Converting EPUB to AZW3/MOBI

[eschwartz]

If the books were painstakingly tweaked and you still *need* the old copies, that is versioning.

If you need a version per device, I cannot help you... note that calibre doesn't support that anyway.

[cybmole]

I call it archiving.
versioning is wanting two or more files of a given format, not one of each.

this is not versioning, it's diversity
1 x file.original_epub
1 x file.epub
1 x file.mobi
1 x file.azw
with all 4 selectable as source formats for conversion

versioning would be
file at date a.epub
file at date b.epub
....

the original_FMT files behaves like an archive format in most instances already e.g. it cannot be opened in calibre viewer, or in calibre editor....

[itimpi]

For archive purposes, why not attach a RAR file as an additional format, and then store any Archive versions within the RAR file. That way you can have Calibre keep the archived version with the current book entry, and it is easy to get things out of the archive by double-clicking it when required. Not a perfect solution but definitely workable.

If you do not want to pay for a RAR license then use something like 7-Zip instead. Avoid storing any archive files as a ZIP file as this can get confused with the way Calibre handles HTML files.

[cybmole]

i have rar & 7z, but that sort of work-around will confuse the database integrity checks,

Once calibre sees that original_FMT has been deleted, it's not then going to know that it is back again when manually extracted from the rar or via a rename. Maybe a library repair will pick it up again but what a rigmarole, compared with turning on a select epub as source format conversion option.
Even though in practice, I've not ever really needed the library mainenence tools, why deliberately do stuff that will mess them about.

Luckily I mostly prefer reading my books on Kindle Fire in their epub format via mantano app, vs reading an azw version in the native kindle app, so this is an user-choice-principle argument.
(I can't customise left handed page turn controls in kindle app , like I can in Mantano)

[itimpi]

Originally Posted by cybmole

i have rar & 7z, but that sort of work-around will confuse the database integrity checks,

I find it does not as long as you add it as a book format via Calibre.

Once calibre sees that original_FMT has been deleted, it's not then going to know that it is back again when manually extracted from the rar or via a rename. Maybe a library repair will pick it up again but what a rigmarole, compared with turning on a select epub as source format conversion option.
Even though in practice, I've not ever really needed the library mainenence tools, why deliberately do stuff that will mess them about.

It easy to open the RAR (via Calibre) and then drag-drop a file from the RAR archive back into Calibre so that it knows it is there again.

[eschwartz]

Originally Posted by cybmole

I call it archiving.
versioning is wanting two or more files of a given format, not one of each.

this is not versioning, it's diversity
1 x file.original_epub
1 x file.epub
1 x file.mobi
1 x file.azw
with all 4 selectable as source formats for conversion

versioning would be
file at date a.epub
file at date b.epub
....

the original_FMT files behaves like an archive format in most instances already e.g. it cannot be opened in calibre viewer, or in calibre editor....

versioning.

1x book.epub
1x book.old-version_epub

Play the semantic games all you want, I don't really care. You have two versions of one format (as can be ascertained by examining the contents and finding an EPUB).
Regardless of whether you call it versioning or archiving, you will get no help from calibre. Because calibre is not meant to do it. So any degree to which calibre does so anyway is to be taken as a blessing.

Adding new features -- because that is what this is -- is something to treat as... <drumroll>... a Feature Request!
Regardless of how *reasonable* you think it is, the fact remains that Kovid rejected your request -- and really, I agree with that decision.

And again, it seems like the root of the problem is your desire for a customized EPUB per device. Leaving aside the issue of how wrong that mindset is (ask the people who literally make their living off EPUB design), calibre does not support this.

Why not just make a different database record for each of your versions?

[Section8]

I also tweak books when I buy them and like to have an "archived" copy to fall back on if I screw up. I just have a separate, physical library called "originals" and copy new books there before touching them.

[BetterRed]

Originally Posted by itimpi

I find it does not as long as you add it as a book format via Calibre.

It easy to open the RAR (via Calibre) and then drag-drop a file from the RAR archive back into Calibre so that it knows it is there again.

