Fuse df columns

[mariusbarth]

Added new function fuse_df() to move degrees of freedom in an apa_results_table from column to variable label of statistic column.

• I chose the function name fuse_df() instead of fuse_df_columns() because I think "columns" is a tricky word and the pluralization might also lead to some confusion. However, I don't have a strong opinion on this choice.
• I also updated the glm, glht, and emmeans methods. I could retire apa_glm_res()` along the way. I believe the complete lm/glm code could be refactored easily using the canonize/beautify workflow.
• Some checks are failing (probably because of upstream changes in the effectsize package), but this is probably best addressed on the development or a separate branch.

We could add fuse_df() to the examples section of the

• [x] lm/glm,
• [x] glht, and
• [x] emmeans methods

after deciding on the function name...

[mariusbarth]

I just addressed virtually all the changes that you suggested in your review. What is left is 1) deciding on the function name, 2) adding function to the examples sections (where it makes sense), and 3) using is_glm in arrange_regression(). This is because, from reading the code, it was my impression that arrange_regression() will become superfluous as soon as we refactor the lm/glm method.

[mariusbarth]

transmute_df()?

https://www.merriam-webster.com/dictionary/transmute

or df_to_label()?

[crsh]

Excellent, thank you!