addins for boilerplate and merge refactor groups + {poof} tricks

[moodymudskipper]

Addins would replace code by %refactor_*% calls with rm()boilerplate if relevant.
We should also have an addin to merge refactored code by call type, non refactored code would be integrated on both sides if it's between calls of same type.

This is a good use case for {poof} action groups too, they should be built on top of addins.

  {
    # MoM change
    series_vol_1_mom <- smooth_series(series_vol_1, 1)
    series_vol_2_mom <- smooth_series(series_vol_2, 1)
  } %refactor_chunk% {
    series_vol_1_mom <- map_dfc(series_vol_1, ~ . / lag(.) - 1)
    series_vol_2_mom <- map_dfc(series_vol_2, ~ . / lag(.) - 1)
    # ~ for {refactor}:
    series_vol_1_mom <- as.data.frame(series_vol_1_mom)
    series_vol_2_mom <- as.data.frame(series_vol_2_mom)
  }

  {
    series_vol_1_cum <- moving_sum(series_vol_1_mom, 12, 0)
    series_vol_2_cum <- moving_sum(series_vol_2_mom, 12, 0)
  } %refactor_chunk% {
    series_vol_1_cum <- map_dfc(series_vol_1_mom, roll::roll_sum, 12)
    series_vol_2_cum <- map_dfc(series_vol_2_mom, roll::roll_sum, 12)
    # ~ for {refactor}:
    series_vol_1_cum <- as.data.frame(series_vol_1_cum)
    series_vol_2_cum <- as.data.frame(series_vol_2_cum)
  }

==>

  {
    # MoM change
    series_vol_1_mom <- smooth_series(series_vol_1, 1)
    series_vol_2_mom <- smooth_series(series_vol_2, 1)

    series_vol_1_cum <- moving_sum(series_vol_1_mom, 12, 0)
    series_vol_2_cum <- moving_sum(series_vol_2_mom, 12, 0)
  } %refactor_chunk% {
    series_vol_1_mom <- map_dfc(series_vol_1, ~ . / lag(.) - 1)
    series_vol_2_mom <- map_dfc(series_vol_2, ~ . / lag(.) - 1)

    series_vol_1_cum <- map_dfc(series_vol_1_mom, roll::roll_sum, 12)
    series_vol_2_cum <- map_dfc(series_vol_2_mom, roll::roll_sum, 12)
    # ~ for {refactor}:
    series_vol_1_mom <- as.data.frame(series_vol_1_mom)
    series_vol_2_mom <- as.data.frame(series_vol_2_mom)

    series_vol_1_cum <- as.data.frame(series_vol_1_cum)
    series_vol_2_cum <- as.data.frame(series_vol_2_cum)
  }