Release 1.3.0 (#102)

* adding covariate-based geometric anisotropy

* try to make CI happy

* change interface

* Update develop_kappa_covariates.R

* Update develop_kappa_covariates.R

* Update internal.R

* add `get_cor` as internal function

* alpha-version for `spatial_cor`

* make CI happy

* robustify `spatial_cor`

* Create develop_barrier_spde.R

* add `triangle_formula` for barriers

* Update develop_kappa_covariates.R

* Update develop_kappa_covariates.R

* allow offset in `triangle_covariates`

* fix `make_eof_ram` to work as intended for multivariate models

e.g., having a separate response map and shared indices across variables (h/t Jon Reum for identifying the issue)

* update NEWS

* fix bug in triangle_formula offset implementation

* fix another bug in triangle_formula offset ...

... where previously the offset was mapped off at 0

* another bug in triangle_formula offset ...

... arising when no formula is given

* Update fit.R

* further updates to scratch examples

* Vertex covariates barrier (#88)

* Refactor add_vertex_covariates()

- Add barrier option
- Faster gstat IDW and *much* faster RANN IDW

* Visualize vertex covariate example

Also suppress gstat message

* Add triangle covariates and barrier TRUE/FALSE

* add_vertex_covariates -> add_mesh_covariates()

And add tests

* Add to triangles example

* make CI happy

* fix predict for spatially_varying using factors

---------

Co-authored-by: Sean Anderson <[email protected]>
Co-authored-by: Jim Thorson <[email protected]>

* use `size_i` for binomial size and REPORT negloglik_i

* make CI happy

* make CI happy

* add strength of barrier-effect to tinyVASTcontrol

* update docs

* more options controlling SE calcs

* make CI happy

* add math note for RSR

* Update project.R

* Update develop_project.R

* Update develop_project.R

* Create test_geometric_anisotropy.R

* reparameterize extra H values

* bug fix #92

h/t Anthony Charsley for flagging it.

* add `development` to `project` calls to `tinyVAST`

* add `PESC_example_red_grouper` from VAST ...

... as prep for a multi-prey single-predator diet analysis vignette (h/t @holeynik)

* make CI happy

* port over VAST PESC example as vignette

* update docs for `spatial_varying`

* bug fix for #97

h/t Peri Gerson

* update `profile` docs

* add data for forecasting vignette

* adding forecasting vignette

* further explore geometric anisotropy

* change TMB code for aniso-covariates

* adding notes on edge-covariates

* allowing NAs in vertex_covariates ...

... as interface for stratum-specific geometric anisotropy

* sablefish surveyjoin anistropy example

* add sponge-coral-fish web-only vignette

* fix roxygen docs

* Update DESCRIPTION

* make check_win_devel() happy

* shrink alaska_sponge_coral_fish.rda

* make CRAN check happy

* Update probabilitistic-forecasts.Rmd

* make `pkgdown::build_site` happy

* still making `pkgdown::build_site` happy

* updates to vignettes

* vignette updates

* Update VAST.Rmd

---------

Co-authored-by: Jim Thorson <[email protected]>
Co-authored-by: Sean Anderson <[email protected]>

Dev (#87)

* adding geographic SAR with geometric anisotropy option

* generalize SAR input to sfc_GEOMETRY

* add error if spatial_domain doesn't match options

* fix bug in lognormal deviance for delta models

* use weights as N trials for family = "binomial"

* fix typos in previous commit

* add integrated test for Binomial family

* add deviance-residuals for Binomial

* fix percent_deviance for non-default weights

* add `check_tinyVAST_version`

* Update fit.R

* adding `simulate_conditional_gmrf`

* Update tinyVAST.cpp

* adding `project`

* add `supress_user_warnings` to `control(.)`

* fix bugs in `project`

* fix typos

* fix another bug

* Update project.R

* update scratch script for projections

* fix typo in simulate_sfnetworks docs

* expand `project` arguments

* fix bug in last commit

* fix another bug

* Update develop_project.R

* eliminate ... argument to

ensure misnamed arguments are obvious

* fix cAIC example

* update `project`

* make CI happy

* remove `dsem::make_matrices` from `project` ...

