Option to skip current_repodata.json

[dholth]

Checklist

• I added a descriptive title
• I searched open requests and couldn't find a duplicate

What is the idea?

current_repodata is slow, requires conda and isn't used by new solvers. Add option to skip this feature.

It will need to be refactored into a separate function.

Why is this needed?

This is the slowest part of conda-index, and only conda-with-classic-solver uses this file. If we separate this feature then we can even use conda-index without depending on conda and its version parsing / dependency resolution features.