Suggested Performance Improvement - choose markdown engine once (DRAFT)

[slodge]

Prework

• Read and agree to the code of conduct and contributing guidelines.
• If there is already a relevant issue, whether open or closed, comment on the existing thread instead of posting a new issue.

Description

We've got one app in test where we were seeing very slow interactive gt rendering times.

I believe the majority of this slowness came from markdown cell rendering - and when I've tried simpler reproduction projects I've generated some profvis outputs that demonstrated this:

However... in my real project, profvis is pointing to utils::packageVersion as being a significant performance issue - e.g. see:

... but I haven't yet gotten similar results in my test repro projects (yet). I'm not sure why this is - but I'll see if I can work out why "soon" - e.g. perhaps something to do with renv (used in my real project but not in my test repro one)

I'm pausing investigating this for a moment now - but thought I'd open this draft issue/PR now - is this change something gt would be interested in?

Reproducible example

• Post a minimal reproducible example so the maintainer can troubleshoot the problems you identify. A reproducible example is:
  • Runnable: post enough R code and data so any onlooker can create the error on their own computer.
  • Minimal: reduce runtime wherever possible and remove complicated details that are irrelevant to the issue at hand.
  • Readable: format your code according to the tidyverse style guide.

Expected result

• Todo - once I get a nice repro

Session info

• Todo - once I get a nice repro

[rich-iannone]

Many have noted that interactive tables (tables using {reactable} as the backend) are substantially slower than static tables. Do you find that standard interactive tables (i.e., no formatting or styling) are also a bit slow? If so, this whole thing should probably be given much more time and attention.

[slodge]

Many have noted that interactive tables (tables using {reactable} as the backend) are substantially slower than static tables. Do you find that standard interactive tables (i.e., no formatting or styling) are also a bit slow? If so, this whole thing should probably be given much more time and attention.

The tables I've been looking at this work were slow in non-interactive and really slow in interactive.

As one of those perf traces shows, some of this slowness was in utils::packageVersion calls.

Beyond that, a lot of slowness was in some map-based engineering where our authors had written md/html/passthrough content using md or using htmltools tags - e.g. the sort of code suggested in https://community.rstudio.com/t/create-interactive-links-in-gt-table-in-rmarkdown/70266/2

Some of the markdown columns were particularly slow... let me rejig my test script....

[slodge]

Discussion moved to #1528

fmt_markdown	md

(Just leaving these images here as I'm not sure how github cleans things up)