Mobi -> Mobi "conversion"

Jellby · 07-06-2009, 06:15 AM

Hello,

I have a mobipocket file that has every paragraph with hard-coded left alignment, and I'd like to change it and remove the explicit alignment. I don't want to change anything else in the file (metadata, etc.). What would be the best way to achieve this?

After trying unsuccessfully with MobiPerl I managed to install Calibre, and tried with mobi2oeb, edit the .html, and oeb2mobi, but something is not fine here:

1. The original .mobi file is 2.6MB, the converted one is 1.6MB

2. The converted file adds vertical space above every paragraph. Indeed, when I convert again this new file to oeb, the .html file has "margin-top: 1em" for every paragraph, which was not present initially.

3. The original language (according to mobi-meta) is ENGLISH, the language of the converted file is NEUTRAL

So, suggestions or advice for altering only the text in the mobipocket file and in a controlled way?

EDIT: I've tried with version 0.5.11, by the way.

JSWolf · 07-06-2009, 07:19 AM

Try mobi2oeb and then edit the HTML and try rebuilding it backinto a Mobipocket format eBook.

Jellby · 07-06-2009, 07:39 AM

Quote:

Originally Posted by JSWolf

Try mobi2oeb and then edit the HTML and try rebuilding it backinto a Mobipocket format eBook.

That's what I did, or I don't quite understand... I said I "tried with mobi2oeb, edit the .html, and oeb2mobi, but something is not fine here"

JSWolf · 07-06-2009, 08:05 AM

Is this file one you can attach here?

Jellby · 07-06-2009, 08:12 AM

No, sorry

Jellby · 07-06-2009, 08:40 AM

... but I can post this one, which shows similar problems.

a.mobi is just my Mobipocket version of "The Picture of Dorian Gray". It was created with MobiPerl's html2mobi, from a mobi-flavoured HTML (no mobipocket creator involved).

a-dir is what resulted from

Code:

mobi2oeb a.mobi -o a-dir

b.mobi is the result of

Code:

oeb2mobi a-dir/The\ Picture\ of\ Dorian\ Gray.opf -o b.mobi

b-dir is the result of

Code:

mobi2oeb b.mobi -o b-dir

I expected a.mobi and b.mobi to be almost the same, but a.mobi is 322KB and b.mobi 548KB.

I also expected a-dir and b-dir to have very similar contents, but not even the images are the same (they have different sizes). And the HTML in a-dir has simple ... tags, while the one in b-dir has ...

kovidgoyal · 07-06-2009, 11:23 AM

Don't use oeb2mobi, use any2mobi on the OPF file. Also make sure you have justification turned off.

Jellby · 07-06-2009, 12:09 PM

Quote:

Originally Posted by kovidgoyal

Don't use oeb2mobi, use any2mobi on the OPF file. Also make sure you have justification turned off.

I tried this:

Code:

any2mobi --no-justification -o b.mobi a-dir/The\ Picture\ of\ Dorian\ Gray.opf

(after deleting the justification from the <style> and <body> tags of the HTML in a-dir)

And the result is a bit better, but still:

1. b.mobi is quite larger than a.mobi
2. The images in both a-dir/images and b-dir/images are not the same size
3. The HTML in b-dir still has instead of simply , which is cumbersome and "wrong" (I don't want explicit top margin or indent).
4. The language of b.mobi is NEUTRAL instead of ENGLISH

Wouldn't it be possible to have a couple of commands to just unpack and pack a mobi file with no data loss? (I admit I have no real knowledge of the internal structure of mobipocket format.)

tompe · 07-06-2009, 12:44 PM

Quote:

Originally Posted by Jellby

Wouldn't it be possible to have a couple of commands to just unpack and pack a mobi file with no data loss? (I admit I have no real knowledge of the internal structure of mobipocket format.)

MobiPerl tried to take this approach but I do not think it is perfect.

Calibre uses a totally different approach since it converts to a middle format and then have backends for the different output formats.

Jellby · 07-06-2009, 12:51 PM

Quote:

Originally Posted by tompe

MobiPerl tried to take this approach but I do not think it is perfect.

Alas, no. For the particular mobi file I wanted to modify, it gave an empty html file, probably because of the compression.

tompe · 07-06-2009, 12:57 PM

Quote:

Originally Posted by Jellby

Alas, no. For the particular mobi file I wanted to modify, it gave an empty html file, probably because of the compression.

Yes, that is usually the reason.

One obvious feature of Calibre would be to just change the compression.

kovidgoyal · 07-06-2009, 01:21 PM

Having specialised tools for every format quickly leads to insanity. Instead calibre takes the approach that you write your HTML once and it will do its best to render it as faithfully as possible in all the formats.

There is probably some CSS lurking around somewhere in your decompiled MOBI that's causing the margins. Try setting --override-css "p {margin-top: 0pt}"

Jellby · 07-06-2009, 01:47 PM

Quote:

Originally Posted by kovidgoyal

Having specialised tools for every format quickly leads to insanity. Instead calibre takes the approach that you write your HTML once and it will do its best to render it as faithfully as possible in all the formats.

And I think that's a good approach. But my needs here are a bit different. I'm not trying to create a mobipocket file from my custom HTML, all I want is minimally change the HTML (or whatever) that is already in the mobipocket file. Maybe the mobipocket file has unsupported tags, or unclosed elements... but I don't care, I don't want them corrected (that's good in general, but that's not what I want now, I mean).

