[GUI Plugin] Modify ePub

[JSWolf]

Originally Posted by kerb

His comment is at the very end of the blog entry
http://dsandrei.blogspot.co.uk/2012/...19243870487287

Incorrectly posted link. Please repost a correct link.

[PeterT]

Try http://dsandrei.blogspot.co.uk/2012/...19243870487287

[Barb-B]

Originally Posted by Barb-B

Kiwidude - I suppose I may not be the only person who suggested it, but I was thrilled to see your latest modification, allowing a command line method to use the plugin tools without adding a book to Calibre library. Thanks SO much!

I only found it because I came looking for the url to give someone I wanted to point here. I'll be sure to recommend a contribution as well...

@Kiwidude - I put off trying to figure out what to do with this new item until recently. I haven't been able to get it to work. Attached is a screenshot showing my PowerShell input and messages, along with the folder where I put the me.py file and the epub I wanted to fix. It worked just fine if I did it the manual import to Calibre way. Any chance you could help me understand what I've done wrong? Thanks!

[JSWolf]

Originally Posted by Barb-B

@Kiwidude - I put off trying to figure out what to do with this new item until recently. I haven't been able to get it to work. Attached is a screenshot showing my PowerShell input and messages, along with the folder where I put the me.py file and the epub I wanted to fix. It worked just fine if I did it the manual import to Calibre way. Any chance you could help me understand what I've done wrong? Thanks!

The only thing I can think of is maybe the ePub has DRM and if so, dropping it into Calibre (with the proper plugins) is stripping the DRM and allowing Modify ePub to work.

[Barb-B]

@JSWolf - thanks for the suggestion. But I can assure it's DRM-free. Maybe Kiwi will still have some additional clue to offer...

[cybmole]

I put a retail epub into calibre, then used the plugin only to strip out ADOBE files & links i.e. ticked only those 2 boxes. The output looked fine in calibre & in sigil but when opened on Sony reader, it caused reader to crash & restart. No error message ( no invalid page error) - just white screen, couple of seconds delay then restart is observed

I am wondering why - I scrolled through this thread but saw no obvious explanation. it is time consuming to debug as it needs a physical connect device, send to device, eject... sequence.

I can work around it, so I am really just flagging this out of curiosity

I have checked
1. the unaltered retail epub opens OK on Sony reader.
2 the tweaked epub is now also OK after doing a calibre epub to epub convert (after stripping the adobe stuff.)

I also observe that this epub has an embedded set of charis fonts - dunno if that is relevant

I am left wondering what the plug-in could have done, or what I could accidentally done to cause the fail on Sony. I did what I usually do, which was to strip unwanted TOC, dedication, blurb... pages before using the plug-in, but that is usually harmless

[JSWolf]

I've seen the code after Modify ePub does it's thing. Nothing I've seen would crash a 650. I suggest trying again without ePub > ePub converting.

[cybmole]

ok - the crash is probably not worth pursuing. I must have done other manual damage via sigil before the modify epub. I re-tried only the modify step on the original source & the output opened OK in ADE.

[JSWolf]

Originally Posted by cybmole

ok - the crash is probably not worth pursuing. I must have done other manual damage via sigil before the modify epub. I re-tried only the modify step on the original source & the output opened OK in ADE.

My 650 has only rebooted once and locked up once in it's life. I don't recall what caused the crash, but the lock was after adding a few books. It locked during the cover thumbnail creation. A simple reset and the 650 did create the thumbnails.

[cybmole]

it was not a lock up - it was a repeatable crash to restart triggered by trying to open that epub. so something in the epub code was causing it.
I do still have the file - as original-epub after running the calibre convert which fixed it; but with no clue as to what to look for I don't think it is a worthwhile exercise.
If I copy it to desktop, rename it as epub, and drop it into ADE I get " ADE has stopped working...".

I can open it OK in sigil & run validate with Flight Crew...
no significant errors - just some unreachable stuff e.g.
This OPS document is reachable but not present in the OPF <spine>. "Reachable" means that a reference of some kind that points to this resource exists in the epub.
& some complaint about content.opf version number

I can't post the problem book file due to copyright rules. I suppose I can PM or email it for diagnostics if anyone is interested but it is 4Mb as it contains many pictures

[jackie_w]

Originally Posted by cybmole

I did what I usually do, which was to strip unwanted TOC, dedication, blurb... pages before using the plug-in, but that is usually harmless

When you say 'strip' you did make sure that any files you physically removed were also removed from the opf <manifest>, <spine> and <guide> sections, didn't you? Retaining references to non-existent files would cause problems.

[cybmole]

could be - I don't know how to edit .ofp - usually I just delete unwanted toc, dedication, epigraph... XHTML entries...

[jackie_w]

Originally Posted by cybmole

could be - I don't know how to edit .ofp - usually I just delete unwanted toc, dedication, epigraph... XHTML entries...

It's possible that's your problem, then. If you're doing it via Sigil, I expect Sigil does some automatic opf-tidying-up as you remove chunks and save the changes. But I don't know it well enough to know in detail exactly what it does and doesn't clean up.

The <manifest> is a list (id, url, media-type) of all the files in the epub,
the <spine> is an ordered list of ids of the html files which will be displayed when you read the epub,
and the <guide> is... I'm not sure what it's purpose is in an epub but I think the items in there are used by Kindle formats.

[theducks]

Originally Posted by jackie_w

It's possible that's your problem, then. If you're doing it via Sigil, I expect Sigil does some automatic opf-tidying-up as you remove chunks and save the changes. But I don't know it well enough to know in detail exactly what it does and doesn't clean up.

The <manifest> is a list (id, url, media-type) of all the files in the epub,
the <spine> is an ordered list of ids of the html files which will be displayed when you read the epub,
and the <guide> is... I'm not sure what it's purpose is in an epub but I think the items in there are used by Kindle formats.

My observation
OPF (section) Cleanup does not always get done... Unless:

Manifest: you move an existing file. (you can move it back after a move)
Spine: rename the file
Guide: assign a semantic to a different file

[cybmole]

yes I've wasted too much time on this already.
I love a mystery so I took a copy of the faulty version & tried deleting stuff bit by bit & dropping it back into ADE e.g. removed cover, notes, index... eventually it worked but when I started over & tried to replicate that, I failed.
I saw that with the fails, ADE could not produce a sensible thumbnail, but as I say, I cannot replicate the issue reliably so time to move on I think.
At some point I opened & closed the sigil metadata editor but repeating that action alone was insufficient.

The calibre epub to epub convert is a good alternative to banging your head against the wall - it seems to fix stuff that otherwise is not obvious.

I am inclined now to NOT blame the modify epub plugin - I think user error i.e. me, is a far more likely culprit