Package index • tflspec

ARD spec and ARD code

The analyses of a study as a workbook – its datasets, analysis sets, analysis data and one cards / cardx call per analyses row – and the short R program it stands for (tfl_ard_code()), or that program run (tfl_build_ard()): one study ARD.

tfl_read_ard_spec() tfl_ard_spec(): Read and check an ARD definition workbook
tfl_ard_spec_template(): An empty ARD spec
tfl_ard_code(): The R code that makes the study's ARD
tfl_build_ard(): Make the study's ARD from its definition
tfl_ard_for(): One output's part of the study ARD
tfl_ard_spec_hash(): A fingerprint of one output's ARD definition

Methods, statistics and functions of one’s own

What a method may name: the keywords, any cards / cardx function and its arguments, an analysis written as R (custom), a function of the study’s own.

tfl_ard_methods(): The methods an analysis row may name
tfl_ard_statistics(): The statistics an ARD analysis may ask for
tfl_ard_functions(): The ARD functions an analysis may call
tfl_ard_args(): The arguments of one ARD function, and how each is filled in
tfl_ard_as_custom(): An analysis as R: the code of a custom analysis
tfl_ard_function_template(): Start an ARD function of one's own from a template
tfl_ard_function_info(): What the ARD functions of one's own in some R files are

The functions the programs call

The code of the functions the generated programs call as they run (the code lists on a data, the ids in front, the formats, the study ARD), which a study keeps in a file of its own: its programs run without tflspec.

tfl_helpers_code(): The functions a study's generated programs call

Data, analysis data and populations

The datasets, populations and analysis_data sheets of the ARD spec say what is read and how the analysis data are made (see tfl_read_ard_spec()); these write the line that reads a dataset and give data to try things on.

tfl_read_data_code(): The code that reads one dataset of a data catalog
tfl_read_adam(): Read ADaM datasets
tfl_example_adam(): Example ADaM-like data for trying tflspec

An oncology sample study

Simulated ADaM datasets of one oncology study, XYZ-ONC-01 (Drug X vs Control), that agree with one another; its SAP and CRF outline are in system.file("sample", "onco", package = "tflspec").

adsl: Oncology sample study XYZ-ONC-01: subject level (ADSL)
adex: Oncology sample study XYZ-ONC-01: exposure (ADEX)
adae: Oncology sample study XYZ-ONC-01: adverse events (ADAE)
adlb: Oncology sample study XYZ-ONC-01: laboratory (ADLB)
advs: Oncology sample study XYZ-ONC-01: vital signs (ADVS)
adrs: Oncology sample study XYZ-ONC-01: tumour response (ADRS)
adtr: Oncology sample study XYZ-ONC-01: tumour sizes (ADTR)
adtte: Oncology sample study XYZ-ONC-01: time to event (ADTTE)

Document import

A study’s define (Define-XML 2.0 / 2.1, or the workbook it is made from) read by rule: its datasets, variables and code lists, and a report’s code lists and labels as codelists rows.

tfl_read_define(): Read a Define-XML (2.0 or 2.1)
tfl_read_define_spec(): Read a define spec workbook (an import profile)
tfl_define_profile(): An import profile for a define spec workbook
tfl_define_codelists(): A define's code lists as a report's codelists rows

Spec files (YAML)

A study’s spec as files: one report a file, its content: a list of items, one function call each; read, layered over the study’s defaults, checked against the schema, written as code, and converted from the workbooks.

tfl_read_study_spec() tfl_read_outputs(): Read a study's YAML spec
tfl_read_output() tfl_write_output(): Read and write a report's YAML file
tfl_resolve_output(): A report over the layers it extends
tfl_write_study_part(): Write one part of a study's YAML spec
tfl_item_code() tfl_item_written() tfl_item_own(): One content item: its code, its written arguments, its own part
tfl_schema(): The spec schema: the arguments a spec item may give, and their types
tfl_args_to_r() tfl_args_from_r(): An item's arguments as R, and R as an item's arguments
tfl_schema_check(): Check one spec item against the schema
tfl_schema_help(): The help of a spec item's argument
tfl_content_code(): The code of a report's content, from its spec file
tfl_content_run(): Make a report's content, from its spec file, without its code
tfl_convert_sheets_to_yaml() tfl_yaml_from_sheets(): Convert a study's definition sheets to the YAML spec
tfl_write_spec_xlsx(): Write the flat view of a study's YAML spec as a workbook
tfl_import_spec_xlsx(): Take an edited flat-view workbook back into the YAML spec
tfl_sheets_from_yaml() tfl_sheets_from_spec(): The YAML spec as sheets

