[Metadata Source Plugin] Goodreads

[ownedbycats]

I usually run Extract ISBN plugin before doing the metadata download. That seems to help get the right edition.

[kiwidude]

Originally Posted by ownedbycats

When downloading a book that combined entries from Amazon and Goodreads, the blank GR description overrode the Amazon one. I posted about it here (more details) and Kovid said this:

Is this what is happening here?

FYI I have posted back on that thread. However as Kovid rightly pointed out there is not actual merging of results going on at least in the example I tried to replicate.

[kiwidude]

Release Notes
https://github.com/kiwidude68/calibr...odreads-v1.7.2

Not convinced there ever was an issue to fix at this point but this code should ensure that if similar comes up again it definitely is not being caused by the Goodreads plugin handling of empty comments fields.

[Rellwood]

Is there any chance of putting in the option to deselect the title/author/series in the plugin for filling in the metadata field rather than having to deselect them for all?

I want to keep Amazon's title and Authors (Amazon authors don't space between periods) but keep Goodread's series since Amazon tacks "The" and "Series" at the beginning and ends.

The plugin may use the title/author/id for searching but whether or not it fills the field should be an option.

[kiwidude]

Originally Posted by Rellwood

Is there any chance of putting in the option to deselect the title/author/series in the plugin for filling in the metadata field rather than having to deselect them for all?

I want to keep Amazon's title and Authors (Amazon authors don't space between periods) but keep Goodread's series since Amazon tacks "The" and "Series" at the beginning and ends.

The plugin may use the title/author/id for searching but whether or not it fills the field should be an option.

Unless Im missing something you can already do that just like you can for any other metadata plugin. Just select the Goodreads plugin and change its configuration options.

[DNSB]

Originally Posted by kiwidude

Unless Im missing something you can already do that just like you can for any other metadata plugin. Just select the Goodreads plugin and change its configuration options.

Looking at Rellwood's post, it seems that automatically combining the output of the two metadata downloads is what they want to do without having to do two metadata download passes.

[kiwidude]

Originally Posted by DNSB

Looking at Rellwood's post, it seems that automatically combining the output of the two metadata downloads is what they want to do without having to do two metadata download passes.

I thought that is just down to the limitations of how calibre matches books - and unless the titles match exactly you are SOL. Or are they saying that setting the title/author to null means that calibre will "automatically" merge the results it finds? Even if you could that sounds way too dangerous/misleading for all the times that the goodreads search finds the wrong book and you wouldn't know it because calibre would present the results of the Amazon title/author...

[Rellwood]

The goodreads and Amazon plugins themselves don't allow you to deselect the title/author/series.

If I could go into the Amazon plugin and select title/author and the Goodreads plugin to only select the series.

The main edit metadata field that mass selects the fields can be filled in but the plugins themselves could be edited to download only the fields i want from each one. This would allow all the plugins to get the fields but only selected fields will be chosen from each plugin (i hope this makes sense).

They can fetch the data but I only want specific fields filled in by specific plugins.

[kiwidude]

Originally Posted by Rellwood

The goodreads and Amazon plugins themselves don't allow you to deselect the title/author/series.

If I could go into the Amazon plugin and select title/author and the Goodreads plugin to only select the series.

The main edit metadata field that mass selects the fields can be filled in but the plugins themselves could be edited to download only the fields i want from each one. This would allow all the plugins to get the fields but only selected fields will be chosen from each plugin (i hope this makes sense).

They can fetch the data but I only want specific fields filled in by specific plugins.

Ahhh now I understand what you mean. Actually Goodreads does let you control whether the series is downloaded, but not the title/authors. And I'm almost certain (without checking the code) that is for the simple reason that the calibre code requires the plugins to pass title/authors back in order for it's merging algorithm to work (and to present the results to the user).

I could be wrong as not looked at the calibre code but I suspect it would require a significant calibre metadata API change to support what you want. Or it could simply be the case that I could just add title/author to the set of configurable fields and it just works - would have to experiment...

[Rellwood]

Originally Posted by kiwidude

Or it could simply be the case that I could just add title/author to the set of configurable fields and it just works - would have to experiment...

That would be appreciated if possible, but NOT a major big deal to break your mind over.

[Rellwood]

Originally Posted by kiwidude

I thought that is just down to the limitations of how calibre matches books - and unless the titles match exactly you are SOL. Or are they saying that setting the title/author to null means that calibre will "automatically" merge the results it finds? Even if you could that sounds way too dangerous/misleading for all the times that the goodreads search finds the wrong book and you wouldn't know it because calibre would present the results of the Amazon title/author...

It's only dangerous if you don't already know what book you are updating the metadata for. Honestly, that isn't going to be the case. Everyone knows what the title/author of a book is before they update the metadata and if they don't they can at least find some kind of identifier within it to help the plugin. The concern I am having is the fact that Amazon and Goodreads are the main sources for me and their metadata is in conflict with each other and how I like my library.

Also, Calibre will use title/author/isbn(if available) no matter what fields you choose to fill in. Even if you deselect every available field they will still come back with a book and the id's. The id field will get changed.

[kiwidude]

Originally Posted by Rellwood

It's only dangerous if you don't already know what book you are updating the metadata for. Honestly, that isn't going to be the case. Everyone knows what the title/author of a book is before they update the metadata and if they don't they can at least find some kind of identifier within it to help the plugin.

The problem is the website search engines, if you *don’t* help them out before your metadata download by setting goodreads ids etc. Which I am willing to bet is how the majority of the plugin users work with their books - they rely on a title/author search match to pull back the right metadata. Which can sometimes go wrong due to for instance that specific book not yet being on the goodreads website and it presents it’s nearest match instead.

It is why I always recommend never using the calibre bulk metadata download option. There are certain times with calibre where you have to eyeball the data you are accepting if there is the possibility it is wrong. The situation I was talking about was if by setting title/author to none prior to where it merges the results of the search and letting it show you as one line item (if it worked that way) - it would create a lot of garbage entries over time for people who don’t prepare the correct goodreads id first.

[ownedbycats]

Originally Posted by kiwidude

The problem is the website search engines, if you *don’t* help them out before your metadata download by setting goodreads ids etc. Which I am willing to bet is how the majority of the plugin users work with their books - they rely on a title/author search match to pull back the right metadata. Which can sometimes go wrong due to for instance that specific book not yet being on the goodreads website and it presents it’s nearest match instead.

Does this indicate that the GR plugin doesn't use ISBNs? I thought it did.

[kiwidude]

Originally Posted by ownedbycats

Does this indicate that the GR plugin doesn't use ISBNs? I thought it did.

It will only use ISBN if your book has one in the metadata already - which in most cases people won't have, unless it either (a) was embedded in the book metadata already or (b) they prepopulated it either manually or using Extract ISBN.

The logic of the plugin is to:
- Use goodreads id if it has one
- If it doesn't, try to find the goodreads id based on either ISBN or ASIN identifiers if it has them
- If no goodreads id found by any of the above means, fall back to a title/author search.

[kiwidude]

Release Notes:
https://github.com/kiwidude68/calibr...odreads-v1.7.3

Thanks to MartinCa and busches for the fixes and PRs, much appreciated!