Update functions that relied on 'flextable' and convert to 'gt'

[sbreitbart-NOAA]

What is the feature?

• Update functions that relied on 'flextable' and convert to 'gt'

How have you implemented the solution?

• Update ID_tbl_width_class() to work with gt tables, vs flextables
• Create ID_tbl_length_class() to ID table length, to decide whether to split them across pages lengthwise
• Update create_tables_doc() to properly format and split tables with varying widths and lengths in tables qmd
• Update/create tests
• Update render_lg_table()
• Update captions for split tables by removing individual column names shown, and instead state number of table vs total number of tables
• Edit add_child() to add line to tables chunk in skeleton to add "clearpage" latex command, which ensures all tables will be plotted before any other section starts (otherwise, Figures and Appendix titles appeared between tables)
• Other minor fixes like removing stale, duplicate license

Does the PR impact any other area of the project, maybe another repo?

• Now it better connects with stockplotr's table exporting functionality

These reports can be tagged. ✅

[github-actions]

Checklist

• PR base branch is accurate
• Is the code concise?
• Comments are clear and useful.
• Can you remove or combine any arguments?
• Do argument contain defaults (if appliable)?
• Code is documented and example provided (Roxygen).
• Did you make a test (testthat)?
• Was this tested under multiple scenarios?
• Did you run devtools::check()?

[github-actions]

Code Metrics Report

Coverage	Code to Test Ratio	Test Execution Time
40.0%	1:0.3	30s

Code coverage of files in pull request scope (67.6%)

Files	Coverage
R/ID_tbl_length_class.R	93.7%
R/ID_tbl_width_class.R	94.7%
R/add_child.R	95.2%
R/create_tables_doc.R	58.1%
R/export_split_tbls.R	100.0%
R/render_lg_table.R	96.4%

---

Reported by octocov

[Schiano-NOAA]

On initial review, the code looked good. It was really cool seeing one of the functions at 100% testing coverage. I'm not sure what, but something is going wrong in the code for splitting the tables. I tested this out on a table_landings that I exported which looked fine in the view, but when it was added to a report and rendered, it resulted in the table plus split tables containing all NA.

Here is what the table looks like straight from stockplotr::table_landings()

[break in rows because it was a long table]

and here is the issue in the rendered report:

The above table in the rendered report only displays the values up to 1917 while the data goes until 2024

the table continues with all NAs for an additional 3 pages.

Here is the workflow I used:

stockplotr::convert_output(file = "C:/Users/samantha.schiano/Documents/Stock Assessment Workflow/SS_output_examples/Sablefish/Results-2025/Report.sso", save_dir = file.path(getwd(), "conout.rda"))

load(file.path(getwd(),"conout.rda"))
stockplotr::table_landings(out_new, make_rda = TRUE)

create_template(model_results = file.path(getwd(), "conout.rda"))

[Schiano-NOAA]

I got a new error when running the code with your recent change

processing file: sar_Petrale_skeleton.qmd
  |....................................       |  85% [tables]                  

Error in `gt::gt_split()`:
! argument "data" is missing, with no default
Backtrace:
    x
 1. +-gt::grp_pull(gt::gt_split(row_every_n = 38), 1)
 2. | \-gt:::extract_gt_tbl_from_gt_group(data = data, which = which)
 3. |   \-gt:::get_indices_gt_tbls(data = data)
 4. |     \-base::nrow(data[["gt_tbls"]])
 5. \-gt::gt_split(row_every_n = 38)
 6.   \-gt:::stop_if_not_gt_tbl(data = data)
 7.     \-gt:::is_gt_tbl(data = data)

Quitting from sar_Petrale_skeleton.qmd:225-234 [tables]
                                                                                                                     
Execution halted
WARN: Error encountered when rendering files

I am not sure what could be wrong with my system. I think it's got to be something with the code, but I guess I will have to debug this issue if you can't replicate it :/

[sbreitbart-NOAA]

I got a new error when running the code with your recent change

processing file: sar_Petrale_skeleton.qmd
  |....................................       |  85% [tables]                  

Error in `gt::gt_split()`:
! argument "data" is missing, with no default
Backtrace:
    x
 1. +-gt::grp_pull(gt::gt_split(row_every_n = 38), 1)
 2. | \-gt:::extract_gt_tbl_from_gt_group(data = data, which = which)
 3. |   \-gt:::get_indices_gt_tbls(data = data)
 4. |     \-base::nrow(data[["gt_tbls"]])
 5. \-gt::gt_split(row_every_n = 38)
 6.   \-gt:::stop_if_not_gt_tbl(data = data)
 7.     \-gt:::is_gt_tbl(data = data)

Quitting from sar_Petrale_skeleton.qmd:225-234 [tables]
                                                                                                                     
Execution halted
WARN: Error encountered when rendering files

I am not sure what could be wrong with my system. I think it's got to be something with the code, but I guess I will have to debug this issue if you can't replicate it :/