Table spec and table plan

A table as a workbook: its roles and the rtfreporter plan verbs it stands for, as a plan (tfl_table_plan()) or as code (tfl_table_code()); a plan written in code back to a workbook.

tfl_table_spec(): A workbook-shaped definition of how an ARD becomes a table
tfl_table_spec_template(): Scaffold a definition workbook from an ARD
tfl_table_plan(): A table's plan from its definition
tfl_normalize_code(): The line that normalizes a table's ARD
tfl_table_code(): The code of a table's plan, from its definition
tfl_as_table_spec(): Write a plan as a table definition workbook

Template SPEC

A report kind from a few parameters: one definition gives the table plan (tpl_*(), the A use) and the report’s items (tfl_template_expand(), the B use). Layout templates (summary) and display templates on them (demog). A report’s YAML file may be a template SPEC, template: {name, args}: tfl_resolve_output() resolves it, its program is the template’s call, and tfl_template_to_spec() writes it out as its items.

tfl_templates(): The templates
tfl_template_expand(): Expand a template into a report's items
tfl_template_to_spec(): A template report written out as its items
tpl_demog(): Demographics: a table plan from the template demog
tpl_summary(): Continuous variables by a column key: a table plan from the layout summary
tpl_ae_soc_pt(): Adverse events by SOC and PT: a table plan from the template ae_soc_pt
tpl_hierarchical(): A hierarchy's counts: a table plan from the layout hierarchical
tpl_disposition(): Disposition: a table plan from the template disposition
tpl_shift() tpl_lab_shift(): Shift tables: table plans from the templates shift and lab_shift
tpl_listing(): A listing: a listing plan from the layout listing
tpl_km_table(): Kaplan-Meier estimates: a table plan from the template km_table
tpl_vs_change(): A vital sign's change: a table plan from the template vs_change
tpl_orr(): Objective response: a table plan from the template orr
tpl_ae_overview(): Overview of adverse events: a table plan from the template ae_overview
tfl_read_template_sheets(): Read reports' template SPECs from a workbook
tfl_write_template_yaml(): Write a report's template SPEC as its YAML file

Template SPEC, a company’s templates

A company’s templates are an R package of its own: it registers them when it is loaded (tfl_template(), usually a tflspec template with the company’s defaults), exports the function its reports’ programs call (tfl_template_run()) and tests that no parameter is the look (tfl_test_templates()). tfl_use_company_templates() writes one to start from.

tfl_template(): Register a template
tfl_param(): A template parameter
tfl_template_params(): A template's parameters
tfl_template_run(): Run a template: its table plan
tfl_test_templates(): Test templates: no parameter is the look
tfl_use_company_templates(): Write a company's template package to start from

Report spec, titles, footnotes, headers and tokens

The page, the running header and footer, the titles and footnotes, the tokens: the document around a table, a listing or a figure.

tfl_read_report_spec(): Read a report definition: the table and everything around it
tfl_report(): Build a report's RTF document from its definition
tfl_report_code(): The code of a report's document, from its definition
tfl_report_setup_code(): The study's setup code for its report programs
tfl_report_tokens(): A report's tokens
tfl_report_path(): Where a report's RTF file goes

Listings

Two sheets keyed by output_id, the pages and the program.

tfl_listing_spec() tfl_read_listing_spec() tfl_write_listing_spec(): A listing definition: read, write and check it
tfl_listing(): A listing's pages from its definition
tfl_listing_code(): The code that makes a listing from its definition
tfl_as_listing_spec(): A listing written in code, as a listing spec

Figures

A figure as a YAML design – its data steps, statistics and layers – and the ggplot2 script it stands for.

