Show more precise debugging steps

[etiennebacher]

This is related to #283. It is not at all complete, I just did a small test on the "Assumption checks" part to see how to customize the error message to give more precise steps. I implemented an error manually (log("a")) just to simulate an error.

foo <- lm(mpg ~ drat, mtcars)
model_dashboard(foo, check_model_args = list(show_dots = FALSE))

Here's the current output:

And here's the new type of error messages:

@easystats/core-team What do you think? Should it be generalized to the other parts of the dashboard?

[codecov-commenter]

Codecov Report

Merging #286 (71f165c) into main (f70f38e) will increase coverage by 0.41%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main     #286      +/-   ##
==========================================
+ Coverage   22.91%   23.32%   +0.41%     
==========================================
  Files           9        9              
  Lines         371      373       +2     
==========================================
+ Hits           85       87       +2     
  Misses        286      286              

Impacted Files	Coverage Δ
R/model_dashboard.R	100.00% <100.00%> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[strengejacke]

Nice, that's a great idea!

[IndrajeetPatil]

Love it!

Having a precise error message will be really helpful, both for the users and for us.

[bwiernik]

For the last part, can we also direct them to include the function call and not only the error text (is that accessible from the error object we are printing?)? I don't know that just getting non-numeric argument to mathematical function without the context would be useful