cwbtools (version 0.3.3)

cwb_corpus_dir: Manage directories for indexed corpora

Description

The Corpus Workbench (CWB) stores the binary files for structural and positional attributes in an individual 'data directory' (referred to by argument data_dir) for each corpus. The data directories will typically be subdirectories of a parent directory called 'corpus directory' (argument corpus_dir). Irrespective of the location of the data directories, all corpora available on a machine are described by so-called (plain text) registry files stored in a so-called 'registry directory' (referred to by argument registry_dir). The functionality to manage theses directories is used as auxiliary functionality by higher-level functionality to download and install corpora.

Usage

cwb_corpus_dir(registry_dir)

cwb_registry_dir()

cwb_directories(registry_dir = NULL, corpus_dir = NULL)

create_cwb_directories(prefix = "~/cwb", ask = interactive(), verbose = TRUE)

use_corpus_registry_envvar(registry_dir)

Arguments

registry_dir

Path to the directory with registry files.

corpus_dir

Path to the directory with data directories for corpora.

prefix

The base path that will be prefixed

ask

A logical value, whether to prompt user before creating directories.

verbose

A logical value, whether to output status messages.

Details

cwb_corpus_dir will make a plausible suggestion for a corpus directory where data directories for corpora reside. The procedure requires that the registry directory (argument registry_dir) is known. If the argument registry_dir is missing, the registry directory will be guessed by calling cwb_registry_dir. The heuristic to detect the corpus directory is as follows: First, directories in the parent directory of the registry directory that contain "corpus" or "corpora" are suggested. If this does not yield a result, the data directories stated in the registry files are evaluated. If there is one unique parent directory of data directories (after removing temporary directories and directories within packages), this unique directory is suggested. cwb_corpus_dir will return a length-one character vector with the path of the suggested corpus directory, or NULL if the heuristic does not yield a result.

cwb_registry_dir will return return the system registry directory. By default, the environment variable CORPUS_REGISTRY defines the system registry directory. If the polmineR-package is loaded, a temporary registry directory is used, replacing the system registry directory. In this case, cwb_registry_dir will retrieve the directory from the option 'polmineR.corpus_registry'. The return value is a length-one character vector or NULL, if no registry directory can be detected.

cwb_directories will return a named character vector with the registry directory and the corpus directory.

create_cwb_directories will create a 'registry' and an 'indexed_corpora' directory as subdirectories of the directory indicated by argument prefix. Argument ask indicates whether to create directories, and whether user feedback is asked for before creating the directories. The function returns a named character vector with the registry and the corpus directory.

use_corpus_registry_envvar is a convenience function that will assist users to define the environment variable CORPUS_REGSITRY in the .Renviron-file. making it available across sessions. The function is intended to be used in an interactive R session. An error is thrown if this is not the case. The user will be prompted whether the cwbtools package shall take care of creating / modifying the .Renviron-file. If not, the file will be opened for manual modification with some instructions shown in the terminal.