Type: Package
Title: A Companion to the Multi-CAST Collection
Version: 2.0.0
URL: https://multicast.aspra.uni-bamberg.de/
Description: Provides a basic interface for accessing annotation data from the Multi-CAST collection, a database of spoken natural language texts edited by Geoffrey Haig and Stefan Schnell. The collection draws from a diverse set of languages and has been annotated across multiple levels. Annotation data is downloaded on request from the servers of the University of Bamberg. See the Multi-CAST website https://multicast.aspra.uni-bamberg.de/ for more information and a list of related publications.
License: CC BY 4.0
Encoding: UTF-8
LazyData: true
Depends: R (≥ 3.0.0),
RoxygenNote: 7.1.1
NeedsCompilation: no
Packaged: 2021-02-22 00:06:07 UTC; Nils
Author: Nils Norman Schiborr [aut, cre]
Maintainer: Nils Norman Schiborr <nils-norman.schiborr@uni-bamberg.de>
Repository: CRAN
Date/Publication: 2021-02-22 19:20:02 UTC

Count clauses in a multicastR table

Description

mc_clauses counts the number of clause units (bounded by the <##> or <#> GRAID annotation symbols) in a multicastR table.

Usage

mc_clauses(x, bytext = FALSE, printToConsole = FALSE)

Arguments

x

A data.frame in multicastR format. This table minimally requires the corpus and graid columns with the names of the corpora and the GRAID annotation values, respectively, as well as the text column if bytext is set to TRUE.

bytext

Logical. If FALSE, calculate the number of clause units for each corpus. If TRUE, count for each text separately. FALSE by default.

printToConsole

Logical. If TRUE, prints the table to the console (using message). FALSE by default.

Value

A data.frame with the corpus, text (if bytext is TRUE), the number of valid clause units in each corpus (nClause), the total number of clause units (nAll), the number of clause units not analyzed (nNC), and the percentage the later make up of the total (pNC).

See Also

multicast, mc_index, mc_metadata, mc_referents, mc_clauses

Examples

## Not run: 
  # count clause units in the most recent version
  # of the Multi-CAST data, by corpus
  n <- mc_clauses(multicast())

  # count by text instead
  m <- mc_clauses(multicast(), bytext = TRUE)

  # number of clauses units in the whole collection
  sum(n$nClauses)

## End(Not run)

Access the Multi-CAST version index

Description

mc_index downloads a tabular index of the versions of the Multi-CAST corpus data from the servers of the University of Bamberg. The value in the leftmost version column may be passed to the multicast method for access to earlier versions of the annotations.

Usage

mc_index()

Value

A data.frame with five columns:

[, 1] version

Version key. Used for the vkey argument of other functions in this package.

[, 2] date

Publication date in YYYY-MM-DD format.

[, 3] corpora

Number of corpora (languages).

[, 4] texts

Number of texts.

[, 5] size

Total file size in kilobytes.

See Also

multicast, mc_metadata, mc_referents, mc_clauses

Examples

## Not run: 
  # retrieve version index
  mc_index()

## End(Not run)

Access the Multi-CAST metadata

Description

mc_metadata downloads a table with metadata on the texts and speakers in the Multi-CAST collection from the servers of the University of Bamberg.

Usage

mc_metadata(vkey = NULL)

Arguments

vkey

A four-digit number specifying the requested version of the metadata. Must be one of the version keys listed in the first column of mc_index, or empty. If empty, the most recent version of the metadata is retrieved automatically.

Value

A data.frame containing metadata on the Multi-CAST collection. The table has the following eight columns:

[, 1] corpus

The name of the corpus.

[, 2] text

The title of the text.

[, 3] type

The text type, either TN 'traditional narrative', AN 'autobiographical narrative', or SN 'stimulus-based narrative'.

[, 4] recorded

The year (YYYY) the text was recorded.

[, 5] speaker

The identifier for the speaker.

[, 6] gender

The speaker's gender.

[, 7] age

The speaker's age at the time of recording. Approximate values are prefixed with a c.

