search

anansi-project / comicinfo

ComicInfo.xml's new home
https://anansi-project.github.io/docs/category/comicinfo
MIT License
136 stars 8 forks source link

Added SeriesSort tag #16

Open majora2007 opened 2 years ago

majora2007 commented 2 years ago

Implements support for #4 by adding a new tag, SeriesSort, which allows users to specify sort order for series with ComicInfo.

majora2007 commented 2 years ago

Any chance to merge this? Kavita added support for this a bit ago.

gotson commented 2 years ago

@lordwelch @ajslater WDYT?

lordwelch commented 2 years ago

Looks good to me, coincidentally I just had a discussion on why this tag is necessary if not useful for most types of media, in ComicTagger I won't have support for a while, I want to get some other changes merged first

ajslater commented 2 years ago

Most sorting with English seems easily done semantically. For instance, the examples given in #4 are easy to do algorithmically. In Codex If i were to solve for the example cases, I'd make a series.sort_name column and remove two to four common english articles from the first token of series.name. I imagine that would enable more consistent sorting than scattershot tags.

I wonder if there are more useful non-english use cases.

That said, if you think it's useful, sure, why not. lordwelch's attribute idea is technically where i might go, but I see the wisdom in gotson's consistency argument.

gotson commented 2 years ago

Most sorting with English seems easily done semantically. For instance, the examples given in #4 are easy to do algorithmically. In Codex If i were to solve for the example cases, I'd make a series.sort_name column and remove two to four common english articles from the first token of series.name.

100% agree, but that works only if you are managing english content only, or if you can detect that the content is english language (the language tag could help). But even so, you can have japanese manga with english titles, for which you may want to also handle sorting.

I imagine that would enable more consistent sorting than scattershot tags.

The idea is more to decouple the processing from the metadata. If an app wants to perform it algorithmically, fine. But if you want to fill it yourself, and expect the apps to handle it if present, that's another more flexible solution.

That said, if you think it's useful, sure, why not. lordwelch's attribute idea is technically where i might go, but I see the wisdom in gotson's consistency argument.

Let's keep the 💩 consistent 🤣