region_matrix_ops: Get IDs and Counts for Region Matrices.

In RcppCWB: 'Rcpp' Bindings for the 'Corpus Workbench' ('CWB')

Get IDs and Counts for Region Matrices.

Description

Get IDs and Counts for Region Matrices.

Usage

region_matrix_to_ids(
  corpus,
  p_attribute,
  registry = Sys.getenv("CORPUS_REGISTRY"),
  matrix
)

region_matrix_to_count_matrix(
  corpus,
  p_attribute,
  registry = Sys.getenv("CORPUS_REGISTRY"),
  matrix
)

region_matrix_context(
  corpus,
  registry = Sys.getenv("CORPUS_REGISTRY"),
  matrix,
  p_attribute,
  s_attribute,
  boundary,
  left,
  right
)

ranges_to_cpos(ranges)

Arguments

corpus	a CWB corpus
p_attribute	a positional attribute
registry	registry directory
matrix	a regions matrix
s_attribute	If not NULL, a structural attribute (length-one character vector), typically indicating a sentence ("s").
boundary	Structural attribute (length-one character vector) that serves as a boundary and that shall not be transgressed.
left	An integer value, number of strucs to move to the left.
right	An integer value, number of strucs to move to the right.
ranges	A two-column integer matrix of ranges (left and right corpus positions in first and second column, respectively).

Details

ranges_to_cpos() will turn a matrix of ranges into an integer vector with the individual corpus positions covered by the ranges.

Examples

# Scenario 1: Get full text for a subcorpus defined by regions
m <- get_region_matrix(
  corpus = "REUTERS", s_attribute = "places",
  strucs = 4L:5L, registry = get_tmp_registry()
  )
ids <- region_matrix_to_ids(
  corpus = "REUTERS", p_attribute = "word",
  registry = get_tmp_registry(), matrix = m
  )
tokenstream <- cl_id2str(
  corpus = "REUTERS", p_attribute = "word",
  registry = get_tmp_registry(), id = ids
  )
txt <- paste(tokenstream, collapse = " ")
txt

# Scenario 2: Get data.frame with counts for region matrix
y <- region_matrix_to_count_matrix(
  corpus = "REUTERS", p_attribute = "word",
  registry = get_tmp_registry(), matrix = m
  )
df <- as.data.frame(y)
colnames(df) <- c("token_id", "count")
df[["token"]] <- cl_id2str(
  "REUTERS", p_attribute = "word",
  registry = get_tmp_registry(), id = df[["token_id"]]
  )
df[order(df[["count"]], decreasing = TRUE),]
head(df)

RcppCWB documentation built on July 9, 2023, 7:40 p.m.