simple tagger based on text excerpt and available context from data-frames with edam vocabularies

Usage

flat_tagger(
  txt,
  nterms = 20,
  model = "claude-sonnet-4-5",
  provider = "anthropic",
  prompt = read_prompt("flat_tagger.txt"),
  ...
)

Arguments

txt: a text string for analysis
nterms: integer(1) approximate number of EDAM terms to return. Defaults to 20.
model: character(1) model identifier for the selected provider; defaults to "claude-sonnet-4-5" (Anthropic)
provider: character(1) LLM provider; see llm_env_var for supported values and the required environment variable for each. Defaults to "anthropic".
prompt: character(1) instruction text sent to the LLM. Defaults to the contents of inst/prompts/flat_tagger.txt; must contain two %s/%d placeholders for nterms and txt. Use read_prompt("flat_tagger.txt") to inspect the default.
...: parameters passed to the underlying chat_* function via llm_chat

Value

a data.frame with columns id (EDAM CURIE) and lbl (term label)

Note

This function as of Nov 7 2025 will routinely hallucinate associations and terms.

Examples

if (interactive()) {
txt = "The Voyager package is an R/Bioconductor software designed for exploratory spatial 
data analysis (ESDA) of spatial single-cell omics datasets. It operates on the 
SpatialFeatureExperiment (SFE) S4 class, allowing users to perform a wide range of spatial 
statistical analyses directly within a biological context. Univariate global spatial statistics 
supported include Moran's I for measuring spatial autocorrelation, permutation testing 
for assessing significance, and correlograms for examining spatial correlation structure. 
Bivariate spatial statistics implemented in Voyager comprise Lees L statistic 
and cross variograms for evaluating spatial associations between two 
variables. In addition, Voyager provides tools for multivariate analysis using methods 
such as MULTISPATI PCA, which integrates spatial structure into principal component 
analysis, and Anselins recent multivariate local Gearys C" 
flat_tagger(txt, nterms=12)
}