arframe

Regulatory-grade clinical tables, listings, and figures in RTF, PDF, and HTML from a single specification.

arframe is an R package for pharmaceutical submissions. You describe the output once; arframe renders RTF, PDF, and HTML — change the file extension, the output adapts.

tbl_demog |>
  fr_table() |>
  fr_titles("Table 14.1.1", "Demographics and Baseline Characteristics") |>
  fr_cols(
    characteristic = fr_col("", width = 2.5),
    placebo        = fr_col("Placebo", align = "decimal"),
    zom_50mg       = fr_col("Zomerane 50mg", align = "decimal"),
    zom_100mg      = fr_col("Zomerane 100mg", align = "decimal"),
    total          = fr_col("Total", align = "decimal"),
    group          = fr_col(visible = FALSE),
    .n = c(placebo = 45, zom_50mg = 45, zom_100mg = 45, total = 135)
  ) |>
  fr_hlines("header") |>
  fr_render("Table_14_1_1.rtf")  # Change to .pdf — same output, different format

Why arframe?

The pharmaverse has excellent tools for data derivation (admiral), analysis (Tplyr, rtables, tern), and summarisation (gtsummary, tfrmt). But no single package closes the last mile: taking a finished analysis table and producing submission-grade RTF, PDF, and HTML from the same specification.

• r2rtf produces RTF only. Zero PDF.
• tfrmt + docorator produces good PDF. RTF is documented as experimental/lossy.
• gt / gtsummary produces HTML-first output. RTF page headers are not field-code-based.
• None produce all three formats from one spec with running page headers, decimal alignment, and study-wide font control.

arframe is a native rendering engine — it writes RTF 1.9.1 control words and tabularray/LaTeX directly. There is no intermediate gt or pandoc layer. Both RTF and PDF are first-class outputs, produced from the same fr_spec without a format-specific pipeline.

Capability	arframe
Native RTF rendering	Direct RTF 1.9.1 — no intermediate HTML or gt layer
Native PDF rendering	Direct tabularray/XeLaTeX — no RMarkdown/Quarto pipeline
HTML preview	Self-contained HTML with paper simulation and viewer integration
Single spec → RTF + PDF + HTML	Same fr_spec renders to .rtf, .pdf, or .html
Font family (Courier New, Times New Roman)	fr_page(font_family = "Courier New") — first-class parameter via fontspec
Automatic decimal alignment	AFM font metrics for 15 stat types — no manual markers
Auto column width	Physical width from AFM metrics for 12 font variants
Group-aware pagination	Keeps SOC groups together, repeats headers, adds continuation text
Page headers/footers	3-slot layout with live field tokens: {thepage}, {total_pages}, {program}
Study-wide config	_arframe.yml auto-discovered — one file sets font, margins, headers for all programs
Inline markup	fr_super(), fr_bold(), fr_italic(), fr_dagger() in any cell
Dependencies	7 imports — no dplyr, tidyr, purrr, or gt

Installation

# Install from GitHub
# install.packages("pak")
pak::pak("vthanik/arframe")

Quick start

arframe ships with synthetic CDISC ADaM datasets so every example runs out of the box.

A table in 3 lines

library(arframe)

tbl_demog |> fr_table() |> fr_render(tempfile(fileext = ".rtf"))

A production demographics table

# Set shared formatting once — inherited by all tables
fr_theme(
  font_size   = 9,
  font_family = "Times New Roman",
  orientation = "landscape",
  hlines      = "header",
  header      = list(bold = TRUE, align = "center"),
  n_format    = "{label}\n(N={n})",
  pagehead    = list(left = "TFRM-2024-001", right = "CONFIDENTIAL"),
  pagefoot    = list(left = "{program}",
                     right = "Page {thepage} of {total_pages}")
)

n_itt <- c(placebo = 45, zom_50mg = 45, zom_100mg = 45, total = 135)

spec <- tbl_demog |>
  fr_table() |>
  fr_titles(
    "Table 14.1.1",
    "Demographics and Baseline Characteristics",
    "Intent-to-Treat Population"
  ) |>
  fr_cols(
    .width = "fit",
    characteristic = fr_col("", width = 2.5),
    placebo        = fr_col("Placebo", align = "decimal"),
    zom_50mg       = fr_col("Zomerane 50mg", align = "decimal"),
    zom_100mg      = fr_col("Zomerane 100mg", align = "decimal"),
    total          = fr_col("Total", align = "decimal"),
    group          = fr_col(visible = FALSE),
    .n = n_itt
  ) |>
  fr_rows(group_by = "group", blank_after = "group") |>
  fr_footnotes("Percentages based on number of subjects per treatment group.")

