[new-model] Implement censored shifted Wald Model

[GidonFrischkorn]

I want to add the censored shifted Wald model for modeling reaction times and proportion correct in tasks with few errors. In such task the DDM is typically not well identified as the distribution for error responses has few observations which leads to problems with accurately estimating the drift rate and boundary separation.

The model is described by Miller et al (2018): https://journals.sagepub.com/doi/10.1177/0146621617710465

And there is an existing implementation for STAN by Tom Faulkenberry: https://github.com/tomfaulkenberry/censoredShiftedWald

My goal is to implement the STAN functions with a parametrization in terms of cognitive processes: that ist drift, boundary, and ndt. For this, I should be able to reuse most of the code by Tom Faulkenberry.

If you @venpopov have any suggestions for the implementation feel free to comment, here.

Steps for implementation:

• fill in model_info
• implement STAN functions
• specify check_data.cswald
• specify check_model.cswald
• specify bmf2bf.cswald if necessary
• specify check_formula.cswald
• specify configure_model.cswald
• add function to generate initial values to allow for proper start of sampling
• specify distribution functions for the cswald model
• check that pp_check and bridgesampling works for cswald
• if possible: optimize sampling and speed for model estimation