... to ensure that model is more stable in terms of dependencies

* add informative error for `predict` using SAR domain...

... for when `newdata` includes locations that are not within the SAR domain

* update model-description vignette ...

... to describe SVCs and also add references included in Zotero "tinyVAST vignettes" collection

* Update _pkgdown.yml

* address CRAN check warning

* fix CRAN check error

* add logo

* run pkgdown for favicon

* add tests for parsing functions in formula

* fix bug resulting in nonconverged model ...

... when dsem arrow-and-lag specification was missing whitespace around double-headed arrow (h/t Jon Reum for flagging it)

* Update predict.R

* edit docs

* Update internal.R

* Update internal.R

* fix test-smooths syntax for mgcv::gam

---------

Co-authored-by: Jim Thorson <[email protected]>

Dev (#75)

* Update DESCRIPTION

* adding geometric_anisotropy

* add VAST_1.2.0.tar.gz

* more scratch helper scripts

* try to fix rhub check

* Update NEWS.md

* change semantic version to a bug fix 1.1.1

* trying to make CRAN check happy ...

... by using donttest for slow examples

Add te/ti and fix small bug in stream-network precision (#74)

* adding `term_covariance`

* fix bug in last commit

* Update _pkgdown.yml

* edit example for term_covariance ...

... to not require package:tweedie

* update `term_covariance` example

* add RSR estimator

* Update NEWS.md

* add SE for RSR

* fix bug in last commit

* Update CITATION

* adding cv::cv(.) methods and integrated test

* fix bug in last commit

* add error check for cv conflict

* try to make CI happy

* fix integrated test for cv::cv ...

... which apparently doesn't allow local definition of input `mesh`

* Update test-basic-fits.R

* update citation

* experiment to make R_CMD_check happy

* another attempt to make R_CMD_check happy

* another attempt

* another attempt

* different way to test cv::cv as example

* add CI comparing tinyVAST vs. sdmTMB

* fix CI in last commit

* fix error-check for control$tmb_par

* revert last commit

* changing check for control$tmb_par, which wasn't working

* fix issue in last commit

* found and fixed problem with control$tmb_par

* add CIs for DSEM vs. tinyVAST

* make CI happy

* further improvement in CI for lagged DSEM

* Update test-dsem.R

* still making CI happy with last commit

* small changes

* add t2 and ti splines

* add untested comparison of wiggliness pars for te / ti tests

* add back s(X, by=F) to CI

* loosely correlation-test for te and ti CI

* make CI happy

* add test for specifying m=1 in spline s(.)

* small fix in message for ti/te

* update stream-network Q ...

... both description in the vignette, a scratch script, and in sfnetwork.R

* fix bug in last commit

* fix bug in stream-network model

* Create run_tests_locally.R

* prep for CRAN update

* make check_win_devel happy ...

... by simplifying examples to address NOTE

---------

Co-authored-by: Jim Thorson <[email protected]>

Dev (#69)

* add newdata for sample_variable

* attempts to fix R checks valgrid and clang-UBSAN

* Create R-CMD-check-valgrind2.yaml

* Delete R-CMD-check-valgrind2.yaml

* Create R-CMD-check-valgrind3.yaml

* try to fix error

* eliminate all examples

* fix last test

* attempt fix for valgrind

* Update R-CMD-check-valgrind3.yaml

* try to fix NA in ivector

* Create rhub.yaml

* remove dontrun examples ...

... to see if that fixes valgrind

* move tests to scratch

* experiments for valgrind

* trying to turn off all tests for exploring valgrind

* Revert "eliminate all examples"

This reverts commit 17417eb.

* Create test.R

* replace GMRF(Q).Quadform with manual ..