# Render both formats from the same spec
fr_render(spec, tempfile(fileext = ".rtf"))
fr_render(spec, tempfile(fileext = ".pdf"))

An AE table with SOC/PT hierarchy

tbl_ae_soc |>
  fr_table() |>
  fr_titles(
    "Table 14.3.1",
    "Treatment-Emergent Adverse Events by SOC and Preferred Term",
    "Safety Population"
  ) |>
  fr_page(continuation = "(continued)") |>
  fr_cols(
    soc       = fr_col(visible = FALSE),
    pt        = fr_col("System Organ Class\n  Preferred Term", width = 3.5),
    row_type  = fr_col(visible = FALSE),
    placebo   = fr_col("Placebo", align = "decimal"),
    zom_50mg  = fr_col("Zomerane\n50mg", align = "decimal"),
    zom_100mg = fr_col("Zomerane\n100mg", align = "decimal"),
    total     = fr_col("Total", align = "decimal"),
    .n = c(placebo = 45, zom_50mg = 45, zom_100mg = 45, total = 135)
  ) |>
  fr_rows(group_by = "soc", indent_by = "pt") |>
  fr_styles(
    fr_row_style(rows = fr_rows_matches("row_type", value = "soc"), bold = TRUE)
  ) |>
  fr_render(tempfile(fileext = ".rtf"))

Pipeline verbs

Every verb takes an fr_spec, returns a modified fr_spec. Verb order doesn't matter — arframe resolves everything at render time.

Verb	Purpose
fr_table() / fr_listing() / fr_figure()	Entry points — create spec from data or plot (figures support a list of plots with per-page metadata tokens)
fr_cols() + fr_col()	Column labels, widths, alignment, visibility, N-counts
fr_header()	Header styling (bold, align, valign, colors)
fr_titles()	Table titles (1-4 lines, with inline markup)
fr_footnotes()	Footnotes with placement control ("every" or "last")
fr_rows()	group_by (with list form for label), group_keep, page_by (with list form for visible), indent_by (multi-level), blank_after, sort_by, suppress, wrap
fr_page()	Paper size, orientation, margins, font, continuation text
fr_pagehead() / fr_pagefoot()	Running headers/footers with token substitution
fr_hlines() / fr_vlines() / fr_grid()	Horizontal/vertical rules (8 presets)
fr_spans()	Spanning column headers
fr_styles()	Cell/row/column conditional formatting
fr_spacing()	Gaps between structural sections
fr_render()	Render to RTF, PDF, HTML, or LaTeX

Key features

Multi-format rendering

fr_render(spec, "output.rtf")   # Native RTF
fr_render(spec, "output.pdf")   # Native PDF via XeLaTeX
fr_render(spec, "output.html")  # Self-contained HTML preview
fr_render(spec, "output.tex")   # LaTeX source

Interactive preview

# In RStudio/Positron — just print the spec to see it in the Viewer panel
spec
# In Rmd/Quarto/pkgdown — tables render inline automatically via knit_print

Automatic N-counts in headers

spec |> fr_cols(
  .n = c(placebo = 45, zom_50mg = 45, zom_100mg = 45, total = 135),
  .n_format = "{label}\n(N={n})"
)

Page tokens

spec |>
  fr_pagehead(left = "Protocol TFRM-2024-001", right = "CONFIDENTIAL") |>
  fr_pagefoot(left = "{program}", right = "Page {thepage} of {total_pages}")

Inline markup

# Works in titles, footnotes, and data cells
spec |> fr_titles("Table 14.1.1{fr_super('a')}", "Demographics")
spec |> fr_footnotes("{fr_super('a')} Post-hoc analysis")

