Is there a tool like Sigil to edit mobi ebooks? - Page 5

DiapDealer · 08-15-2014, 01:17 PM

Note that if you're talking about old-fashioned mobis, there are no classes to rename. Because neither CSS nor class names are used in their internal markup. KindleUnpack will take those files and produce a single html-ish file you can work with. No need to convert them to htmlz. If you're talking about the newer KF8 files, then calibre edit or KindleUnpack will produce editable markup with the "original" markup and class names. You can still have the issue of multiple HTML files in that case, but I'd like to point out that both calibre's editor and Sigil both have a "Merge" feature (at least I think calibre's editor has that feature ... I know it can Split files) that allows you to combine HTML files without converting.

eschwartz · 08-15-2014, 01:27 PM

I think an excellent solution would be a calibre editor plugin that opens any type of ebook for editing, by first using the Input plugins to unpack the ebook into an OEB directory, then passing the OEB into the editor.

DaleDe · 08-15-2014, 06:24 PM

Quote:

Originally Posted by DiapDealer

Note that if you're talking about old-fashioned mobis, there are no classes to rename. Because neither CSS nor class names are used in their internal markup. KindleUnpack will take those files and produce a single html-ish file you can work with. No need to convert them to htmlz. If you're talking about the newer KF8 files, then calibre edit or KindleUnpack will produce editable markup with the "original" markup and class names. You can still have the issue of multiple HTML files in that case, but I'd like to point out that both calibre's editor and Sigil both have a "Merge" feature (at least I think calibre's editor has that feature ... I know it can Split files) that allows you to combine HTML files without converting.

True enough about the old mobi and even the newer KF8 file which is compiled with pointers to objects, no named classes or named anything else. Compiled means there is nothing to look up as everything is in memory and simply pointed to. There is only compiled binary. KindleUnpack has to make up names for practically everything including the file names themselves, image names and CSS entries since all of that was lost when the Mobi was compiled. By comparison Epub is easy, it is never compiled, only compressed.

Dale

DiapDealer · 08-15-2014, 06:43 PM

Quote:

Originally Posted by DaleDe

True enough about the old mobi and even the newer KF8 file which is compiled with pointers to objects, no named classes or named anything else. Compiled means there is nothing to look up as everything is in memory and simply pointed to. There is only compiled binary. KindleUnpack has to make up names for practically everything including the file names themselves, image names and CSS entries since all of that was lost when the Mobi was compiled. By comparison Epub is easy, it is never compiled, only compressed.

Dale

Surprisingly enough, the class names in the markup (and their corresponding css entries) do not have to be invented when "unpacking" KF8. Everything else, yes.

Freeshadow · 08-17-2014, 02:38 PM

I remember having read somewhere that the Kindlegen tool packs (at least parts of) the source epubs into the ebooks. It was mentioned in the context of debloating the books. AFAIR If you used an epub as source it is added as a whole to the contents of the Kindle ebook. If you pointed to a folder of a decompressed epub, only the parts needed for structure are added, so that you don't double the content. Both would explain the preservation of internal names.

DiapDealer · 08-17-2014, 04:56 PM

Quote:

Originally Posted by Freeshadow

I remember having read somewhere that the Kindlegen tool packs (at least parts of) the source epubs into the ebooks. It was mentioned in the context of debloating the books. AFAIR If you used an epub as source it is added as a whole to the contents of the Kindle ebook. If you pointed to a folder of a decompressed epub, only the parts needed for structure are added, so that you don't double the content. Both would explain the preservation of internal names.

Yes, but the source portion appended to the Kindlegen output doesn't make it all the way to an end user. It's stripped out of the file that someone gets when purchasing a Kindlebook. So unless users are using calibre-edit or KindleUnpack on their very own kindlegen/kindle-previewer generated output, the appended source just isn't there.

Freeshadow · 08-17-2014, 10:14 PM

Oh, interesting. According to that, kindle books sold directly by the creator are more bloated, as if they were when aquired via amazon?
And - if it is so, is there a way to strip the home brewed ones?

DaleDe · 08-17-2014, 10:29 PM

Quote:

Originally Posted by DiapDealer

Surprisingly enough, the class names in the markup (and their corresponding css entries) do not have to be invented when "unpacking" KF8. Everything else, yes.

Thanks for that information. It is strange they would preserve this when they don't need to.

