Wrap an arbitrary embedding function as a securecontext embedder

Generic plug-in point for external embedders. fn must be a function that takes a character vector and returns a numeric matrix with one row per input text and dims columns. The function is called inside embed_texts() and therefore inherits its tracing span if a trace is active.

Usage

embed_custom(fn, dims, name = "custom")

Arguments

fn

A function function(texts) -> matrix where ncol(result) == dims and nrow(result) == length(texts).

dims

Integer, the dimensionality of the embedding space.

name

Character scalar, an informational label used in diagnostic output. Defaults to "custom".

Value

A securecontext_embedder object.

Details

Unlike embed_tfidf(), no local corpus is needed; you are responsible for ensuring the underlying model produces the dimensionality you declare.

Examples

fake <- function(texts) matrix(runif(length(texts) * 8L), ncol = 8L)
emb <- embed_custom(fake, dims = 8L, name = "fake")
embed_texts(emb, c("hello", "world"))
#>           [,1]       [,2]      [,3]       [,4]      [,5]      [,6]      [,7]
#> [1,] 0.9805397 0.05144628 0.6958239 0.03123033 0.3008308 0.4790245 0.7064338
#> [2,] 0.7415215 0.53021246 0.6885560 0.22556253 0.6364656 0.4321713 0.9485766
#>           [,8]
#> [1,] 0.1803388
#> [2,] 0.2168999