Four-tier defaults

+----------------------------------------------------------------------+
| 4. Per-table verbs         fr_page(font_size = 10)          <- wins  |
| 3. Session theme           fr_theme(font_size = 9)                   |
| 2. Project config          _arframe.yml                              |
| 1. Package defaults        inst/defaults/_arframe.yml       <- lowest|
+----------------------------------------------------------------------+

Recipes: reusable pipeline fragments

company_layout <- fr_recipe(
  fr_page(orientation = "landscape", font_size = 9),
  fr_pagehead(left = "TFRM-2024-001", right = "CONFIDENTIAL"),
  fr_pagefoot(left = "{program}", right = "Page {thepage} of {total_pages}"),
  fr_hlines("header")
)

tbl_demog |> fr_table() |> fr_apply(company_layout) |> fr_titles("Table 14.1.1", "Demographics")

Validation and QC

fr_validate(spec)                        # Pre-render checks
fr_get_titles(spec)                      # Programmatic inspection
fr_style_explain(spec, row = 1, col = "total")  # Debug style cascade

Built-in datasets

Dataset	Description
adsl, adae, adtte, adcm, advs	Synthetic CDISC ADaM datasets (135 subjects)
tbl_demog	Demographics summary (Table 14.1.1)
tbl_ae_soc	AE by SOC/PT (Table 14.3.1.2)
tbl_ae_summary	Overall AE summary (Table 14.3.1.1)
tbl_disp	Subject disposition (Table 14.1.3)
tbl_tte	Time-to-event summary (Table 14.2.1)
tbl_cm	Concomitant medications (Table 14.4.1)
tbl_vs	Vital signs (Table 14.3.6)

Architecture

arframe writes output directly — no intermediate gt or huxtable layer:

fr_spec → finalize_spec() → RTF backend  → .rtf file
                            ├→ LaTeX backend → .tex → XeLaTeX → .pdf
                            └→ HTML backend  → .html file

• RTF: Writes RTF 1.9.1 control words directly. Cell-level formatting, decimal alignment, \trkeep group protection, R-side pagination.
• PDF: Generates tabularray LaTeX, compiles with XeLaTeX. Falls back to Adobe open-source fonts (Source Serif 4, Source Sans 3, Source Code Pro — SIL OFL) on Linux/Docker without Microsoft fonts. Set ARFRAME_FONT_DIR to a directory of .ttf/.otf files for project-local fonts without system-wide installation.
• HTML: Self-contained document with paper simulation (orientation-aware page dimensions, margins). Auto-preview in RStudio/Positron viewer via print(). Inline rendering in Rmd/Quarto via knit_print().
• Font metrics: Real Adobe Font Metrics (AFM) for 12 font variants — accurate column width estimation without rendering.

Documentation

• Get Started — Your first table to production-ready output
• Columns & Headers — Widths, alignment, N-counts, decimal alignment
• Titles & Footnotes — Titles, footnotes, inline markup
• Rows & Page Layout — Grouping, pagination, page chrome
• Rules, Spans & Styles — Borders, spanning headers, conditional formatting
• Table Cookbook — 8 complete ICH E3 table programs
• Listings & Figures — Patient listings and embedded figures
• Automation & Batch — Themes, recipes, YAML config, batch rendering
• Architecture — Internal design and backend interface

Related packages

arframe sits at the output layer of the pharmaverse — it receives whatever the analysis layer produces and renders it to files. It does not replace upstream packages; it completes them.

Package	Role	Relationship to arframe
admiral	ADaM dataset creation	Upstream — arframe renders admiral outputs
Tplyr / rtables	Table summarisation	Upstream — arframe renders their wide-format output
cards / cardx	ARD creation	Upstream — fr_wide_ard() converts ARD to arframe input
gtsummary	High-level summaries	Upstream — arframe renders the same data gtsummary summarises
tfrmt	ARD display spec + mock shells	Complementary — tfrmt for SAP-phase shell review; arframe for production output
r2rtf	RTF-only rendering	Parallel — r2rtf is RTF-specialist; arframe adds PDF + HTML from the same spec
docorator	gt-based PDF decoration	Parallel — docorator wraps gt objects; arframe is a native engine without a gt layer

License

MIT