tfl_fig_design() tfl_write_fig_design() tfl_read_fig_design() tfl_fig_design_code() tfl_check_fig_design(): A figure design
tfl_fig_content_from_design(): A figure design as a figure report's content
tfl_fig_forest_analyses(): The analyses of a forest plot: the figure's own ARD
tfl_fig_templates() tfl_fig_template(): Figure templates
tfl_fig_parts(): The pieces of a figure design
tfl_fig_objects(): The objects a figure's data steps read and make
tfl_fig_page_size(): A figure's size on its report's page
tfl_fig_uses_ard(): Whether a figure's design reads an ARD
tfl_fig_add_layer(): Add a layer to the figure designs' catalog
tfl_fig_calls(): Extension functions a figure design knows
tfl_fig_compat(): ggplot2 3.5 / 4.0 differences for figure designs
tfl_fig_data_name(): The name of a figure's data in its program
tfl_r() tfl_unset() tfl_call() tfl_item() tfl_output(): R objects of the YAML spec

Figure style

Every look-and-feel value of a study’s figures, in one place.

tfl_fig_style() tfl_fig_style_template() tfl_read_fig_style(): The figure style standard
tfl_fig_setup_code() tfl_check_fig(): The helper script of a study's figure programs
tfl_time_unit_days(): The days in a week, a month and a year
tfl_fig_palettes(): Colour palette presets

Figure spec (sheets) and quick figures

The older sheet-based figure spec and its quick figures, kept while the design is decided (Discussion #63).

tfl_fig_spec(): Build a spec from data frames
tfl_fig_spec_template(): Write an Excel spec template with drop-down lists built from ADaM data
tfl_read_fig_spec(): Read an Excel plot spec
tfl_example_fig_spec(): Example plot spec
tfl_fig_code(): Generate ggplot2 code from a plot spec
tfl_write_fig_code(): Write generated code to .R files
tfl_fig_schema(): The arguments of every figure type, described
tfl_fig_types(): Styles of the quick API
tfl_fig_catalog(): Catalogue of clinical figure types
tfl_fig_list_template(): Write an Excel plot list template
tfl_fig_list_code(): Generate code for every row of a plot list
tfl_fig_km(): Kaplan-Meier plot code (ggsurvfit)
tfl_fig_waterfall(): Waterfall plot code
tfl_fig_swimmer(): Swimmer plot code
tfl_fig_forest(): Forest plot code
tfl_fig_mean(): Mean over time code
tfl_fig_box(): Box plot code
tfl_fig_bar(): Bar chart code for rates and category percentages
tfl_fig_scatter(): Scatter plot code
tfl_fig_individual(): Individual profile code (spaghetti / spider)
tfl_fig_pk(): PK concentration-time plot code
tfl_fig_ae_dot(): AE dot plot code
tfl_fig_butterfly(): AE butterfly plot code
tfl_fig_edish(): eDISH plot code
tfl_fig_sankey(): Sankey diagram code for treatment sequences
tfl_fig_sunburst(): Sunburst code for treatment sequences

Sankey diagrams and sunbursts

tfl_plot_sankey(): Plot Sankey Diagram with Rectangular Nodes and Bezier Polygon Links
tfl_plot_sankey_batch(): Batch Create Sankey Plots for Subgroup Analysis
tfl_sankey_data() tfl_sunburst_data(): Sankey / sunburst input from treatment-line data
tfl_plot_sunburst(): Plot a Static Sunburst of Treatment Sequences

CDISC ARS export and import

The specs as CDISC’s Analysis Results Standard (JSON, and CDISC’s Excel template to read), run by siera; anyone’s ARS back as specs.

tfl_ars(): The specs as a CDISC ARS reporting event
tfl_write_ars_json(): Write a CDISC ARS reporting event as JSON
tfl_write_ars_xlsx(): Write a CDISC ARS reporting event as CDISC's Excel template
tfl_ars_unmapped(): What the ARS does not say
tfl_ars_ard(): Make the ARD of a CDISC ARS reporting event with siera
tfl_read_ars_json(): Read a CDISC ARS reporting event from JSON
tfl_ars_to_specs(): A CDISC ARS reporting event as tflspec specs

The YAML spec (one file a report)

The study’s definition as YAML – study.yml, spec/data.yml, spec/defaults.yml, spec/patterns/, spec/toc.yml, one spec/outputs/<id>.yml a report – its objects, the layering of a report over its defaults, the one-time conversion of the sheets, and the flat view as a workbook.

