search

pacific-hake / hake-assessment

:zap: :fish: Build the assessment document using latex and knitr
MIT License
13 stars 6 forks source link

t.v. non-parametric selectivity added twice #659

Closed aaronmberger-nwfsc closed 4 years ago

aaronmberger-nwfsc commented 4 years ago

Looks like we have two equivalent models in t.v. sensitivity plot:

Sensitivity models group 4

-----------------------------------------------------------------------------

sens.model.dir.names.4 <- c("2020.01.27_tvSelect_phi_extralow", "2020.01.28_tvSelect_phi_low", "2020.01.29_tvSelect_phi_high", "2020.01.26_semiPara_tvSelect_sig0.695", "2020.01.26_semiPara_tvSelect_sig0.695") sens.model.names.4 <- c("Phi t.v. selectivity (0.21)", "Phi t.v. selectivity (0.70)", "Phi t.v. selectivity (2.10)", "Semi-Parametric t.v selectivity (0.695)", "Semi-Parametric t.v. selectivity (1.0)")

We only compare 0.695.

aaronmberger-nwfsc commented 4 years ago

Chris and I made the decision to just quickly run this missing sensitivity (missing relative to last year). So you will need to:

  1. download model 2020.01.37_semiPara_tvSelect_sig1.0 from the google drive
  2. run build(.model_list = "2020.01.37_semiPara_tvSelect_sig1.0")

I have changed the bit of text that describes this run in the sensitivity section. I still have the correlation part to do (doing now), but that is another issue.

cgrandin commented 4 years ago

I did it. Worked perfectly first time.

andrew-edwards commented 4 years ago

Damn, I thought build() would just find the one missing the .Rdata file, as everything is FALSE. (This is sooo last week...). And I get an error:

Error in fetch_retrospectives(model$retropath, retrospective_yrs) : 
  object 'retros_list' not found

Ugh - and I can't recover the .Rdata file from my recycle bin.It was for 2020.01.23_age1Survey.Rdata

cgrandin commented 4 years ago

Just run build again with both models as a vector:

build(.model_list = c("2020.01.23_age1Survey", "2020.01.37_semiPara_tvSelect_sig1.0"))
andrew-edwards commented 4 years ago

Aha, my retros were only to 10 but now it thinks they're to 20. But we don't need them for appendix. Have deleted the retros folder and rebuilt the .RData file.

andrew-edwards commented 4 years ago

Ooh, I get to close. Realised I didn't finish the QAQC on sensitivity input files (did the base model ones a couple of weeks back). Mean to go back and do them once we had the new model. My bad. Figured we might spot any glaring issues, like this one!