search

pfmc-assessments / indexwc

Estimate indices of abundance for west coast fish species
2 stars 2 forks source link

[BUG] error in sdmTMB call within run_index_comparison.R #4

Closed iantaylor-NOAA closed 2 years ago

iantaylor-NOAA commented 2 years ago

Is there an existing issue for this?

Current Behavior

Based on suggestion via chat with @kellijohnson-NOAA and @chantelwetzel-noaa, I've been working through sandbox/run_index_comparison.R. The call to sdmTMB in these lines https://github.com/kellijohnson-NOAA/indexwc/blob/main/sandbox/run_index_comparison.R#L209-L227 is producing the error Error: class(formula) not identical to "formula"

I note that running class(formula) (on formula object created starting on line 39 of the file) returns [1] "list" Any suggestions?

Expected Behavior

A successful completion of the sdmTMB estimation.

Steps To Reproduce

  1. install the packages loaded at the top of the file, including the dev branch of VAST required to get VAST_v14_0_1
  2. change the setwd() command to point to sandbox on my local computer
  3. add source("my_apply_epsilon.R") near the top to make that function available
  4. add recursive = TRUE to dir.create(dir, showWarnings = FALSE, recursive = TRUE) on line 91
  5. run the script

Environment

- OS:Windows
- Node:?
- npm:?

Anything else?

No response

ericward-noaa commented 2 years ago

Ian -- do you have a shorter reproducible example? I have all these packages updated, but the script crashes out for me in a few places before getting to the sdmTMB call (e.g. 'species' not found at the start of the loop). Or send the R image you're working with?

formula can be a formula, but also a list for delta models -- but I can see what the issue might be

chantelwetzel-noaa commented 2 years ago

I my be off base here, but I am unclear as to what line 39 returns when you run the script in the original issue. Do you mean that it returns a list: 

> formula
[[1]]
catch_mt ~ 0 + as.factor(Year) + pass_scaled + (1 | vessel_scaled)

[[2]]
catch_mt ~ 0 + as.factor(Year) + pass_scaled + (1 | vessel_scaled)

or is it just returning

[1] "list"

Although there is addition code that could modify the formula call after line 39, the returned object should be a list of length two with the same formula for each element. If you are getting the correct list then ignore this comment.

seananderson commented 2 years ago

Is it possible you have an old version of sdmTMB? At one point formula had to be a formula. Now it can be a formula or list of formulas (if you want different formulas for 2 delta model components).

iantaylor-NOAA commented 2 years ago

Thanks everyone for coming to my rescue! @seananderson, you hit the nail on the head. That's embarrassing. I had sdmTMB version 0.0.21.9004 which seems to be from March 1, 2022. I didn't recall installing it and just assumed it was newly installed along with the other packages I added recently.

The update is taking a while and I need to quit for the day but will assume that the newer version will solve the issue.

In case they are useful for other purposes, I just pushed the changes required to source the script on my computer (sorry @ericward-noaa, I clearly left a few out). Sourcing the modified file initially reproduced the error and will hopefully now reproduce the lack of error. The commit to a fork is here: https://github.com/kellijohnson-NOAA/indexwc/compare/main...iantaylor-NOAA:indexwc:debug_issue_4.

iantaylor-NOAA commented 2 years ago

Follow-up: updating sdmTMB indeed resolved the problem. Sorry for the confusion.