object-builders: Object builders

In quanteda: Quantitative Analysis of Textual Data

Object builders

Description

Functions to build or re-build core objects, or to upgrade earlier versions of these objects to the current format.

Usage

build_dfm(
  x,
  features,
  docvars = data.frame(),
  meta = list(),
  class = NULL,
  ...
)

rebuild_dfm(x, attrs)

upgrade_dfm(x)

build_tokens(
  x,
  types,
  padding = TRUE,
  docvars = data.frame(),
  meta = list(),
  class = NULL,
  ...
)

rebuild_tokens(x, attrs)

upgrade_tokens(x)

build_corpus(x, docvars = data.frame(), meta = list(), class = NULL, ...)

rebuild_corpus(x, attrs)

upgrade_corpus(x)

build_dictionary2(x, meta = list(), class = "dictionary2", ...)

rebuild_dictionary2(x, attrs)

upgrade_dictionary2(x)

build_fcm(
  x,
  features1,
  features2 = features1,
  meta = list(),
  class = "fcm",
  ...
)

rebuild_fcm(x, attrs)

upgrade_fcm(x)

Arguments

x	an input corpus, tokens, dfm, fcm or dictionary object.
features	character for feature of resulting dfm.
docvars	data.frame for document level variables created by make_docvars(). Names of documents are extracted from the docname_ column.
meta	list for meta fields
class	class labels to be attached to the object.
...	values saved in the object meta fields. They overwrite values passed via meta. If not specified, default values in make_meta() will be used.
attrs	a list of attributes to be reassigned
types	character for types of resulting the tokens object.
padding	logical indicating if the tokens object contains paddings.
features1	character for row feature of resulting fcm.
features2	character for column feature of resulting fcm iff. different from feature1

Examples

quanteda:::build_tokens(
    list(c(1, 2, 3), c(4, 5, 6)),
    docvars = quanteda:::make_docvars(n = 2L),
    types = c("a", "b", "c", "d", "e", "f"),
    padding = FALSE
)
quanteda:::build_corpus(
    c("a b c", "d e f"),
    docvars = quanteda:::make_docvars(n = 2L),
    unit = "sentence"
)

quanteda documentation built on Sept. 11, 2024, 6:08 p.m.