Calibredb export

[phossler]

Is there any 'selectivity' to control which subset of files are exported?

For example, I wanted to export epubs to a single folder, but only if the Tag=Science Fiction.


http://manual.calibre-ebook.com/cli/...libredb-export

It looks like I can list IDs, but that's not an easy thing to do

Export the books specified by ids (a comma separated list) to the filesystem. The export operation saves all formats of the book, its cover and metadata (in an opf file). You can get id numbers from the list command.

Thanks

[kovidgoyal]

As the help says, you get the list of ids by using

calibredb list

for example if you want to get the ids for books by Asimov, you use

calibredb list --search author:asimov

[eschwartz]

https://github.com/kovidgoyal/calibr...b8f5fac295a6b8
New command: `calibredb search "some-search"` to output a list of ids to use in some other command.

Thanks for the new option, Kovid!
(Much easier than commandline json parsing.)

[kovidgoyal]

You dont actually need to parse JSON to get a list of ids from calibredb list, you can just do, for example:

calibredb list --search whatever -f title --separator , | cut -d, -f1 | tail -n +2 | awk -vORS=, '{ print $1 }' | sed 's/,$/\n/'

Or on windows, where none of the unix tools are available, you can just use calibre-debug

calibredb list --search whatever -f title  | calibre-debug -c "import sys; print ','.join(filter(None, [line.partition(' ')[0].strip() for line in sys.stdin][1:]))"

But, I realize this kind of thing is hard for most people, hence calibredb search

[phossler]

Originally Posted by kovidgoyal

As the help says, you get the list of ids by using

calibredb list

for example if you want to get the ids for books by Asimov, you use

calibredb list --search author:asimov

Actually I saw that in the manual but I couldn't figure out how to pass it to 'calibredb export' to only export epub books with certain tags

I did try various command lines, but

calibredb list --search format:=epub and tags:="science fiction"

seems to list non-Science Fiction and then error out

[kovidgoyal]

You need to quote things properly, like this

calibredb list -f title --search 'format:=epub and tags:="science fiction"'

And note that the windows console dies if you output too much too it, which is the cause of the IOError you got.

[phossler]

Originally Posted by kovidgoyal

You need to quote things properly, like this

calibredb list -f title --search 'format:=epub and tags:="science fiction"'

Sorry to be a pest, but I must be doing SOMETHING wrong.

Nothing lists now

I used your example and variations

Thanks

[kovidgoyal]

I forgot, this is windows, I always forget that windows cmd.exe has insane quoting rules -- here you go

calibredb list -f title --search ^"format:=epub and tags:=^"science fiction^"^"

[phossler]

1. Thanks

2. I NEVER would have gotten that, ... but

calibredb list -f title --search ^"format:=epub and tags:=^"science fiction^"^"

is only finding tags = "Science" and not "Science Fiction"

Reading the manual about bracketing parameters with spaces with quotes, and assuming that the ^ is treated as a special escape character, what you provided looks like it should work.

It works great on a 'one word tag', but not one with a space

Can you provide a tweak for cmd.exe's insane quoting rules that works with spaces?

[kovidgoyal]

Not right now as I dont have access to a windows machine and this is likely to require some trial and error

[phossler]

Thanks -- no problem

I've already trialed and errored some variations that seems reasonable (to me anyways)

I suppose I could just re-tag as "ScienceFiction" but that feels like cheating

(I don't suppose that there's any chance calibredb is not parsing the command line correctly is there???)

[kovidgoyal]

No, no chance

IIRC you can use PowerShell which has better quoting rules, so my original recommendation would work

calibredb list -f title --search 'format:=epub and tags:="science fiction"'

[phossler]

1. The new calibredb search option works well and the parameters are intuitive (even using cmd.exe)

>calibredb search "tags:science fiction" format:mobi > SF_MOBI.TXT will return the intersection of "science fiction" and mobi (ID=515 in the screen shot)

Now I just have to figure out how to pass the txt file as a list of IDs to caliber export

2. As a 'By the way' the calibredb list --search ..... does seem to operate differently (other screen shot)

Thanks for the new CLI option

[kovidgoyal]

That is because everything after search is assumed to be part of the search expression, whereas when you use --search it has to consume only a *single* argument, which is why you have to enclose the entire search expression in quotes.

I have no idea if this works in cmd.exe, but in most modern shells you can simply do this:

calibredb export `calibredb search whatever`

[chaley]

Originally Posted by phossler

Now I just have to figure out how to pass the txt file as a list of IDs to caliber export

The following works in Win 7 dos box, of course with appropriate changes for your configuration.

for /f "usebackq" %a in (`calibredb search "tags:science fiction"`) do calibredb export %a --to-dir=bar

And yes it is indeed ugly, but it works.