Bug in bottom_up()

[GeorgeAthana]

tourism_full <- tsibble::tourism %>% aggregate_key((State/Region)*Purpose, Trips = sum(Trips))
tourism_fit_ets <- tourism_full %>%
filter(year(Quarter)<=2015) %>%
model(
ets=ETS(Trips),
) %>%
reconcile(
bu_ets = bottom_up(ets),
wls_ols = min_trace(ets, method="ols")
)

tourism_fc_ets <- tourism_fit_ets %>%
forecast(h="2 years")

tourism_fc_ets %>%
filter(is_aggregated(State),is_aggregated(Region),!is_aggregated(Purpose)) %>%
autoplot(tourism_full, level=NULL) +
facet_wrap(vars(Purpose), scales = "free_y")

tourism_fc_ets %>%
filter(!is_aggregated(State),is_aggregated(Region),is_aggregated(Purpose)) %>%
autoplot(tourism_full, level=NULL) +
facet_wrap(vars(State), scales = "free_y")

[mitchelloharawild]

Issue is rooted in redundant nodes, fixed in 251fcc2