tfl_r() tfl_unset() tfl_call() tfl_item() tfl_output(): R objects of the YAML spec
tfl_plot(): A plot of a figure's content: its name and its items
tfl_read_output() tfl_write_output(): Read and write a report's YAML file
tfl_read_study_spec() tfl_read_outputs(): Read a study's YAML spec
tfl_resolve_output(): A report over the layers it extends
tfl_convert_sheets_to_yaml() tfl_yaml_from_sheets(): Convert a study's definition sheets to the YAML spec
tfl_write_spec_xlsx(): Write the flat view of a study's YAML spec as a workbook
tfl_import_spec_xlsx(): Take an edited flat-view workbook back into the YAML spec

Reading and writing workbooks, the TOC

Each spec has its own reader and writer; several specs may share one workbook. A company’s table of contents becomes report specs, and an ARD a JSON / YAML / XPT copy.

tfl_write_ard_spec(): Write an ARD definition to a workbook
tfl_read_table_spec(): Read an ARD table definition from a workbook
tfl_write_report_spec() tfl_write_table_spec(): Write a table or report definition to a workbook
tfl_write_specs(): Write several specs to one workbook
tfl_spec_columns(): What each column of a spec workbook means
tfl_read_toc(): Read a table of contents (TOC) as report specs
tfl_write_ard(): Write the study ARD as JSON, YAML or XPT
tfl_read_ard(): Read an ARD written by tfl_write_ard()

Checks

What went wrong, said before a report is made: an ARD and what a table reads from it, a function of one’s own, the ARS, a figure, a report’s own code.

tfl_check_ard(): Check an ARD before a report uses it
tfl_check_code(): Check a report's own code against the rules
tfl_check_ard_function(): Try an ARD function of one's own
tfl_ard_conditions(): What went wrong while an ARD was made
tfl_check_ars(): Check a CDISC ARS reporting event
tfl_fig_advice() tfl_fig_apply_fix(): Advice on a figure design
tfl_check_fig_spec(): Check a plot spec against ADaM data

Review of the whole definition

Every report’s problems at once, one row each: what cannot be used, what is probably wrong and what to set by hand; with the facts of the data and of each report’s ARD.

tfl_review_spec(): Review a study's definition
tfl_review_rules(): The rules of the review
tfl_review_templates(): The sentences of the review beyond its rules
tfl_data_facts(): What the data hold, for the review
tfl_ard_facts(): What an ARD holds, for the review

AI manual and kit

The manual of the installed version, and the kit to attach to a chat that drafts a study’s specs from its TOC and shells; each file it returns reviewed, what the AI can correct apart from what the person decides.

tflspec_ai_manual(): The AI assistant manual that ships with this package
tfl_ai_unpack(): Write an AI assistant's answer into a study's spec files
tfl_adam_metadata(): The ADaM metadata to give an AI assistant
tfl_ai_review(): Review the spec an AI assistant wrote

AI-assisted drafting

A prompt for an AI chat that drafts tables’ report files (the code spec) from a company’s TOC / SHELL workbook and an ADaM summary, the study’s TOC, or a figure design; the answer read and checked by tflspec’s own checks, and written as files or shown as what it changes; nothing is sent anywhere.

tfl_ai_read_shell(): A TOC / SHELL workbook as text for an AI chat
tfl_ai_adam_summary(): The ADaM datasets as a summary for an AI chat
tfl_ai_batches(): The TOC's tables in batches, one family each
tfl_ai_write_requests(): Write a study's table requests, one file a request
tfl_ai_prepare(): Prepare a study's AI table requests in one call
tfl_ai_collect_answers(): Read the answers of a study's requests, check them, write the report files
tfl_ai_study_init(): Start a study folder for drafted report files
tfl_ai_tasks(): The tasks AI-assisted drafting knows
tfl_ai_context(): The study facts a drafting prompt carries
tfl_ai_schema(): The schema text of a drafting task
tfl_ai_prompt(): A drafting prompt for an AI chat
tfl_ai_parse(): Read a drafting answer
tfl_ai_check(): Check a drafting answer in its context
tfl_ai_repair_prompt(): The prompt that sends an answer's problems back
tfl_ai_diff(): What a drafting answer changes

Helpers

tflspec_code_options: How the generated programs name folders and packages
print(<tfl_code>): Code object returned by the quick API