tests/testthat/_snaps/tagging.md

tagging requires optional arguments to be named

Code
  tagging(input, tagger, "legacy")
Condition
  Error in `tagging()`:
  ! `...` must be empty.
  x Problematic argument:
  * ..1 = "legacy"
  i Did you forget to name an argument?
Code
  tagging_batch(input, tagger, "legacy")
Condition
  Error in `tagging_batch()`:
  ! `...` must be empty.
  x Problematic argument:
  * ..1 = "legacy"
  i Did you forget to name an argument?

tagging rejects an invalid batch value

Code
  tagging(input, tagger, batch = "data.frame")
Condition
  Error in `tagging()`:
  ! `batch` must be one of "list" or "flatten", not "data.frame".

tagging snapshots invalid inputs

Code
  tagging("测试", segmenter)
Condition
  Error in `tagging()`:
  ! `jiebar` must be a `jieba_tagger` object created with `worker(type = "tag")`.
Code
  tagging(NA_character_, tagger)
Condition
  Error in `tagging()`:
  ! `code` must be a non-empty character vector without missing values.
Code
  tagging("测试", tagger, format = "bad")
Condition
  Error in `tagging()`:
  ! `format` must be one of "vector", "data.frame", or "legacy", not "bad".
Code
  tagging_batch("测试", tagger, batch = "bad")
Condition
  Error in `tagging_batch()`:
  ! `batch` must be one of "list" or "flatten", not "bad".


Try the jiebaRS package in your browser

Any scripts or data that you put into this service are public.

jiebaRS documentation built on Oct. 13, 2026, 5:07 p.m.