Dale

DaleDe · 08-17-2014, 10:31 PM

Quote:

Originally Posted by Freeshadow

Oh, interesting. According to that, kindle books sold directly by the creator are more bloated, as if they were when aquired via amazon?
And - if it is so, is there a way to strip the home brewed ones?

Yes there is. It is called Kindle strip and available the same Kindle forum and Kindle Unpack. I always use it before uploading a mobi file to our upload area for exactly that reason.

Dale

eschwartz · 08-17-2014, 10:35 PM

If they are sold directly by the creator, they cannot have been made by kindlegen -- it is against the kindlegen licensing.

Here in the MR library, of course, they aren't sold but freely offered. So kindlegen is permitted for usage.

Doitsu · 08-18-2014, 04:01 AM

Quote:

Originally Posted by DaleDe

Yes there is. It is called Kindle strip and available the same Kindle forum and Kindle Unpack. I always use it before uploading a mobi file to our upload area for exactly that reason.

Kindlestrip is no longer required, if you use KindleGen with the -dont_append_source parameter.

DaleDe · 08-18-2014, 12:17 PM

Quote:

Originally Posted by Doitsu

Kindlestrip is no longer required, if you use KindleGen with the -dont_append_source parameter.

Good point although it may be useful if KindleGen is automatically started by another application.

Dale

KevinH · 08-19-2014, 10:48 AM

Hi All,

Just to add fuel to the fire ... we now have a very rough but working python 2.7 plugin capability already added to Sigil (look for an alpha in two or three weeks or so) that will support "edit", "input", and "output" python plugins and has many of the features of the calibre edit style plugins.

The python interpreter itself will not come as part of Sigil (how many python interpreters does one machine really need ... ;-) , but will rely on the user installing ActiveState's Active Python 2.7 release on his machine if they already do not have a working python 2.7 version (Most Mac OS X and Linux users already have python 2.7 installed by default).

Once plugins are working, our first order of business will be to add a Sigil KindleUnpack plugin to give Sigil the capability to read non-drm'd Kindle ebooks directly. This is easy to do as it pretty much already exists for calibre. The idea is Sigil KindleUnpack plugin will load the "source" if the source is included in the mobi, otherwise the default epub created from the KF8/AZW3 side of the mobi will be used, but if no KF8, it will decode the old mobi html into more modern xhtml and allow it to be imported.

Similarly, using python, we can create an gui interface to kindlegen and run kindlegen as part of an output plugin for Sigil and have it purposefully include the source archive so that nothing is lost in the cycle (mobi -> epub -> mobi).

As time allows, there will also be a Python 3 plugin capability and a Lua plugin capability added to Sigil.

Our long term plan is to add support to Sigil for as many of the epub 3 features (and metadata) that are supported by current level of WebKit/WebView in Qt 5.

So as user_none (the maintainer of Sigil) has said .. development is not dead, it was dormant .... it just needed some volunteers to move it forward. Now development has started again but is moving forward slowly. Hopefully we can gather Cpp volunteers and Python Plugin developers to continue improving it.

Hope this will heat up the flames of the debate!

Take care,

KevinH

KevinH · 08-22-2014, 03:30 PM

Hi,

If anyone is interested in the progress (it is working on Mac OS X and Linux) you can checkout the source from:

https://github.com/kevinhendricks/Sigil

Hopefully after some code reviews and more testing, binary alpha builds will be made available for interested people.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Release : New Tool to Edit ePubs TOC, Edit with Sigil and keep you TOC	Nigol	ePub	105	10-29-2012 12:40 PM
RPATHS - edit them by hand? a helpful tool	twobob	Kindle Developer's Corner	14	10-01-2012 09:28 PM
ePub Validation Fail after Sigil edit	dm1839	Sigil	7	10-16-2011 08:59 PM
whishlist: button/command edit in Sigil	booksonthemove	Calibre	3	03-03-2010 07:30 AM

08-15-2014, 01:17 PM	#61
DiapDealer Grand Sorcerer Posts: 28,044 Karma: 199464182 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	Note that if you're talking about old-fashioned mobis, there are no classes to rename. Because neither CSS nor class names are used in their internal markup. KindleUnpack will take those files and produce a single html-ish file you can work with. No need to convert them to htmlz. If you're talking about the newer KF8 files, then calibre edit or KindleUnpack will produce editable markup with the "original" markup and class names. You can still have the issue of multiple HTML files in that case, but I'd like to point out that both calibre's editor and Sigil both have a "Merge" feature (at least I think calibre's editor has that feature ... I know it can Split files) that allows you to combine HTML files without converting.

