The TOC's tables in batches, one family each

A chat drafts a few reports well and many badly: the table prompt (tfl_ai_prompt()) is made for 5 to 10 reports a request. This splits the TOC's tables into batches of one family each – disposition, demographics, exposure, adverse events, laboratory, vital signs, efficacy, PK, other, by the words of the title (English or Japanese) – so that one request holds tables that look alike, and a reviewed table of a family can be the example for the rest of it (tfl_ai_context()'s examples).

Usage

tfl_ai_batches(x, size = 8L, types = "table", families = NULL)

Arguments

x

A tfl_ai_read_shell() result (its toc), or a data frame with output_id and title (and optionally type, family).

size

The reports a batch holds at most.

types

The kinds of report kept (type; a report whose kind is not known is kept); NULL for all.

families

Your own families: a named character vector of regular expressions (Perl) tried on the title before the built-in ones.

Value

A data frame: output_id, title, family, batch (ae-1, ae-2 ...), in the TOC's order within a batch; split(x$output_id, x$batch) gives the requests.

Examples

f <- system.file("ai", "example", "onco_toc_shell.xlsx", package = "tflspec")
b <- tfl_ai_batches(tfl_ai_read_shell(f), size = 2)
b
#>   output_id
#> 1  T-14-1-1
#> 2  T-14-3-1
#> 3  T-14-3-5
#> 4  T-14-2-1
#>                                                                        title
#> 1                                   Demographic and Baseline Characteristics
#> 2 Treatment-Emergent Adverse Events by System Organ Class and Preferred Term
#> 3  Shift from Baseline in Laboratory Normal Range Indicator at Cycle 2 Day 1
#> 4                          Best Overall Response and Objective Response Rate
#>         family          batch
#> 1 demographics demographics-1
#> 2           ae           ae-1
#> 3         labs         labs-1
#> 4     efficacy     efficacy-1
split(b$output_id, b$batch)
#> $`ae-1`
#> [1] "T-14-3-1"
#> 
#> $`demographics-1`
#> [1] "T-14-1-1"
#> 
#> $`efficacy-1`
#> [1] "T-14-2-1"
#> 
#> $`labs-1`
#> [1] "T-14-3-5"
#>