Return variable names and dimensions

[rok-cesnovar]

Submission Checklist

• Run unit tests
• Declare copyright holder and agree to license (see below)

Summary

Attempt at fixing #240

For now, this just returns the variable names and dimensions as part of metadata in read_cmdstan_csv. We should add it in fit, but not sure on the name.

Example:

> f <- cmdstanr_example("logistic", method = "sample")
> csv <- read_cmdstan_csv(f$output_files())
> f1$metadata$model_params_dims
$lp__
[1] 0

$alpha
[1] 0

$beta
[1] 3

> f1$metadata$model_params_named
[1] "lp__"  "alpha" "beta"

Copyright and Licensing

Please list the copyright holder for the work you are submitting
(this will be you or your assignee, such as a university or company):
Rok Češnovar, Univ. of Ljubljana

By submitting this pull request, the copyright holder is agreeing to
license the submitted work under the following licenses:

• Code: BSD 3-clause (https://opensource.org/licenses/BSD-3-Clause)

[rok-cesnovar]

@jgabry we can hold off with this for after beta if you want. I am done for today, but if you have any feedback I can finish this tomorrow.

Mostly want ideas for names. Does metadata$model_params_dims and metadata$model_params_named work?
Would fit$draws_dims() work? Or should this be a function on draws rather than a part of fit?

[codecov-commenter]

Codecov Report

Merging #256 into master will increase coverage by 0.07%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #256      +/-   ##
==========================================
+ Coverage   90.78%   90.85%   +0.07%     
==========================================
  Files          11       11              
  Lines        2311     2329      +18     
==========================================
+ Hits         2098     2116      +18     
  Misses        213      213              

Impacted Files	Coverage Δ
R/read_csv.R	96.28% <100.00%> (+0.04%)	⬆️
R/utils.R	70.85% <100.00%> (+2.53%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update fb3703f...4dd2c18. Read the comment docs.

[jgabry]

Ok cool thanks I'll take a look.

[mitzimorris]

Mostly want ideas for names. Does metadata$model_params_dims and metadata$model_params_named work?
Would fit$draws_dims() work? Or should this be a function on draws rather than a part of fit?

hi, could we co-ordinate w/ CmdStanPy on names? we went through this a while back - last thing I implemented there was stan-dev/cmdstanpy#252 - called everything "stan_variable".

[jgabry]

Yes good point @mitzimorris!

[jgabry]

@mitzimorris Does stan_var_dims() in CmdStanPy say that a variable

real alpha

has dim 0 or 1?

[jgabry]

@rok-cesnovar I pushed a few edits and I think this is ok now other than changing the names. Are you ok with changing to "stan_variable", which is what @mitzimorris said CmdStanPy uses?

[mitzimorris]

@mitzimorris Does stan_var_dims() in CmdStanPy say that a variable

real alpha

has dim 0 or 1?

property stan_var_dims returns a Python dict which maps
Stan program variable names to variable dimensions.
Scalar types have int value '1'. Structured types have list of dims,
e.g., program variable vector[10] foo has entry ('foo', [10]).

given fit object bernoulli_fit, stan_var_dims returns {'theta': 1}
given fit object logistic_fit, stan_var_dims returns {'beta': (2,)}

ps. thanks for asking - noticed that doc doesn't match output. damn.

[rok-cesnovar]

@jgabry changed to stan_var_dims and stan_variables. Also added a few more tests. I think its good to go.

[jgabry]

Looks good! This is good to go, just one super nitpicky thing (sorry!): I think it's slightly weird to use stan_variables but then use stan_var_dims instead of stan_variable_dims. I would prefer using either stan_vars and stan_var_dims or stan_variables and stan_variable_dims. I think I prefer the latter, but I'm fine either way, as long as it's consistent. @mitzimorris and @rok-cesnovar what do you think?

[rok-cesnovar]

I would prefer using either stan_vars and stan_var_dims or stan_variables and stan_variable_dims.

I agree. That annoyed me a bit as well to be honest but thought that was discussed and agreed on earlier. I prefer the verbose version as well.

[mitzimorris]

agreed. will change CmdStanPy accordingly.

[rok-cesnovar]

@jgabry renamed. Good to merge?

[jgabry]

Yeah looks good. Will merge now.