Drop ggpubr dependency

[grantmcdermott]

Closes #210.

This PR simply drops ggpubr and replaces one or two functions with with "native" ggplot2 equivalents.

The plots are slightly different, but I personally doubt that anyone will really notice. More importantly, this small change leads to over 30(!) fewer dependencies.

pkgload::load_all("~/Documents/Projects/did-1/")
#> ℹ Loading did
data(mpdta)

out <- att_gt(
  yname = "lemp",
  gname = "first.treat",
  idname = "countyreal",
  tname = "year",
  xformla = ~1,
  data = mpdta,
  est_method = "reg"
)
ggdid(out, ylim = c(-.25, .1))

es <- aggte(out, type = "dynamic")
ggdid(es)

^{Created on 2024-11-04 with reprex v2.1.1}

PS. If you want to continue down this road, then there are more places for trimming. E.g. If you already have data.table, then you probably get away with importing dplyr and tidyr. But I'll leave that for you to signal interest.

[pedrohcgs]

I will check this in a bit. But @grantmcdermott, you know the automatic tests are not passing.... =P

[grantmcdermott]

I will check this in a bit. But @grantmcdermott, you know the automatic tests are not passing.... =P

Well, the R CMD tests are all passing. It's the mysterious "R-Package-Test" workflow that's failing...

More seriously, I think it's just a minor undeclared dependency issue. We just need to add broom (or even backports) to Suggests. Let me try that and then retrigger the test suite.

[grantmcdermott]

Ugh, looks like I made things worse. Serves me right for doing this on my phone before getting the kids to school. I'll take a proper look later and ping you when it's ready.

[pedrohcgs]

Sounds good!!!

[pedrohcgs]

I think it may be the data.table because I merged a PR that updated the code and leverages the data.table now.

[pedrohcgs]

To be clear, the tests are also passing locally here

[pedrohcgs]

We are good! It is not passing!