Get Books showing if result is in library

[user_none]

I want to add the ability for results in Get Books to be marked if they are already in the users library. The idea being if a user searches for an author they can at a glance see what books they already have by the author.

My first thought is to use the comparison features of the Find Duplicates plugin. However, I'm concerned about performance and I'm also unsure about the best way to do this. The big thing that makes me think I need to use Find Duplicates is the fuzzy matching.

My initial thought is to do a soundex search over each book in the users library using author and title each time Get Books is opened. Then store a hash of the title and authors. Then compute the hash for each result as they come in and check if there is an existing hash from the library.

What I'm looking for is advice on the best way to go about this?

[kovidgoyal]

How fuzzy do you want to get? There is a find_identical_books() method in db2 that does a little bit of fuzzy matching.

But yeah the performance of this on larger libraries (where it is more useful) will be bad enough that you will likely need to run it in a separate thread like the drm status.

[user_none]

Originally Posted by kovidgoyal

How fuzzy do you want to get?

Fuzzy enough that these match:

Toll the Hounds (Malazan Book of the Fallen Series #8) by Steven Erikson

Toll the hounds by Steven Erikson

and

The Hobbit by J RR Tolkien

Hobbit by J. R. R. Tolkien

I'll look at find_identical_books() in db2 and see if it matches enough tests. I know it can't be 100% accurate and it can't be too fuzzy.

I can easily run in a thread and have the result appear dynamically like covers and additional info does.

[kovidgoyal]

Sounds like you'll need to use find duplicates in that case. Since find duplicates is supposed to be merged into the calibre code base anyway, I'd suggest you go ahead and do that (if it is ok with kiwidude).

[kiwidude]

I had replied to user_none with a PM while I was on holiday - if he or someone else is willing to do the work to merge the Find Duplicates plugin then please go for it. It isn't a high priority to do so for me given it is stable as is. As my available Calibre dev time is going to continue to be significantly reduced for at least the next little while I would rather spend the little I have doing any fixes for my other plugins...

[Starson17]

Originally Posted by user_none

Fuzzy enough that these match:

Toll the Hounds (Malazan Book of the Fallen Series #8) by Steven Erikson
Toll the hounds by Steven Erikson

and

The Hobbit by J RR Tolkien
Hobbit by J. R. R. Tolkien

I'll look at find_identical_books() in db2 and see if it matches enough tests. I know it can't be 100% accurate and it can't be too fuzzy.

I can easily run in a thread and have the result appear dynamically like covers and additional info does.

find_identical_books() is mine (from AutoMerge and Copy To Library), and predates kiwidude's Find Duplicates plugin by a lot. It won't match either of those. It will fail on the first due to the content of the parenthetical in the title. The case differences wouldn't be a problem. It will fail on the second due to the differences in the author name. Periods and other punctuation gets stripped from titles, but not author names. The "The" would get stripped for those who haven't Tweaked their indefinite articles to another language.

It looks like merging kiwidude's code is your best option to have those match.

As you do so, think about whether kiwidude's code could/should be combined with find_identical_books. Perhaps the user should get an option to control how aggressive the automated duplicate finding should be for AutoMerge, Copy To Library and Get Books?