g3_iterative: `Error in if (is.na(attr(lik_init, "nll"))) { : argument is of length zero`

[MikkoVihtakari]

I get the error in the title when trying to run my model through g3_iterative(). The error stems from this line and, further, from here and results from adfun$report(adfun$par) not returning a list element named nll (see the output below). Has something been changed somewhere or is this issue specific to my model?

adfun <- gadget3::g3_tmb_adfun(tmb_model, tmb_param, type = 'Fun')
res <- adfun$report(adfun$par)
names(adfun$report(adfun$par))

[1] "nll_adist_surveyindices_log_Juv_SI_1__weight"       "adist_surveyindices_log_SI_48_obs__wgt"            
 [3] "adist_surveyindices_log_SI_48_model__wgt"           "nll_cdist_sumofsquares_RussianS_ldist__weight"     
 [5] "nll_cdist_sumofsquares_TrawlNor_ldist__num"         "cdist_sumofsquares_OtherNor_ldist_model__num"      
 [7] "nll_cdist_sumofsquares_RussianS_ldist__num"         "nll_cdist_sumofsquares_OtherNor_ldist__num"        
 [9] "nll_cdist_sumofsquares_EggaS_ldist__weight"         "cdist_sumofsquares_TrawlRus_ldist_obs__num"        
[11] "cdist_sumofsquares_EggaN_aldist_male_obs__num"      "nll_cdist_sumofsquares_EggaN_aldist_male__num"     
[13] "cdist_sumofsquares_RussianS_sexdist_obs__num"       "cdist_sumofsquares_RussianS_ldist_model__num"      
[15] "cdist_sumofsquares_TrawlNor_sexdist_model__num"     "adist_surveyindices_log_Juv_SI_2_model__params"    
[17] "nll_adist_surveyindices_log_SI_28_47__weight"       "cdist_sumofsquares_OtherNor_ldist_obs__num"        
[19] "cdist_sumofsquares_RussianS_ldist_obs__num"         "cdist_sumofsquares_RussianS_sexdist_model__num"    
[21] "nll_cdist_sumofsquares_EggaN_aldist_female__weight" "nll_cdist_sumofsquares_EggaS_matp__weight"         
[23] "nll_cdist_sumofsquares_RussianS_sexdist__num"       "cdist_sumofsquares_EggaN_aldist_female_model__num" 
[25] "nll_cdist_sumofsquares_EggaS_ldist__num"            "cdist_sumofsquares_EcoS_ldist_model__num"          
[27] "nll_adist_surveyindices_log_Juv_SI_2__wgt"          "cdist_sumofsquares_EggaN_ldist_model__num"         
[29] "nll_cdist_sumofsquares_EcoS_ldist__num"             "nll_adist_surveyindices_log_Juv_SI_1__wgt"         
[31] "nll_cdist_sumofsquares_OtherNor_ldist__weight"      "cdist_sumofsquares_TrawlNor_ldist_obs__num"        
[33] "cdist_sumofsquares_WinterS_ldist_obs__num"          "cdist_sumofsquares_EcoS_ldist_obs__num"            
[35] "cdist_sumofsquares_TrawlNor_sexdist_obs__num"       "cdist_sumofsquares_EggaS_ldist_obs__num"           
[37] "nll_cdist_sumofsquares_EggaN_ldist__num"            "nll_cdist_sumofsquares_EcoS_sexdist__weight"       
[39] "nll_adist_surveyindices_log_SI_48__wgt"             "nll_cdist_sumofsquares_TrawlNor_sexdist__weight"   
[41] "nll_cdist_sumofsquares_TrawlRus_sexdist__num"       "nll_cdist_sumofsquares_TrawlNor_sexdist__num"      
[43] "cdist_sumofsquares_EcoS_sexdist_obs__num"           "adist_surveyindices_log_SI_28_47_model__params"    
[45] "adist_surveyindices_log_SI_28_47_obs__wgt"          "nll_cdist_sumofsquares_EggaN_ldist__weight"        
[47] "cdist_sumofsquares_TrawlRus_sexdist_model__num"     "cdist_sumofsquares_OtherNor_sexdist_obs__num"      
[49] "nll_cdist_sumofsquares_OtherNor_sexdist__num"       "cdist_sumofsquares_EggaS_ldist_model__num"         
[51] "cdist_sumofsquares_EggaS_matp_obs__num"             "adist_surveyindices_log_Juv_SI_1_model__params"    
[53] "nll_adist_surveyindices_log_SI_28_47__wgt"          "cdist_sumofsquares_WinterS_ldist_model__num"       
[55] "nll_cdist_sumofsquares_EggaS_matp__num"             "cdist_sumofsquares_TrawlNor_ldist_model__num"      
[57] "cdist_sumofsquares_TrawlRus_ldist_model__num"       "cdist_sumofsquares_OtherNor_sexdist_model__num"    
[59] "nll_understocking__wgt"                             "nll_cdist_sumofsquares_TrawlRus_sexdist__weight"   
[61] "adist_surveyindices_log_SI_28_47_model__wgt"        "nll_cdist_sumofsquares_OtherNor_sexdist__weight"   
[63] "nll_adist_surveyindices_log_Juv_SI_2__weight"       "nll_adist_surveyindices_log_SI_48__weight"         
[65] "nll_cdist_sumofsquares_EggaN_matp__num"             "nll_cdist_sumofsquares_EcoS_sexdist__num"          
[67] "nll_cdist_sumofsquares_EcoS_ldist__weight"          "adist_surveyindices_log_Juv_SI_1_obs__wgt"         
[69] "nll_cdist_sumofsquares_RussianS_sexdist__weight"    "adist_surveyindices_log_SI_48_model__params"       
[71] "nll_cdist_sumofsquares_WinterS_ldist__weight"       "nll_cdist_sumofsquares_TrawlNor_ldist__weight"     
[73] "nll_cdist_sumofsquares_EggaN_aldist_male__weight"   "cdist_sumofsquares_EggaN_matp_model__num"          
[75] "nll_cdist_sumofsquares_EggaN_aldist_female__num"    "cdist_sumofsquares_EggaS_matp_model__num"          
[77] "cdist_sumofsquares_EggaN_aldist_female_obs__num"    "nll_cdist_sumofsquares_TrawlRus_ldist__num"        
[79] "nll_cdist_sumofsquares_TrawlRus_ldist__weight"      "adist_surveyindices_log_Juv_SI_2_obs__wgt"         
[81] "adist_surveyindices_log_Juv_SI_1_model__wgt"        "nll_cdist_sumofsquares_EggaN_matp__weight"         
[83] "cdist_sumofsquares_EggaN_matp_obs__num"             "cdist_sumofsquares_EggaN_ldist_obs__num"           
[85] "adist_surveyindices_log_Juv_SI_2_model__wgt"        "nll_cdist_sumofsquares_WinterS_ldist__num"         
[87] "cdist_sumofsquares_EggaN_aldist_male_model__num"    "cdist_sumofsquares_TrawlRus_sexdist_obs__num"      
[89] "cdist_sumofsquares_EcoS_sexdist_model__num"        

[willbutler42]

I assume you're using the latest github version of g3? If so, this is likely due to a change in how reporting is handled in models. I can push through a fix tomorrow.

[MikkoVihtakari]

Yup, using the newest one as far as I can tell.

[MikkoVihtakari]

As we discussed on Teams, this stems from recent updates to gadget3. g3_iterative() works on the CRAN version of gadget3. I'll keep on using that version for now.

[willbutler42]

g3_iterative should work with the latest version of g3 now (see ef409be). This is not an ideal fix, but it does the job for the time being.