Which data did you use / can you add the code here? I'll try to reproduce your error. If I can't let's cowork on this next week

[Schiano-NOAA]

@sbreitbart-NOAA it's happened with multiple examples. The most current one the code was

stockplotr::table_landings(stockplotr::example_data, make_rda = TRUE)
create_template()

then render to get this error

==> quarto preview sar_species_skeleton.qmd --to pdf --no-watch-inputs --no-browse



processing file: sar_species_skeleton.qmd
  |....................................       |  85% [tables]                  

Error in `gt::gt_split()`:
! argument "data" is missing, with no default
Backtrace:
    x
 1. +-gt::grp_pull(gt::gt_split(row_every_n = 38), 1)
 2. | \-gt:::extract_gt_tbl_from_gt_group(data = data, which = which)
 3. |   \-gt:::get_indices_gt_tbls(data = data)
 4. |     \-base::nrow(data[["gt_tbls"]])
 5. \-gt::gt_split(row_every_n = 38)
 6.   \-gt:::stop_if_not_gt_tbl(data = data)
 7.     \-gt:::is_gt_tbl(data = data)

Quitting from sar_species_skeleton.qmd:225-234 [tables]
                                                                                                                     
Execution halted
WARN: Error encountered when rendering files

I will upload my document I made to a github repo. Maybe see if you can pinpoint the error or render it?

[sbreitbart-NOAA]

@Schiano-NOAA I noticed that my last commit missed adding in a few pipes, which broke the code. I've fixed that now and am able to produce reports with all kinds of tables (per width and length). Here's my code, if you want to try it. I understand if you'd rather wait to discuss this before trying it once more!

# All table scenarios (length, width)
scenarios <- list(
  list(name = "reg_reg",    rows = 25, cols = 2),
  list(name = "reg_long",   rows = 45, cols = 2),
  list(name = "wide_reg",   rows = 25, cols = 5),
  list(name = "wide_long",  rows = 45, cols = 5),
  list(name = "ewide_reg",  rows = 25, cols = 10),
  list(name = "ewide_long", rows = 45, cols = 10)
)

# Locations to sample from
location_pool <- c("Providence Rhode Island", "Narragansett Bay", "Point Judith Pond", "Block Island Sound",
  "Cape Cod Bay", "Georges Bank", "Gulf of Maine", "Montauk New York", "Hudson Canyon",
  "Long Island Sound", "Chesapeake Bay", "Outer Banks NC", "Florida Keys", "Gulf of Mexico",
  "Monterey Bay CA", "Puget Sound WA", "Kodiak Alaska", "Bering Sea", "Hawaiian Islands",
  "Sargasso Sea", "Buzzards Bay MA", "Stellwagen Bank", "Vineyard Sound", "Great South Channel",
  "Casco Bay Maine", "Grand Banks", "Scotian Shelf", "Delaware Bay", "Pamlico Sound NC",
  "Charleston Bump SC", "Gray Reef GA", "St Johns River FL", "Mobile Bay AL", 
  "Flower Garden Banks", "Channel Islands CA", "Columbia River Estuary", 
  "Gulf of the Farallones", "Aleutian Islands", "Chukchi Sea", "Beaufort Sea")

tbl_path <- fs::path(getwd(), "tables")
dir.create("tables")
if (!dir.exists(tbl_path)) dir.create(tbl_path)

# Iterate through scenarios and save tables as rdas
for (s in scenarios) {
  
  mat <- matrix(
    sample(location_pool, s$rows * s$cols, replace = TRUE), 
    nrow = s$rows, 
    ncol = s$cols
  )
  
  df <- as.data.frame(mat)
  colnames(df) <- paste0("Col_", 1:s$cols)
  
  test_table <- df |> gt::gt()
  
  rda <- list(
    "table" = test_table,
    "caption" = paste("Test caption for", s$name)
  )
  
  file_name <- paste0(s$name, "_table.rda")
  save(rda, file = fs::path(tbl_path, file_name))
  }

rm(test_table, rda, s, scenarios, mat, df, location_pool, tbl_path)

create_template()
quarto::quarto_render(here::here("report", "sar_species_skeleton.qmd"))

[Schiano-NOAA]

Awesome! I tried it out with the fixes you pushed and I am still getting the NA tables. I made a repo if you want to try and fork it then render what I made.

I will try the code you added above.

If you are able to render/make it fine based on my example, then it must be my computer, but we will still need to figure out next week what is going on to advise users when this happens. Thanks for keeping at it!

[Schiano-NOAA]

Ugh yeah still getting the same issue with your code too. We can go through the source code on my machine next Tuesday to see what is going on

[sbreitbart-NOAA]

@Schiano-NOAA I forked your repo, downloaded the ft-to-gt branch, updated the tables_dir in the tables doc, and was able to render both the tables qmd and the skeleton without issue. Hopefully we can sort this out next week!