[FEA]: ST_distance improvement: consolidate existing APIs

[isVoid]

Is this a new feature, an improvement, or a change to existing functionality?

Improvement

How would you describe the priority of this feature request

Medium

Please provide a clear description of problem you would like to solve.

There are dependencies between st_distance kernels, as shown in the following diagram:

The current implementation of st_distance kernels are siloed and lackes reusing overlapping logics. A refactor should happen to facilitate these dependencies.

Second, several implementation in kernels are not load balanced if the input is skewed. This can lead to unwanted performance degradation. As experimented in #1011, it is possible to create a perfectly load balanced kernel for any input geometries. It should be adopted for other distance API too.

Describe any alternatives you have considered

No response

Additional context

No response

Tasks

Preview Give feedback

1. Refactor st_distance kernels to increase modularity #1060
   +0
   
2. Refactor ST_Distance column API to ensure API homogeneity. #1123
   +0
   

[isVoid]

Originally, #1061 is considered part of this milestone issue. However, #1061 is mainly about optimization on existing distance kernel performance, while this issue requests architecture wise improvement. I decided that they are unrelated issues. Thus removing 1061 from the list above.

Since all underlying issue for this milestone is closed, closing this milestone.