[Feature Request]: CATE for instrument variable based models

### Describe the feature you want to propose or implement

Thanks for the great work! I am just using this package for heterogeneous treatment effect using instrument variables (DoubleMLPLIV). For example, estimate price elasticity for different categories using a supplier-side instrument.

But it seems that this repo only provides this function for non-iv models like DoubleMLPLR.

### Propose a possible solution or implementation

I came up with several ideas to modify the data to achieve CATE for each category, like

1) split data into subsets and estimate the ATE for each.

2) Interact the treatment/instrument with category dummy and estimate the effect of constructed treatments using constructed instruments.

### Did you consider alternatives to the proposed solution. If yes, please describe

_No response_

### Comments, context or references

But I think these are not neat solution to this problem. Do you think these features have a chance to be implemented in the future? Thanks again!

The issue names DoubleMLPLIV and DoubleMLPLR but no files or tests. Start by locating the existing CATE support for non-IV models and the DoubleMLPLIV implementation, then determine the intended API and statistical behavior for heterogeneous effects with instruments. Done should include an agreed implementation scope and coverage for the supported CATE use cases.

Written by the indexing model from the issue text.