SMC Mallows fails with a single particle

[osorensen]

Maybe not a big deal, but it should be fixed or an informative error message should be given.

library(BayesMallows)
set.seed(1)
# UPDATING A MALLOWS MODEL WITH NEW COMPLETE RANKINGS
# Assume we first only observe the first four rankings in the potato_visual
# dataset
data_first_batch <- potato_visual[1:4, ]

# We start by fitting a model using Metropolis-Hastings
mod_init <- compute_mallows(
  data = setup_rank_data(data_first_batch),
  compute_options = set_compute_options(nmc = 10000))

# Convergence seems good after no more than 2000 iterations
assess_convergence(mod_init)

burnin(mod_init) <- 2000

# Next, assume we receive four more observations
data_second_batch <- potato_visual[5:8, ]

# We can now update the model using sequential Monte Carlo
mod_second <- update_mallows(
  model = mod_init,
  new_data = setup_rank_data(rankings = data_second_batch),
  smc_options = set_smc_options(n_particles = 1, resampler = "systematic")
)
#> Error in eval(expr, envir, enclos): Not a matrix.

^{Created on 2024-03-25 with reprex v2.1.0}

[osorensen]

Closing because incorporating SMC2 methodology may be a better way to move forward.