@itimpi - that's true if you have WinRAR, you can drag it out of the RAR and drop it into the Book Details panel.** But only if the extension is recognised by calibre as an ebook extension, in this context, it would seem that xxxx.ORIGINAL_EPUB is not an ebook extension.

Also you can't drag files out of a RAR and drop into the Book List via the WinRAR browser - not sure if that's a WinRAR or calibre issue.

So the only way out is to unpack the archive into a folder. If the RAR is in the book folder (which in this context it will be), the default location will be a subfolder in the book folder. You can then drag the ORIGINAL_EPUB from the newly created subfolder into the Book List. What happens next depends on the settings in Add books for dealing with duplicates - at worse you'll have to Merge books. Then you need to trash the subfolder that WinRAR created.

In the context of this discussion that's not what I'd call 'easy', at best I'd called it 'fiddlesome'. It might be overcome if calibre allowed drag/drop of non-ebook files into Book Details, but... the answer on that was... No. But even then it would only work if you have WinRAR.

BR

** BTW, you can't drag/drop ebook files out of a RAR to Book Details with Peazip, nor perhaps 7zip.

[cybmole]

it's also darn near impossible to write a batch file that will go through all the calibre subfolders and move all orig_TMT into rar or 7z containers.
none of the suggested workarounds, here or previously (for which i am grateful BTW), are easy to automate. bulk rename via a batch file process and then sacrifice library integrity/repair checks is looking simplest.. I have not tested the one I linked to before, but I examined the coding in notepad++ & can see how it works.

PS the fact that folks are offerign up workarounds is encouraging, either that implies i'm not totally stuipd to want what I want, & that they have done similar things , or that some folks are super helpful here

[PeterT]

Its TRIVIAL to write such a batch file.....

The jist of it will be

for /r %1 in (*.original_fmt) do some_command_involving_%1

This will search for original_fmt files in all child directories of the current directory.

To see all the options on the for command, in a command window type

for /?

. There are many MANY options to parse the filename (ie the %1).

[cybmole]

back up your claim please:

rather than say it is TRIVIAL, & then walk away - how about you prove it by giving us the batch code to do what' been suggested i.e. to scan an entire calibre library & place each original_FMT file into its own individual .rar or an 7z container of the same name , in the correct book subfolder.

('cos I don't think it is trivial at all but will be happy to be proved wrong/ to learn how )

What was was suggested will require repeatedly invoking the main RAR or 7z .exe and handing it each found file via command line options?, without losing control of overall flow.

you lose marks if I am left with several hundred instances of RAR /7z still open at end of batch run.

and then write the matching one that reverses all of the above, if you can.

PS for a gold star you also have to do all of the above in less time that it would have taken to teak calibre source file select options in convert screen

[PeterT]

Well never having used 7za (the command line tool) it took a few minutes

for /r %1 in (*.original_epub) do 7za a "%~p1""%~n1".7z "%1"

This will create a new 7Z archive within the same folder that the original_epub file was located with the same name as the original_epub that contains a single file.

There are NO processes left running after this command; the only issue is that I'm not sure how to have 7za drop the file after it has been added. This could be done by a second for command

for /r %1 in (*.original_epub) do del "%1"

PS: I only SAW your reply about 10 minutes ago; in that time I located the 7zip command line version, installed it, wrote the batch line, tested it, AND wrote this message.

PPS: Quite likely no need to compress due to the fact that an epub actually IS a zip file (albeit with the possibility that not all components of it might be compressed. So one could add the -x0 switch to the command line

[cybmole]

Impressive,but for 2 gold stars can you automate putting humpty dumpy back together again I.e. make all those orig.fmt files reappear where they used to be

[PeterT]

Sigh..... All that is needed is to use 7za to EXTRACT the file....

for /r %1 in (*.7z) do 7za e "%1" -o"%~p1"

[PeterT]

Just to explain a few things

for /r says scan the current directory and all sub directories
"%1" ensures that if the matched file name contains spaces there are no issues
%~p1 extracts the directory name
%~n1 extracts just the file name (with no path or extension)

AH... I could have simplified

for /r %1 in (*.original_epub) do 7za a "%~p1""%~n1".7z "%1"

to

for /r %1 in (*.original_epub) do 7za a "%~pn1".7z "%1"