Can I use the udpipe annotated results? about korpus HOT 2 CLOSED

i was now able to check this, and at least with english texts it works out of the box:

library(udpipe)
library(koRpus.lang.en)
model <- udpipe_download_model(language="english")
ud_en <- udpipe_load_model(model$file_model)
x <- udpipe_annotate(ud_en, x="This is my sample text.")
x_kRp <- readTagged(
    as.data.frame(x),
    lang="en",
    tagger="manual",
    doc_id=as.data.frame(x)[1,"doc_id"],
    mtx_cols=c(
        token="token",
        tag="xpos",
        lemma="lemma"
    )
)

x_kRp can now be used with koRpus.

from korpus.

hi,

this is currently not supported out of the box. looks like a worthwhile feature request ;)

the tagged data frame would have to be converted into a compatible set of columns and the tagset used would have to be checked for compatibility with koRpus' language packages.

in theory, koRpus::readTagged() should be straight forward, then.

from korpus.