See, with ePUB is easy. I can unzip the .epub file, change "black" to "white" and zip everything again, and I don't care whether or not the XHTML validates or CSS classes are undefined. Well, I was thinking that it would be good to be able to do the same with mobipocket files, if you see what I mean.

Of course, it's your program and you do what you see fit

I just thought that since there are already programs that can parse and create mobipocket files, it shouldn't be too hard to have a program that blindly unpacks and packs mobipocket files.

But anyway, I'm still confused that mobi->oeb->mobi->oeb generates such a "mess". Maybe the mobi->oeb conversion is more or less lossless, but then oeb->mobi tries to fix the HTML? I thougt that mobi->oeb would already fix things so that the subsequent oeb->mobi->oeb would be rather transparent.

kovidgoyal · 07-06-2009, 02:20 PM

I don't actually recall what mobi2oeb and oeb2mobi do (it's ages since I've used an 0.5 version of calibre) but at least mobi2oeb converts mobipocket markup to proper HTML and I believe (though I'm not sure) oeb2mobi converts HTMl to mobipocket markup

AlexBell · 07-08-2009, 09:41 AM

Quote:

Originally Posted by kovidgoyal

Having specialised tools for every format quickly leads to insanity. Instead calibre takes the approach that you write your HTML once and it will do its best to render it as faithfully as possible in all the formats.

There is probably some CSS lurking around somewhere in your decompiled MOBI that's causing the margins. Try setting --override-css "p {margin-top: 0pt}"

I must have missed the announcement that your thesis was accepted.

Congratulations!

Regards, Alex

07-06-2009, 06:15 AM	#1
Jellby frumious Bandersnatch Posts: 7,570 Karma: 20150435 Join Date: Jan 2008 Location: Spaniard in Sweden Device: Cybook Orizon, Kobo Aura	Mobi -> Mobi "conversion" Hello, I have a mobipocket file that has every paragraph with hard-coded left alignment, and I'd like to change it and remove the explicit alignment. I don't want to change anything else in the file (metadata, etc.). What would be the best way to achieve this? After trying unsuccessfully with MobiPerl I managed to install Calibre, and tried with mobi2oeb, edit the .html, and oeb2mobi, but something is not fine here: 1. The original .mobi file is 2.6MB, the converted one is 1.6MB 2. The converted file adds vertical space above every paragraph. Indeed, when I convert again this new file to oeb, the .html file has "margin-top: 1em" for every paragraph, which was not present initially. 3. The original language (according to mobi-meta) is ENGLISH, the language of the converted file is NEUTRAL So, suggestions or advice for altering only the text in the mobipocket file and in a controlled way? EDIT: I've tried with version 0.5.11, by the way. Last edited by Jellby; 07-06-2009 at 07:06 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Transfer Web pages to Kindle - what do you use? "Print to MOBI" an option?	guiyoforward	Amazon Kindle	8	09-06-2010 01:50 AM
EPUB to MOBI: "Failed: Convert book 1 of 1"	Jillo	Introduce Yourself	4	02-20-2010 07:49 PM
Content Mobi Reference ("Huge Collections") on Kindle	Mike L	Amazon Kindle	26	04-28-2009 04:14 PM
"Secure" PDF and "Secure" Mobi docs?	AceHarddrive	iRex	9	05-08-2008 10:13 PM
New ".mobi" domain names are coming in May 2006	Bob Russell	Lounge	3	04-25-2006 06:38 PM

07-06-2009, 07:19 AM	#2
JSWolf Resident Curmudgeon Posts: 80,784 Karma: 150249619 Join Date: Nov 2006 Location: Roslindale, Massachusetts Device: Kobo Libra 2, Kobo Aura H2O, PRS-650, PRS-T1, nook STR, PW3	Try mobi2oeb and then edit the HTML and try rebuilding it backinto a Mobipocket format eBook.

07-06-2009, 08:05 AM	#4
JSWolf Resident Curmudgeon Posts: 80,784 Karma: 150249619 Join Date: Nov 2006 Location: Roslindale, Massachusetts Device: Kobo Libra 2, Kobo Aura H2O, PRS-650, PRS-T1, nook STR, PW3	Is this file one you can attach here?

07-06-2009, 08:12 AM	#5
Jellby frumious Bandersnatch Posts: 7,570 Karma: 20150435 Join Date: Jan 2008 Location: Spaniard in Sweden Device: Cybook Orizon, Kobo Aura	No, sorry

07-06-2009, 11:23 AM	#7
kovidgoyal creator of calibre Posts: 45,609 Karma: 28549044 Join Date: Oct 2006 Location: Mumbai, India Device: Various	Don't use oeb2mobi, use any2mobi on the OPF file. Also make sure you have justification turned off.

07-06-2009, 01:21 PM	#12
kovidgoyal creator of calibre Posts: 45,609 Karma: 28549044 Join Date: Oct 2006 Location: Mumbai, India Device: Various	Having specialised tools for every format quickly leads to insanity. Instead calibre takes the approach that you write your HTML once and it will do its best to render it as faithfully as possible in all the formats. There is probably some CSS lurking around somewhere in your decompiled MOBI that's causing the margins. Try setting --override-css "p {margin-top: 0pt}"

07-06-2009, 02:20 PM	#14
kovidgoyal creator of calibre Posts: 45,609 Karma: 28549044 Join Date: Oct 2006 Location: Mumbai, India Device: Various	I don't actually recall what mobi2oeb and oeb2mobi do (it's ages since I've used an 0.5 version of calibre) but at least mobi2oeb converts mobipocket markup to proper HTML and I believe (though I'm not sure) oeb2mobi converts HTMl to mobipocket markup