Add support for `interval_select´ density function type and fix #78

[annhilati]

Description

This PR addresses two issues related to Density Functions to improve parity with Vanilla Minecraft:

1. Fix mul and add neglecting propagation of non-finite values (Fixes [Density Functions] mul and add with 0 neglect propagation of non-finite values #78)
   When parsing mul or add operations where one of the arguments is a constant (like 0), Deepslate previously unconditionally evaluated this as an Ap2 function, which short-circuited and prevented the evaluation of the second argument. This caused discrepancies with Vanilla Minecraft when evaluating non-finite values like Infinity and NaN.
   To match Vanilla's parsing optimizations, this PR introduces a MulOrAdd transformer class. The parser (fromJson) and mapAll methods have been updated to wrap operations involving a constant into this MulOrAdd class, which correctly evaluates the dynamic input first before applying the constant.

2. Add support for interval_select (Fixes [Density Functions] Add support for interval_select #80)
   Added the IntervalSelect class mapping to the interval_select DensityFunction, fully porting the Vanilla Java implementation. This includes accurate compute, mapAll, and min/max implementations.

Changes Made

• Added IntervalSelect class to DensityFunction.ts.
• Added MulOrAdd transformer class to DensityFunction.ts.
• Created a createAp2 helper to safely instantiate MulOrAdd or Ap2 based on whether the inputs are constants.
• Updated DensityFunction.fromJson to parse interval_select and utilize the new createAp2 logic.
• Added tests for interval_select and the 0 * Infinity propagation edge-case in DensityFunction.test.ts.

Note: AI assistance was used to help port the Java implementation and fix the parsing logic, but all changes have been verified against the existing test suite.

[misode]

Thanks for your work on this. I am going to close this and implement this myself, partly because this really should've been split in two PR's. I might take a look at some of your tests though to figure out #78 issue.