Recently learned to run "Extract ISBN" before "Download Metadata". Do U do that too?

[droopy]

I like having a book's metadata:

series
all people involved in book (author, translator, editor) in the author field/column
title
Book description in the Comments field.


I'm not so much a stickler about:
"Publisher", in part because publishing houses merge.
Publishing date (month + year)
Tags (though they help me later, when I'm in a mood for a particular interest (read "tag")

In the past, when I have books that are imported into my Calibre library, I'd select a bunch of them and do "Download Metadata". But now, I do a preliminary step: I find the ISBNs from within the book itself with the Extract ISBN plugin. With this initial step, it seems that the "Download Metadata" runs faster.

Questions:
1. Do you have the same experience?
2. Any other tips on how I can speed up my processing of books?

[thiago.eec]

Originally Posted by droopy

In the past, when I have books that are imported into my Calibre library, I'd select a bunch of them and do "Download Metadata". But now, I do a preliminary step: I find the ISBNs from within the book itself with the Extract ISBN plugin. With this initial step, it seems that the "Download Metadata" runs faster.

Questions:
1. Do you have the same experience?
2. Any other tips on how I can speed up my processing of books?

Almost (or maybe all) the Metadata Source Plugins uses about the same logic:

1) First, look for their own identifier. I you have a identifier set for that specific plugin (skoob, amazon, google, etc), then go directly to the book page.
2) Second, use ISBN search.
3) Third, use TITLE/AUTHOR search. This is the last resource, since it tends to be more slow, because it may find various editions of the same book, and have to decide which one is more appropriated.

[droopy]

Originally Posted by thiago.eec

Almost (or maybe all) the Metadata Source Plugins uses about the same logic:

1) First, look for their own identifier. I you have a identifier set for that specific plugin (skoob, amazon, google, etc), then go directly to the book page.
2) Second, use ISBN search.
3) Third, use TITLE/AUTHOR search. This is the last resource, since it tends to be more slow, because it may find various editions of the same book, and have to decide which one is more appropriated.

Hi thiago thanks for your reply.
About step 1, does that mean it looks for the identifier within the book file itself? In other words, step 1 can be run without an internet connection, the way that Extract ISBN plugin can run without internet?

[droopy]

thiago,
i have now enabled a few more "Metadata source" plugins. Here are the active plugins:
Amazon.com
Big Book Search
Edelweiss
Google
Google Images
Open Library
Overdrive

Even with all those enabled plugins, "Download Metadata" still gave the wrong metadata and cover for one particular PDF. But when I run the Extract ISBN plugin on the same PDF, it was able to get the right ISBN.

[thiago.eec]

Originally Posted by droopy

Hi thiago thanks for your reply.
About step 1, does that mean it looks for the identifier within the book file itself? In other words, step 1 can be run without an internet connection, the way that Extract ISBN plugin can run without internet?

Lets break this down to parts:

a) When creating the e-book, identifier can be set inside it. If that's the case, when you add the book to your library, the Identifiers field will be it filled automatically (Ex.: amazon_br:8543101875), the same way ISBN will be filled if present on e-book file.
b) The Metadata Source Plugin will not extract the Identifier from e-book file. That happens when you add it to your library, if it was correctly set in the file.
c ) So, without internet connection, nothing will happen. When connected, the plugin will search the website based on either its identifier, ISBN or Title/Author, in that order, if they are available.

[thiago.eec]

Originally Posted by droopy

thiago,
i have now enabled a few more "Metadata source" plugins. Here are the active plugins:
Amazon.com
Big Book Search
Edelweiss
Google
Google Images
Open Library
Overdrive

Even with all those enabled plugins, "Download Metadata" still gave the wrong metadata and cover for one particular PDF. But when I run the Extract ISBN plugin on the same PDF, it was able to get the right ISBN.

Extracting the ISBN from file is just a reading operation. It does not have to guess anything. Either it is set or not.

When you run "Download Metadata", plugins will "try" finding a book that matches your data (Identifier, ISBN, Title, Authors). The quality of this data will determine how good the match is. When you get a wrong match, you can try eliminating your current ISBN, so it will try using Title/Authors.

[droopy]

Ah, so my current process of running Extract ISBN first is better, since there's no guesswork involved.

[thiago.eec]

Originally Posted by droopy

Ah, so my current process of running Extract ISBN first is better, since there's no guesswork involved.

If you can set the correspondent Identifier or ISBN, before running Metadata Download, then you will have the best match.

If the extracted ISBN is the correct one, then the plugin will find it. But, the ISBN inside the book might not be correct all the time.

[droopy]

Originally Posted by thiago.eec

If the extracted ISBN is the correct one, then the plugin will find it. But, the ISBN inside the book might not be correct all the time.

Interesting. Do you mean that there may be a typo to the ISBN? Is that what you mean by "not correct"?

[theducks]

Originally Posted by droopy

Interesting. Do you mean that there may be a typo to the ISBN? Is that what you mean by "not correct"?

An ISBN 10 has a MOD11 check digit at the end. Calibre used (waaaay back when) to turn that field Red if there was an error. been so long (almost 10Y) , since I used that, I am not sure

ISBN13 is a standard EAN13 check algorithm

[droopy]

Originally Posted by theducks

An ISBN 10 has a MOD11 check digit at the end. Calibre used (waaaay back when) to turn that field Red if there was an error. been so long (almost 10Y) , since I used that, I am not sure

ISBN13 is a standard EAN13 check algorithm

I'm processing a lot of my books now and I did find a typo. The ISBN in the book is valid, but it's an ISBN for another work of the author. So I had to manually search for the correct ISBN.

[thiago.eec]

Originally Posted by droopy

I'm processing a lot of my books now and I did find a typo. The ISBN in the book is valid, but it's an ISBN for another work of the author. So I had to manually search for the correct ISBN.

That's what I was talking about: an ISBN from another book (but must possibly, from another edition/publisher).

Like @theducks said, if an invalid ISBN is extracted (or manually typed), the box will be red, although it won't prevent you from setting it anyway.

[JSWolf]

Sometimes you'll end up with the ISBN for the pBook version when both are given in the copyright,

[droopy]

Originally Posted by JSWolf

Sometimes you'll end up with the ISBN for the pBook version when both are given in the copyright,

I don't mind. In fact, I often use the ISBN of the pback version. The ISBN for an electronic version may be limited to a particular webstore/publisher (e.g. Kindle Store, Kobo Store, etc.). This unnecessarily limits the metadata downloaded

[davidfor]

Originally Posted by droopy

I don't mind. In fact, I often use the ISBN of the pback version. The ISBN for an electronic version may be limited to a particular webstore/publisher (e.g. Kindle Store, Kobo Store, etc.). This unnecessarily limits the metadata downloaded

In that case, why bother extracting the ISBN? The metadata source plugins work as @thiago.eec said in their first response. If the library already has an ISBN it will be used, but, if there is none, or the ISBN isn't found on the source site, the title and author will be used. I don't really find that having the ISBN helps the accuracy that much.

And for curiosity, I ran Extract ISBN across 50 books that I already had an ISBN for. The plugin didn't find an ISBN in 27 of them. So, I'd be wasting my time for half the books in my library. I find that having the correct title and author is enough to find the book.