08-15-2014, 01:27 PM	#62
eschwartz Ex-Helpdesk Junkie Posts: 19,421 Karma: 85397180 Join Date: Nov 2012 Location: The Beaten Path, USA, Roundworld, This Side of Infinity Device: Kindle Touch fw5.3.7 (Wifi only)	I think an excellent solution would be a calibre editor plugin that opens any type of ebook for editing, by first using the Input plugins to unpack the ebook into an OEB directory, then passing the OEB into the editor.

08-17-2014, 02:38 PM	#65
Freeshadow temp. out of service Posts: 2,803 Karma: 24285242 Join Date: May 2010 Location: Duisburg (DE) Device: PB 623	I remember having read somewhere that the Kindlegen tool packs (at least parts of) the source epubs into the ebooks. It was mentioned in the context of debloating the books. AFAIR If you used an epub as source it is added as a whole to the contents of the Kindle ebook. If you pointed to a folder of a decompressed epub, only the parts needed for structure are added, so that you don't double the content. Both would explain the preservation of internal names.

08-17-2014, 10:14 PM	#67
Freeshadow temp. out of service Posts: 2,803 Karma: 24285242 Join Date: May 2010 Location: Duisburg (DE) Device: PB 623	Oh, interesting. According to that, kindle books sold directly by the creator are more bloated, as if they were when aquired via amazon? And - if it is so, is there a way to strip the home brewed ones?

08-17-2014, 10:35 PM	#70
eschwartz Ex-Helpdesk Junkie Posts: 19,421 Karma: 85397180 Join Date: Nov 2012 Location: The Beaten Path, USA, Roundworld, This Side of Infinity Device: Kindle Touch fw5.3.7 (Wifi only)	If they are sold directly by the creator, they cannot have been made by kindlegen -- it is against the kindlegen licensing. Here in the MR library, of course, they aren't sold but freely offered. So kindlegen is permitted for usage.

08-19-2014, 10:48 AM	#73
KevinH Sigil Developer Posts: 8,160 Karma: 5450818 Join Date: Nov 2009 Device: many	Hi All, Just to add fuel to the fire ... we now have a very rough but working python 2.7 plugin capability already added to Sigil (look for an alpha in two or three weeks or so) that will support "edit", "input", and "output" python plugins and has many of the features of the calibre edit style plugins. The python interpreter itself will not come as part of Sigil (how many python interpreters does one machine really need ... ;-) , but will rely on the user installing ActiveState's Active Python 2.7 release on his machine if they already do not have a working python 2.7 version (Most Mac OS X and Linux users already have python 2.7 installed by default). Once plugins are working, our first order of business will be to add a Sigil KindleUnpack plugin to give Sigil the capability to read non-drm'd Kindle ebooks directly. This is easy to do as it pretty much already exists for calibre. The idea is Sigil KindleUnpack plugin will load the "source" if the source is included in the mobi, otherwise the default epub created from the KF8/AZW3 side of the mobi will be used, but if no KF8, it will decode the old mobi html into more modern xhtml and allow it to be imported. Similarly, using python, we can create an gui interface to kindlegen and run kindlegen as part of an output plugin for Sigil and have it purposefully include the source archive so that nothing is lost in the cycle (mobi -> epub -> mobi). As time allows, there will also be a Python 3 plugin capability and a Lua plugin capability added to Sigil. Our long term plan is to add support to Sigil for as many of the epub 3 features (and metadata) that are supported by current level of WebKit/WebView in Qt 5. So as user_none (the maintainer of Sigil) has said .. development is not dead, it was dormant .... it just needed some volunteers to move it forward. Now development has started again but is moving forward slowly. Hopefully we can gather Cpp volunteers and Python Plugin developers to continue improving it. Hope this will heat up the flames of the debate! Take care, KevinH

08-22-2014, 03:30 PM	#74
KevinH Sigil Developer Posts: 8,160 Karma: 5450818 Join Date: Nov 2009 Device: many	Hi, If anyone is interested in the progress (it is working on Mac OS X and Linux) you can checkout the source from: https://github.com/kevinhendricks/Sigil Hopefully after some code reviews and more testing, binary alpha builds will be made available for interested people.

Advert

Advert