... to avoid calc for log-determinant

* update test.R and add valgrind WSL guide

* trying some cleanup

* re-add CI tests

* Update test.R

* Update test.R

* updates before CRAN submit

* remove R-CMD-check-valgrind3.yaml...

... which wasn't showing anything useful anyway.

* re-add CI tests

* update scratch instructions

* Create run_openblas_via_WSL.R

* Update run_openblas_via_WSL.R

* Create extract_covariance.R

* update test syntax

* suppress warnings about NA values by default

* move SAR to web-only vignettes

* Update NEWS.md

---------

Co-authored-by: Jim Thorson <[email protected]>
Co-authored-by: Sean Anderson <[email protected]>

Dev (#61)

* adding nbinom1 and nbinom2

* fix bug in last commit

* add deviance for nbinom2

* try to make CI happy

* add nbinom1 deviance residuals

* update doc slightly

* update argument names

* Update test-smooths.R

* fix bug in last commit

* fix summary.tinyVAST

* make R_CMD_check happy

* make vignettes and R_CMD_check happy

* update docs

* update web_only vignettes

* add time_term

* print invisibly returns contents

* making R_CMD_check happy

* still making R_CMD_check happy

* update vignettes

* further updates to model-description

* re-order vignettes

* add tibble Suggests

* Update DESCRIPTION

* Update NEWS.md

* add test for time_term

* update test for time_term

* fix time_term test

* trying to make R_CMD_check happy

* add pandoc install

* respond to CRAN comments

* address CRAN onboarding comments

* fix bug in doc updates

* fix another bug in updated docs

* fix another bug in last commit

* add simulate S3 generic

* fix bug in simulate(.)

* update simulate docs ...

... to show DHARMa residuals

* make R_CMD_check happy

* another attempt to make R_CMD_check happy

* still trying to make R_CMD_check happy

* another attempt

* adding tmb_map option

* adding placeholder for summary(.) for 2nd linear predictor

* Update DESCRIPTION

* attempt adding summary(.) for 2nd predictor

---------

Co-authored-by: Jim Thorson <[email protected]>

Fix bugs (#59)

* fix bug in deviance_explained ...

... when control$profile eliminated all parameters

* add ggplot for marginal predictive intervals

* clarify variable naming in spatial.Rmd

* adding errors for tibble

* add abundance-expanded option

* fix bug in last commit

* adding likelihood weights ...

... for use in automated EDF and conditional AIC calculations

* update docs @ issue-52

* fix bug delta-models when using SEM/DSEM notation...

... which erroneously triggered a precautionary error-message without running

* add informative error for extra levels in `data`

* Update deviance_explained ...

... to work in case of all fixed effects being profiled

* add warning about number of components in spatial smoother

* fix bug in delta_SEM specification

* add informative error ...

... when doing bias-correction without random effects

* adding REML option

* fix `integrate_output` bug caught in CI

* fix bug in predict for poisson-link delta model

h/t Jeremy Harris for flagging it

* update version number

* update NEWS

* adding cAIC

* fix bug in last commit

* try to make R_CMD_check happy

* Try to fix test-coverage

* another fix for test-coverage

---------

Co-authored-by: Jim Thorson <[email protected]>

Add svc (#49)

* adding prototype SVC option

* Update fit.R

* fix predict(.) for SVC option

* Update predict.tinyVAST.Rd

* small update

---------

Co-authored-by: Jim Thorson <[email protected]>

Add simulation residuals (#38)

* adding example of DHARMa for sim-residuals

* update DESCRIPTION

---------

Co-authored-by: Jim Thorson <[email protected]>

Change format and code for families argument (#37)

* small update to age-comp example

* adding condition vignette

* updates to vignettes

* adding reload_model(.) and sample_variable(.)

* fix bug in last commi

* fix another bug in last commit

* update family.R

* fix tests from last commit

---------

Co-authored-by: Jim Thorson <[email protected]>