p_attributes: Using Positional Attributes.

CL: p_attributesR Documentation

Using Positional Attributes.

Description

CWB indexed corpora store the text of a corpus as numbers: Every token in the token stream of the corpus is identified by a unique corpus position. The string value of every token is identified by a unique integer id. The corpus library (CL) offers a set of functions to make the transitions between corpus positions, token ids, and the character string of tokens.

Usage

cl_cpos2str(
  corpus,
  p_attribute,
  registry = Sys.getenv("CORPUS_REGISTRY"),
  cpos
)

cl_cpos2id(corpus, p_attribute, registry = Sys.getenv("CORPUS_REGISTRY"), cpos)

cl_id2str(corpus, p_attribute, registry = Sys.getenv("CORPUS_REGISTRY"), id)

cl_regex2id(
  corpus,
  p_attribute,
  regex,
  registry = Sys.getenv("CORPUS_REGISTRY")
)

cl_str2id(corpus, p_attribute, str, registry = Sys.getenv("CORPUS_REGISTRY"))

cl_id2freq(corpus, p_attribute, id, registry = Sys.getenv("CORPUS_REGISTRY"))

cl_id2cpos(corpus, p_attribute, id, registry = Sys.getenv("CORPUS_REGISTRY"))

Arguments

corpus

name of a CWB corpus (upper case)

p_attribute

a p-attribute (positional attribute)

registry

path to the registry directory, defaults to the value of the environment variable CORPUS_REGISTRY

cpos

corpus positions (integer vector)

id

id of a token

regex

a regular expression

str

a character string

Examples

# registry directory and cpos_total will be needed in examples
cpos_total <- cl_attribute_size(
  corpus = "REUTERS", attribute = "word",
  attribute_type = "p", registry = get_tmp_registry()
  )

# decode the token stream of the corpus (the quick way)
token_stream_str <- cl_cpos2str(
  corpus = "REUTERS", p_attribute = "word",
  cpos = seq.int(from = 0, to = cpos_total - 1),
  registry = get_tmp_registry()
  )
  
# decode the token stream (cpos2id first, then id2str)
token_stream_ids <- cl_cpos2id(
  corpus = "REUTERS", p_attribute = "word",
  cpos = seq.int(from = 0, to = cpos_total - 1),
  registry = get_tmp_registry()
  )
token_stream_str <- cl_id2str(
  corpus = "REUTERS", p_attribute = "word",
  id = token_stream_ids, registry = get_tmp_registry()
)

# get corpus positions of a token
token_to_get <- "oil"
id_oil <- cl_str2id(
  corpus = "REUTERS", p_attribute = "word",
  str = token_to_get, registry = get_tmp_registry()
  )
cpos_oil <- cl_id2cpos <- cl_id2cpos(
  corpus = "REUTERS", p_attribute = "word",
  id = id_oil, registry = get_tmp_registry()
)

# get frequency of token
oil_freq <- cl_id2freq(
  corpus = "REUTERS", p_attribute = "word", id = id_oil, registry = get_tmp_registry()
)
length(cpos_oil) # needs to be the same as oil_freq

# use regular expressions 
ids <- cl_regex2id(
  corpus = "REUTERS", p_attribute = "word",
  regex = "M.*", registry = get_tmp_registry()
)
m_words <- cl_id2str(
  corpus = "REUTERS", p_attribute = "word",
  id = ids, registry = get_tmp_registry()
)


RcppCWB documentation built on Sept. 24, 2024, 1:08 a.m.