kwic: Perform keyword-in-context (KWIC) analysis.

Description

Get concordances for the matches for a query / perform keyword-in-context (kwic) analysis.

Usage

kwic(.Object, ...)
# S4 method for context
kwic(
  .Object,
  s_attributes = getOption("polmineR.meta"),
  cpos = TRUE,
  verbose = FALSE
)
# S4 method for slice
kwic(
  .Object,
  query,
  cqp = is.cqp,
  left = getOption("polmineR.left"),
  right = getOption("polmineR.right"),
  s_attributes = getOption("polmineR.meta"),
  region = NULL,
  p_attribute = "word",
  boundary = NULL,
  cpos = TRUE,
  stoplist = NULL,
  positivelist = NULL,
  regex = FALSE,
  verbose = TRUE,
  ...
)
# S4 method for partition
kwic(
  .Object,
  query,
  cqp = is.cqp,
  left = getOption("polmineR.left"),
  right = getOption("polmineR.right"),
  s_attributes = getOption("polmineR.meta"),
  p_attribute = "word",
  region = NULL,
  boundary = NULL,
  cpos = TRUE,
  stoplist = NULL,
  positivelist = NULL,
  regex = FALSE,
  verbose = TRUE,
  ...
)
# S4 method for subcorpus
kwic(
  .Object,
  query,
  cqp = is.cqp,
  left = getOption("polmineR.left"),
  right = getOption("polmineR.right"),
  s_attributes = getOption("polmineR.meta"),
  p_attribute = "word",
  region = NULL,
  boundary = NULL,
  cpos = TRUE,
  stoplist = NULL,
  positivelist = NULL,
  regex = FALSE,
  verbose = TRUE,
  ...
)
# S4 method for corpus
kwic(
  .Object,
  query,
  cqp = is.cqp,
  check = TRUE,
  left = as.integer(getOption("polmineR.left")),
  right = as.integer(getOption("polmineR.right")),
  s_attributes = getOption("polmineR.meta"),
  p_attribute = "word",
  region = NULL,
  boundary = NULL,
  cpos = TRUE,
  stoplist = NULL,
  positivelist = NULL,
  regex = FALSE,
  verbose = TRUE,
  ...
)
# S4 method for character
kwic(
  .Object,
  query,
  cqp = is.cqp,
  check = TRUE,
  left = as.integer(getOption("polmineR.left")),
  right = as.integer(getOption("polmineR.right")),
  s_attributes = getOption("polmineR.meta"),
  p_attribute = "word",
  region = NULL,
  boundary = NULL,
  cpos = TRUE,
  stoplist = NULL,
  positivelist = NULL,
  regex = FALSE,
  verbose = TRUE,
  ...
)
# S4 method for remote_corpus
kwic(.Object, ...)
# S4 method for remote_partition
kwic(.Object, ...)
# S4 method for remote_subcorpus
kwic(.Object, ...)
# S4 method for partition_bundle
kwic(.Object, ..., verbose = FALSE)
# S4 method for subcorpus_bundle
kwic(.Object, ...)

Value

If there are no matches, or if all (initial) matches are dropped due to the application of a positivelist, a NULL is returned.

