Avoiding metadata identify results merge

Ramblurr · 06-04-2014, 03:59 AM

My metadata source plugin is creating identical results that differ only on the publisher field (so title, authors, id, are the same). But the ISBNMerge algorithm in identify.py is merging them all into one result with only the first publisher added getting chosen.

Can I affect this algorithm somehow? Is there some hijinx I can play with my Metadata object to get this algorithm not to perform the merge? It seems changing the title, author, or identifiers field would work, but, naturally, I want these to remain the same, correct, values.

This is because the webservice I'm writing a metadata source plugin for unfortunately conflates the various publishers for all editions of an item.

Example:
See this entry: http://rpggeek.com/rpgitem/44383/trail-of-cthulhu (XML)

The primary publisher is Pelgrane Press, and the two other publishers are responsible for other language translations.

In my metadata source plugin I'm adding three results, one for each publisher. The idea is the user can choose which publisher matches the version they have.

Any advice?

kovidgoyal · 06-05-2014, 12:10 AM

No there's no way a plugin can influence that algorithm. That algorthm exists so that a user with a lot of plugins enabled is not overwhelmed by lots of very similar results. Generally speaking, the metadata download process is not really conducive to allowing a user to pick and choose between slighlty differing individual results as that would be a mess for the most common use cases.

If you really want to allow users to choose individual publishers and the like, your best bet is to to create a user interface plugin that bypasses the normal metadata download process altogether and has its own UI for browsing results. The downside to this approach is that your plugin will not be useable from the edit metadata dialog.

06-04-2014, 03:59 AM	#1
Ramblurr Member Posts: 11 Karma: 10 Join Date: Apr 2010 Device: Kindle	Avoiding metadata identify results merge My metadata source plugin is creating identical results that differ only on the publisher field (so title, authors, id, are the same). But the ISBNMerge algorithm in identify.py is merging them all into one result with only the first publisher added getting chosen. Can I affect this algorithm somehow? Is there some hijinx I can play with my Metadata object to get this algorithm not to perform the merge? It seems changing the title, author, or identifiers field would work, but, naturally, I want these to remain the same, correct, values. This is because the webservice I'm writing a metadata source plugin for unfortunately conflates the various publishers for all editions of an item. Example: See this entry: http://rpggeek.com/rpgitem/44383/trail-of-cthulhu (XML) The primary publisher is Pelgrane Press, and the two other publishers are responsible for other language translations. In my metadata source plugin I'm adding three results, one for each publisher. The idea is the user can choose which publisher matches the version they have. Any advice?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Bulk metadata download and merge tags and select larger cover(s)	hiperlink	Calibre	2	03-26-2014 03:58 AM
Download metadata merge result ?	abuck	Library Management	1	10-11-2013 08:49 AM
Merge books with Edit Metadata bulk option?	Nialls	Calibre	1	02-07-2013 11:35 AM
Merge Formats library into Metadata library? Or, Add Format in bulk?	Sabardeyn	Library Management	5	01-23-2013 06:00 AM
Merge metadata from other lib	drMerry	Library Management	6	05-04-2011 06:41 PM

06-05-2014, 12:10 AM	#2
kovidgoyal creator of calibre Posts: 43,858 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	No there's no way a plugin can influence that algorithm. That algorthm exists so that a user with a lot of plugins enabled is not overwhelmed by lots of very similar results. Generally speaking, the metadata download process is not really conducive to allowing a user to pick and choose between slighlty differing individual results as that would be a mess for the most common use cases. If you really want to allow users to choose individual publishers and the like, your best bet is to to create a user interface plugin that bypasses the normal metadata download process altogether and has its own UI for browsing results. The downside to this approach is that your plugin will not be useable from the edit metadata dialog.

Advert