Add Table 14.1.3 generation script aligned with gtsummary-based TLF style

[Copilot]

This PR adds an R implementation for Table 14-1.03 (Summary of Number of Subjects By Site) based on ADSL, matching the style and structure established in PR #67. It reproduces treatment-wise ITT/Eff/Com counts by pooled/site IDs, including a total row and formatted PDF output.

• Table program implementation

  • Added code/tables/t_14_1_3.r to generate Table 14.1.3 from adsl.json.
  • Derives per-treatment counts for ITT, Eff, and Com using ITTFL, EFFFL, and COMP24FL.
  • Produces grouped rows by SITEGR1 (Pooled ID) and SITEID, with an overall TOTAL row.
  • Preserves missing/dirty site values for display instead of filtering them out.

• Output artifacts and formatting

  • Writes ARD output as data/tables/ard/t_14_1_3.rds using gtsummary + gather_ard().
  • Renders PDF table via gt + docorator with protocol/population headers and note/footer text consistent with the target table style.
  • Keeps pooled ID suppression as a display-only transformation (not in stored ARD data).
  • Applies bold styling to column headers/spanners and keeps label capitalization as ID.

• Repository path/config alignment

  • Extended .Rprofile path list with:
    • table_output = data/tables/pdf
    • table_ard = data/tables/ard
  • Added/updated .gitignore scaffolding under data/tables/, data/tables/ard/, and data/tables/pdf/ for generated outputs.

• Documentation updates

  • Updated AGENTS.md to explicitly require:
    • gtsummary + gather_ard() for ARD generation in TLF scripts.
    • Display of missing/dirty analysis values in tables rather than silently dropping them.

t_14_1_3_ard <- adsl_total %>%
  mutate(
    across(
      c(itt, eff, com),
      ~ if_else(is.na(.x), NA_character_, if_else(.x == 1L, "Y", "N")),
      .names = "{.col}_flag"
    )
  ) %>%
  tbl_summary(
    by = trt01p,
    statistic = all_dichotomous() ~ "{n}",
    include = c(itt_flag, eff_flag, com_flag),
    value = list(c("itt_flag", "eff_flag", "com_flag") ~ "Y")
  ) %>%
  gather_ard()

[bms63]

Created by Copilot

Original from CDISC Pilot

[bms63]

@copilot resolve the merge conflicts in this pull request

[Copilot]

@copilot resolve the merge conflicts in this pull request

Resolved and pushed. I merged origin/main into this branch and fixed all merge conflicts in .Rprofile and table output .gitignore files in commit f92266e.

[DmitryMK]

@bms63 As discussed we are going to use #88, but I had a quick look at this PR.
I'm switching to farming profession as I spent many hours doing this table and GH Copilot got the correct results right away.

Yet there are few things to consider:

1. It did not use gtsummary, so we should put it somewhere in the instructions.
2. It assumed that adding _ard prefix to a data frame, makes it a data frame. We should add an instruction to use gather_ard for the ARD file.
3. Minor style issues, like using "ID" in the labels instead of "Id". Did not use bold style for headers.
4. It dropped missing values/dirty data. E.g., if filtered missing sites. I think sites cannot be missing in ADSL, yet we might want to add a general instruction, that instead of filtering dirty data, we want to show it in the tables.

[bms63]

@copilot can you adjust this code to use the gtsummary R package also please add instructions found into the AGENTS.md file for It assumed that adding _ard prefix to a data frame, makes it a data frame. We should add an instruction to use gather_ard for the ARD file.
Minor style issues, like using "ID" in the labels instead of "Id". Did not use bold style for headers.
It dropped missing values/dirty data. E.g., if filtered missing sites. I think sites cannot be missing in ADSL, yet we might want to add a general instruction, that instead of filtering dirty data, we want to show it in the tables.

[Copilot]

@copilot can you adjust this code to use the gtsummary R package also please add instructions found into the AGENTS.md file for It assumed that adding _ard prefix to a data frame, makes it a data fr...

Implemented in commits 65e0856, 3fed804, and 994655b.

I updated code/tables/t_14_1_3.r to use gtsummary + gather_ard() for ARD output, kept missing/dirty site records in the table instead of filtering them out, and applied header bold styling while keeping Pooled ID/Site ID labels. I also updated AGENTS.md with explicit guidance to use gather_ard() (not _ard naming alone) and to display missing/dirty table records instead of silently dropping them.

Updated table screenshot: