# \[Interface roadmap\] fit objects and \`extract\`

**URL:** <https://discourse.mc-stan.org/t/interface-roadmap-fit-objects-and-extract/10126>\
**Category:** Developers\
**Created:** [August 9, 2019, 12:57pm UTC](https://discourse.mc-stan.org/t/interface-roadmap-fit-objects-and-extract/10126 "2019-08-09T12:57:43Z")\
**Posts on this page:** 1\
**Showing post:** 19

<div class="post-metadata">

**Author:** ![andrewgelman](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.mc-stan.org/andrewgelman/32/1577_2.png) [@andrewgelman](https://discourse.mc-stan.org/u/andrewgelman)\
**Post date:** [August 20, 2019, 9:43pm UTC](https://discourse.mc-stan.org/t/interface-roadmap-fit-objects-and-extract/10126/19 "2019-08-20T21:43:04Z")

</div>

I’d like to have fit$theta because this comes up a lot! I’d also like to do something like foo(fit$theta, median) or foo(fit, “theta”, median) that would grab the median over the simulation draws (or it could be foo(fit, “theta”, mean) or foo(fit, “theta”, quantile, c(0.25, 0.75)) etc. Currently I waste a lot of time in R grabbing simulations, then trying to figure out how to use “apply” or whatever to get summaries. This is a particular issue if theta is a vector or matrix or array. What I’d generally like is pointwise summaries; e.g., if theta is a 2 x 3 x 4 array, then fit$theta would return a n\_sims x 2 x 3 x 4 array, and foo(fit, “theta”, median) would return a 2 x 3 x 4 array.

I’m also thinking the default would be to concatenate all chains together, but of course there should also be an option to grab one chain at a time, which could come up for diagnostics or whatever.

---

_[View the full topic](https://discourse.mc-stan.org/t/interface-roadmap-fit-objects-and-extract/10126)._