[GUI Plugin] Extract ISBN

[ownedbycats]

Look at the Library Codes plugin.

[BetterRed]

Originally Posted by ownedbycats

Look at the Library Codes plugin.

Look at this:

Originally Posted by SidJnsn

. . .
I know that the Library Codes plugin gets LCC and DDC data, but sometimes . . .

BR

[ownedbycats]

I managed to miss that line.

[xochiluvr]

Calibre/plugin question: can this plugin be used to fix errors in other calibre fields? I seem to have a lot of ebooks that were imported with the author and book title fields switched.

[JSWolf]

Originally Posted by xochiluvr

Calibre/plugin question: can this plugin be used to fix errors in other calibre fields? I seem to have a lot of ebooks that were imported with the author and book title fields switched.

This plugin is for finding the ISBN of an eBook.

Where do you get these eBooks where the author and title are switched?

[xochiluvr]

I think it's a matter of importing my dad's old ebooks that have been converted across formats through the years, from who knows where back in the 90's. So I may have a book titled "Isaac Asimov" authored by "Foundation and Earth" or I may have the correct title but the author name is "Asimov, Isaac"

[theducks]

Originally Posted by xochiluvr

I think it's a matter of importing my dad's old ebooks that have been converted across formats through the years, from who knows where back in the 90's. So I may have a book titled "Isaac Asimov" authored by "Foundation and Earth" or I may have the correct title but the author name is "Asimov, Isaac"

2 problems, different corrections methods.
Title<>Author
Select all with this issue: tap 'E' Metadata edit (bulk mode: tick the box to swap them:OK

Authors with commas:
Use the Quality check plugin: Fix section to flip author order of the ones that are wrong.
Some folk like Ln, Fn
Some want Fn Ln

[ownedbycats]

I came across a somewhat amusing false-positive:

I was processing metadata for Halo: Contact Harvest by Joseph Staten. There's an ISBN (9781982111700) listed on the Kobo store page but that's not anywhere in the book itself see the edit.

However, Extract ISBN did detect one of these (from 10% and 29% in):

“48789-20114-AJ,” Avery mumbled.

“48789-20114-AJ,” she said, pulling a COM-pad from the purse, and reading from its screen. “That is your serial number, correct?”

Log:

Spoiler:

Starting job: Extract ISBN for 1 books 
=================================================== 
Title:  Halo: Contact Harvest 
Format: EPUB 
Path:   D:\Documents\My Books\Joseph Staten\Halo_ Contact Harvest (8304)\Halo_ Contact Harvest - Joseph Staten.epub 
--------------------------------------------------- 
  Scanning first 10, then last 5, then remaining 24 files 
      Invalid ISBN match: -000987111 
      Invalid ISBN match: -000614236 
      Invalid ISBN match: -16523.14.82 
      Valid ISBN10: 4878920114 
  Scan time: 1.81 secs 
  The isbn was found in 1.81 secs 
  New ISBN extracted of: 4878920114 
=================================================== 
Scan complete, with 0 failures

This coincidentally ends up being a valid ISBN-10 for what seems to be some kind of comic/manga which is definitely not related to my sci-fi action-game expanded universe novel.

I'm not sure if there's anything that can be done about this (or whether anything should be done). I'm willing to send a copy of the book if you'd like to test further though.

EDIT: To make things even weirder, I just found the real ISBN was in the book, the last line of the book. It wasn't detected though.

[kiwidude]

@ownedbycats - yup nothing I can do about that sort of situation - if it is a 'valid' ISBN number then as I have said dozens of times in this thread there is no way for the plugin to divine whether it is a false positive.

As for why it didnt find the valid one at the end of the book - the plugin first looks at the front, and if it finds a valid ISBN then it will stop scanning rather than going on to look at the back of the book. After all what is the point - it cannot determine which of the two are "more likely" to be "correct" without building in a ton more complexity so it just takes the first valid one it finds.

It's a tool that "most" of the time gives a quick win for people as a starting point, but the programming effort to make it slightly more reliable is way too much given it can never be cast iron guaranteed to be 100% correct with what it comes up with. Blame the publishers.

[The Holy]

The plugin won't extract an ISBN from any PDF I have and simply results in “ISBN not found”, but the description says it checks them too, which makes me think I'm missing something. None of the PDFs are scanned, so what's wrong?

[Wiggo]

Nothing, doesn’t work on Windows.

If you had read back a few pages, you would have found a workaround, but it has other disadvantages.

[The Holy]

What disadvantages would that be?

[theducks]

Originally Posted by The Holy

What disadvantages would that be?

False Positive.
ISBN are patterns, so 13 digits is not different than any other 13 digits without other checks, Even with those, a false one can pass

You might do some research and see how they are assigned and may be printed. OH! Groupings for one

[The Holy]

I think he meant there are disadvantages that come with replacing the pdfinfo.exe included in calibre, so I'm wondering what those are. False positives are rarely an issue for me with this plugin.

[Wiggo]

As I wrote a few pages ago Count Pages doesn’t work with the replacement.
It looks like there’s another workaround for this but never tried.

For me Count Pages is more important than getting ISBN for PDFs.
In PDFs there are almost always more than one ISBN and getting the right one is a lottery