Author Initials and Goodreads Metadata

[davidfor]

Originally Posted by Rellwood

I only wish GR would follow suit and remove the spaces in their author names because import list and gr sync have problems with matching books when their initials have no spaces.

As stated above, the issue isn't Goodreads, but calibre. It passes the returned metadata through a method to clean it. This includes putting a space after any dots in the name. If you don't like the behaviour, you can run from source and make a change on ebooks/metadata/sources/base.py to the method cap_author_token. It's easy to fix the case of adding the space, but harder to remove the extra space if it is already there. Otherwise, I suggest raising a change request and seeing if Kovid or someone else is interested in adding an option or tweak to change this behaviour.

[kovidgoyal]

@davidfor: Individual metadata plugins can override that behavior if they want to, by simply implementing the clean_downloaded_metadata() method themselves.

[davidfor]

True, but that means implementing the full function. The code to fix the authors is elsewhere and it isn't as easy to only override the cleaning of some parts of the metadata.

But, to my mind, this isn't just about Goodreads, it's about consistent metadata download from all sources. Rellwood is picking on Goodreads either because that's the only source being used, or because they hadn't noticed the others did the same thing. If someone has decided they want no spaces between initials, they will want it like that for all books. And as calibre attempts to merge the metadata from multiple source, if the results from one source plugin is inconsistent with the others, will this still happen?

I don't know the full answer. But, I don't like the idea of implementing something like this in each metadata source plugin. They do need to do something, as they do need to produce metadata that calibre can use. But, I tend to think delivering the metadata to common calibre routines to do the cleaning is better. Of course, that means that calibre needs to have the functions and that they are configurable enough to satisfy a reasonable portion of users.

Of course, that needs someone to write the code. I'm actually happy with how it is working at the moment, so it is unlikely to be me

[kovidgoyal]

You dont actually need to re-implement. You simply call the base class method, then further process the authors as needed.

As to the larger point about having consistency, I agree, and I too am happy with the way it works now