Managed to isolate the issue! Seems you might not be supporting the describe-it processes from testthat.

github.com/posit-dev/po...

thanks, did that! :) But I can't seem to get it to run anything, not terminal not in console. I've scrolled through settings and think I have everything that should be enabled, enabled.

Maybe? But I did have the expectation that positron assistant would be more active in agentic mode than I've been able to get it, and it's told me several times that it cannot execute R code, so I was assuming I needed to do more _stuff_ to get it working

Because mine does not seem to manage that and I thought it was because I had not set it up.

I've looked at `mcptools` and `btw` and they seem like the things to use for #rstats MCP servers, but still, I'm not sure how to integrate into #PositronAssistant

2. I'm trying to set up #PositronAssistant to use R as an #MCPserver, and seem to not quite understand how that is done. The "MCP Servers" button in the assistants take me to the Extensions tab, and I don't see any place to set up config etc for the MCP servers for Positron (as I would for Continue)

1. I have a test-suite that currently is not being run properly during testing. I run the tests, it seemingly does something, but there is no output and no tests get marked as failed or successful. It's not skipping, at least it should not. Anyone experienced something like that before?

Cool cool cool 😊 welcome to my sporadic posting 🤣

I'm interested in knowing what prompted you to follow me? Genuine question 😊

It really is such a very powerful and versatile approach, but thinking in so many dimensions is a struggle!

Had a lot of fun figuring that one out 😊

Imagemagick is such an amazing system library, and having it easily callable from #rstats is so convenient!

I don't think I will ever understand it, and I'm kind of happy I don't need to. I just need to get the pieces working.

And thank you to @hadley.nz and Posit. That's some solid piece of stuff you've made, and makes my OAuth and curl headaches less than what they would have been.

httr2.r-lib.org

Could not have don't this without dealing with the nettskjema API, and OAuth.

But OAuth has sooo many ways of working, and every API sets their own constraints and systems on how to authenticate. So every new API I try to work against, is learning a whole new bit of OAuth I knew nothing about...

I work poorly with deadlines, stress shuts me down faster than ever before and my brain just becomes cotton wool. However, having this pipeline set up means I still contribute within my own parameters, and I am no bottleneck.

A daily digest ensures my colleagues always have access to plots, tables and summaries of the responses, so the grant writing goes smoothly also in my absence.

For instance, as I am only working 10% but work deadlines still exist. I made a pipeline using my own nettskjemar package to create a daily digest of the responses from our survey on the current #nfr infrastructure grant proposal we are working on.

www.capro.dev/nettskjemar/