estimate_join fails with uninformative error

[sangeetabhatia03]

incid <- array(c(20, 49, 66, 71, 88, 97, 110, 139, 160, 229, 1, 0, 0, 1, 0, 0, 0, 0, 1, 1), dim = c(10, 1, 2))
si_for_est <- structure(c(0, 0.0109271014298609, 0.0642438529179909, 0.106855472976164,
0.12306728253113, 0.121829045299574, 0.111040351371263, 0.0960718845838475,
0.080195134014859, 0.0652126770986732, 0.0519814334659148, 0.040788378193821,
0.0316014204991031, 0.0242283956789802, 0.0184129440362466, 0.013889008185358,
0.0104092453079867, 0.00775765059724476, 0.00575308532856567,
0.00424792304600063, 0.00312437608196081, 0.00228999572316568,
0.0016731661617051, 0.00121900103725681, 0.000885809191340847,
0.000642160737961417, 0.000464514524466608, 0.000335336927699048,
0.000241632912065137, 0.000173813009963775, 0.000124828171693681,
0, 0.0000103800103182646, 0.000768325053224188, 0.00670878308447795,
0.0230631536967196, 0.0484355618204519, 0.0757802097883441, 0.0977343683866287,
0.110045750786206, 0.112074009114762, 0.105667913970107, 0.09372696086706,
0.0791193104597284, 0.064108758150231, 0.0501885081793932, 0.0381550918711331,
0.0282824135490702, 0.0205074832227862, 0.0145847717721584, 0.0101962247269376,
0.00702000062782433, 0.0047673233358495, 0.00319767110173943,
0.00212087729484105, 0.00139236766621921, 0.000905581626701108,
0.000583940817584487, 0.000373568962290795, 0.000237243004788178,
0.00014964709866497, 0.0000937999537589569), .Dim = c(31L, 2L
), .Dimnames = list(NULL, c("si_distr_ref", "si_distr_variant"
)))

EpiEstim::estimate_joint(
incid, si_for_est, priors = default_priors(), seed = 1,
t_min = 2L, t_max = as.integer(10),
mcmc_control = EpiEstim::default_mcmc_controls()
)

Error in if (any(epsilon < 0)) { : missing value where TRUE/FALSE needed

traceback()
2: draw_R(epsilon_out[, i], incid$local, lambda, priors, t_min = t_min,
t_max = t_max) at gibbs_draws.R#481
1: EpiEstim::estimate_joint(incid, si_for_est, priors = default_priors(),
seed = 1, t_min = 2L, t_max = as.integer(10), mcmc_control = EpiEstim::default_mcmc_controls())

[sangeetabhatia03]

I think the cause is that in this example, R_init is c(4.387277 , NA) and therefore epsilon_init is NA.