Arguments

.Object: A (length-one) character vector with the name of a CWB corpus, a partition or context object.
...: Further arguments, used to ensure backwards compatibility. If .Object is a remote_corpus of remote_partition object, the three dots (...) are used to pass arguments. Hence, it is necessary to state the names of all arguments to be passed explicity.
s_attributes: Structural attributes (s-attributes) to include into output table as metainformation.
cpos: Logical, if TRUE, a data.table with the corpus positions ("cpos") of the hits and their surrounding context will be assigned to the slot "cpos" of the kwic-object that is returned. Defaults to TRUE, as the availability of the cpos-data.table will often be a prerequisite for further operations on the kwic object. Omitting the table may however be useful to minimize memory consumption.
verbose: A logical value, whether to print messages.
query: A query, CQP-syntax can be used.
cqp: Either a logical value (TRUE if query is a CQP query), or a function to check whether query is a CQP query or not (defaults to auxiliary function is.query).
left: A single integer value defining the number of tokens to the left of the query match to include in the context. Advanced usage: (a) If left is a length-one character vector stating an s-attribute, the context will be expanded to the (left) boundary of the region where the match occurs. (b) If left is a named length-one integer vector, this value is the number regions of the structural attribute referred to by the vector's name to the left of the query match that are included in the context.
right: A single integer value, a length-one character vector or a named length-one integer value, with equivalent effects to argument left.
region: An s-attribute, given by a length-one character vector. The context of query matches will be expanded to the left and right boundary of the region where the match is located. If arguments left and right are > 1, the left and right boundaries of the respective number of regions will be identified.
p_attribute: The p-attribute, defaults to 'word'.
boundary: If provided, a length-one character vector stating an s-attribute that will be used to check the boundaries of the text.
stoplist: Terms or ids to prevent a concordance from occurring in results.
positivelist: Terms or ids required for a concordance to occurr in results
regex: Logical, whether stoplist/positivelist is interpreted as regular expression.
check: A logical value, whether to check validity of CQP query using check_cqp_query.

Details

The method works with a whole CWB corpus defined by a character vector, and can be applied on a partition- or a context object.

If query produces a lot of matches, the DT::datatable() function used to produce output in the Viewer pane of RStudio may issue a warning. Usually, this warning is harmless and can be ignored. Use options("polmineR.warn.size" = FALSE) for turning off this warning.

If a positivelist is supplied, only those concordances will be kept that have one of the terms from the positivelist occurr in the context of the query match. Use argument regex if the positivelist should be interpreted as regular expressions. Tokens from the positivelist will be highlighted in the output table.

If a negativelist is supplied, concordances are removed if any of the tokens of the negativelist occurrs in the context of the query match.

Applying the kwic-method on a partition_bundle or subcorpus_bundle will return a single kwic object that includes a column 'subcorpus_name' with the name of the subcorpus (or partition) in the input object where the match for a concordance occurs.

References

Baker, Paul (2006): Using Corpora in Discourse Analysis. London: continuum, pp. 71-93 (ch. 4).

Jockers, Matthew L. (2014): Text Analysis with R for Students of Literature. Cham et al: Springer, pp. 73-87 (chs. 8 & 9).

Examples

Run this code

use("polmineR")
use(pkg = "RcppCWB", corpus = "REUTERS")

# basic usage
K <- kwic("GERMAPARLMINI", "Integration")
if (interactive()) show(K)
oil <- corpus("REUTERS") %>% kwic(query = "oil")
if (interactive()) show(oil)
oil <- corpus("REUTERS") %>%
  kwic(query = "oil") %>%
  highlight(yellow = "crude")
if (interactive()) show(oil)

# increase left and right context and display metadata
K <- kwic(
  "GERMAPARLMINI",
  "Integration", left = 20, right = 20,
  s_attributes = c("date", "speaker", "party")
)
if (interactive()) show(K)

# use CQP syntax for matching
K <- kwic(
  "GERMAPARLMINI",
  '"Integration" [] "(Menschen|Migrant.*|Personen)"', cqp = TRUE,
  left = 20, right = 20,
  s_attributes = c("date", "speaker", "party")
)
if (interactive()) show(K)

# check that boundary of region is not transgressed
K <- kwic(
  "GERMAPARLMINI",
  '"Sehr" "geehrte"', cqp = TRUE,
  left = 100, right = 100,
  boundary = "date"
)
if (interactive()) show(K)

# use positivelist and highlight matches in context
K <- kwic("GERMAPARLMINI", query = "Integration", positivelist = "[Ee]urop.*", regex = TRUE)
K <- highlight(K, yellow = "[Ee]urop.*", regex = TRUE)

# Apply kwic on partition_bundle/subcorpus_bundle
gparl_2009_11_10_speeches <- corpus("GERMAPARLMINI") %>%
  subset(date == "2009-11-10") %>%
  as.speeches(
    s_attribute_name = "speaker", s_attribute_date = "date",
    progress = FALSE, verbose = FALSE
  )
k <- kwic(gparl_2009_11_10_speeches, query = "Integration")

Run the code above in your browser using DataLab

Last chance! 50% off unlimited learning