which indicates the position relative to page. Saved in for instance \database\markup\database\media\books\Miguel de Cervantes\The Story of Don Quixote\The Story of Don Quixote - Miguel de Cervantes_45.epub\1277997910049.445.svg on the reader.

* Highlighted text: Entries indicate start/end for a "marked" section. The reference includes filename, but it appears that start and end in different xhtml-files is not supported (the annotation won't show on the reader).

* Bookmarks: As for highlighted text, but the position indicates a single character only (probably first character of the relevant page)

The positions are saved in a .annot file which looks suspiciously like an xml file: For instance \Digital Editions\Annotations\database\media\books\Miguel de Cervantes\The Story of Don Quixote\The Story of Don Quixote - Miguel de Cervantes_45.epub.annot

The thing is, the marked text is not itself contained in the XML, it needs to be extracted from the files in the epub. An example reference to a highlight is:

<target>
<fragment start="29468/0.html#point(/1/4/1/1)" end="29468/0.html#point(/1/4/1/1:27)"/>
</target>

When I tried to examine the xpointer-ish entry it doesn't seem to be valid for the indicated file, it's probably a reference to an internal rendering of the text. All points seem to start with "/1/4/....." as indicated above, but when I try to find this entry via xpointer in the file I can't use the reference to get to the correct text. For the example it's easy (it's the very first 27 characters of the body), but for instance "29468/0.html#point(/1/4/34/1)" does not refer to the 34th xhtml element of the body. It seems that it ignores certain tags when counting, but I don't know which ones (again, probably it refers to some internal rendering).

In addition I'm not familiar with how I could access the relevant text in an epub from code within Calibre, nor how to handle the content with the Calibre data structure if I did manage to extract it. If a more accomplished Python coder would like to try reverse-engineering it I'll help in any way that I can

Edit: I also found a file \database\cache\cacheExt.xml, which seems to contain references in a more readable format. It also stores up to a hundred characters of text from highlights, but it's possible to highlight more than that on the reader. I suspect this might be that Sony wants to avoid issues with copyright holders. To get these would definitely be an improvement.