Cookiecutter-cmdstanpy: a template for cmdstanpy projects

[ahartikainen]

@Funko_Unko

• Be able to work with fit objects where only some of the chains finished (see above).

ArviZ might work in this case (tbh, there could be errors too)

@Funko_Unko

• Get per chain timings. Did I miss some functionality?

I think these can be accessed from the stdout files?

@Funko_Unko

• Collapse the stan summary to a boolean (i.e. everything alright or not).

Maybe ArviZ summary might be easier to work with?

@Funko_Unko

• Sometimes the ordering of arrays of vectors didn’t seem to be right. But I’m not sure anymore.

Was this after you combined multiple chains? See arviz.InferenceData output, does that look correct?

@Funko_Unko

• Randomly pick a draw to initialize a new run.

I think a simple function could work in this case? (It might be easier to get inits from arviz.InferenceData)

@Funko_Unko

• Inspect the source code and get functions / model blocks.

(source → cpp?)

@Funko_Unko

• Restart sampling from a fit object.

Was there any problems?

@Funko_Unko

• Addendum: I wanted the model to be able to infer some of the data from other parts of the data, like for example the length of an array.

In Stan code?

@Funko_Unko

• Also: being able to sample from the prior.

? (fixed_param + a custom .stan is probably easiest way to do this)

@Funko_Unko

• Also, the sample function does not seem to play nice with numpy s ndarray for the stepsize, and I think I could not simply pass multiple metric s or inits

I think I already fixed these issues. But if this is not yet working, then we need to fix it.