Latest restricted WACUP beta release is build #18980 (April 24th 2024) (x86 & x64 changelogs) | Latest WACUP public preview is build #18980 (April 24th 2024) (x86 only)


NOTE: Beta testers are added in a limited & subjective manner as I can only support so many people as part of the beta test program to keep it useful for my needs.

Unless I think you're going to be helpful, not all requests will be accepted but might still be later on. Remember that beta testing is to help me & the limitations currently works for my needs for this project.

Author Topic: Compatibility with additional/custom tags on Smart Views  (Read 14165 times)

ssechaud

  • Beta Tester
  • Jr. Member
  • **
  • Posts: 8
  • Human
    • View Profile
Re: Compatibility with additional/custom tags on Smart Views
« Reply #15 on: November 23, 2018, 06:41:27 PM »
I really like the idea of letting the user specify which tags the database should index. This would let people who want to, index the tags they want with the accompanied performance hit it would incur while not forcing everyone to receive it.

Der_Rio

  • Beta Tester
  • Hero Member
  • *****
  • Posts: 113
    • View Profile
Re: Compatibility with additional/custom tags on Smart Views
« Reply #16 on: November 23, 2018, 07:24:21 PM »
I have put some thought into the new ml_ll plug-in with how we might be able to define & use custom tags but it's then how to store those values that's the problem

As for the DB have a look at the Entity-Attribute-Value model.
In detail: https://en.wikipedia.org/wiki/Entity%E2%80%93attribute%E2%80%93value_model

As graph: https://www.researchgate.net/figure/The-Main-Concept-of-the-Entity-Attribute-Value-Model-The-entity-attribute-value-serves_fig2_257884193

Rio :)

dro

  • Admin / WACUP Developer
  • Administrator
  • Hero Member
  • *****
  • Posts: 4505
    • View Profile
    • WACUP (Winamp Community Update Project)
Re: Compatibility with additional/custom tags on Smart Views
« Reply #17 on: November 23, 2018, 09:05:53 PM »
The storage side issue is more down to how the nde db works as you need to define the column first before you can use it & also the type (which I assume the likes of sqlite also require - I can't remember how it works off hand).

As I had initially thought about allocating a few empty columns and then via preferences they're stored / accessed as needed from those columns against the user config. Or as I think about it now I could just use a fixed naming for the extra columns so it can be increased dynamically which would also work (which a rebuild of the database would help realign the data / remove the dead space).

Though I think there's scope that some values would always need to be dynamically generated vs being generated on import / update so maybe I'm overthinking it vs what's needed 😉

With regards to the prior post, maybe a full set of options on what tags are read / stored in addition to custom ones would be the best solution. That way it can also help determine the columns able to be shown in the views vs the expected defaults.

Iit might need a bit of thinking about how to handle the filter panes as some would need certain values to be present but I need to make them more flexible anyway vs trying to mimick the existing ml_local selection in the current starting point of the ml_ll implementation.

This has given me a number of thoughts on things to try out so thank you both of you 👍

-dro
« Last Edit: November 23, 2018, 09:07:21 PM by dro »