[, 8] born

The speaker's birth year (YYY). Approximate values are prefixed with a c.

See Also

multicast, mc_index, mc_referents, mc_clauses

Examples

## Not run: 
  # retrieve the most recent version of the Multi-CAST metadata
  mc_metadata()

  # retrieve the lists of referents published in January 2021
  mc_metadata(2101)

  # join the metadata to a table with annotation values
  mc <- multicast()
  merge(mc, mc_metadata(),
        by = c("corpus", "text"))

## End(Not run)

Access the Multi-CAST list of referents

Description

mc_referents downloads a tabular list of all discourse referents occuring in those texts in the Multi-CAST collection that have been annotated with the RefIND scheme (Schiborr et al. 2018). The data are downloaded from the servers of University of Bamberg.

Usage

mc_referents(vkey = NULL)

Arguments

vkey

A four-digit number specifying the requested version of the list of referents. Must be one of the version keys listed in the first column of mc_index, or empty. If empty, the most recent version of the list of referents is retrieved automatically. Note that the first annotations with RefIND were added with version 1905 (May 2019) of Multi-CAST, and hence no lists of referents exist for earlier versions (i.e. 1505 and 1606).

Value

A data.frame containing a list of referents for all texts with RefIND annotations in the Multi-CAST collection. The table has the following eight columns:

[, 1] corpus

The name of the corpus.

[, 2] text

The title of the text.

[, 3] refind

The four-digit referent index, unique to each referent in a text.

[, 4] label

The label used for the referent.

[, 5] description

A short description of the referent.

[, 6] class

The semantic class of the referent. Legend: hum = human, anm = animate, inm = inanimate, bdp = body part, mss = mass, loc = location, tme = time, abs = abstract.

[, 7] relations

Relations of the referent to other referents. Legend: < = set member of (partial co-reference), > = includes (split antecedence), M = part-whole.

[, 8] notes

Annotators' notes on the referent and its properties.

See Also

multicast, mc_index, mc_metadata, mc_clauses

Examples

## Not run: 
  # retrieve the most recent version of the Multi-CAST list of referents
  mc_referents()

  # retrieve the lists of referents published in January 2021
  mc_referents(2021)

  # join the list of referents to a table with annotation values
  mc <- multicast()
  merge(mc, mc_referents(),
        by = c("corpus", "text", "refind"),
        all.x = TRUE)

## End(Not run)

Access Multi-CAST annotation data

Description

multicast downloads corpus data from the Multi-CAST collection (Haig & Schnell 2015) from the servers of the University of Bamberg. As the Multi-CAST collection is continuously evolving through the addition of further data sets and the revision of older annotations, the multicast function takes an optional argument vkey to select earlier versions of the annotation data, ensuring scientific accountability and the reproducibility of results.

Usage

multicast(vkey = NULL)

Arguments

vkey

A four-digit number specifying the requested version of the metadata. Must be one of the version keys listed in the first column of mc_index, or empty. If empty, the most recent version of the metadata is retrieved automatically.

Value

A data.frame with eleven columns:

[, 1] corpus

The name of the corpus.

[, 2] text

The name of the text.

[, 3] uid

The utterance identifier. Uniquely identifies an utterance within a text.

[, 4] gword

Grammatical words. The tokenized utterances in the object language.

[, 5] gloss

Morphological glosses following the Leipzig Glossing Rules.

[, 6] graid

Annotations with the GRAID scheme (Haig & Schnell 2014).

[, 7] gform

The form symbol of a GRAID gloss.

[, 8] ganim

The person-animacy symbol of a GRAID gloss.

[, 9] gfunc

The function symbol of a GRAID gloss.

[, 10] refind

Referent tracking using the RefIND scheme (Schiborr et al. 2018).

[, 11] isnref

Annotations of the information status of newly introduced referents.

Licensing

The Multi-CAST annotation data accessed by this package are published under a Create Commons Attribution 4.0 International (CC-BY 4.0) licence (https://creativecommons.org/licenses/by-sa/4.0/). Please refer to the Multi-CAST website for information on how to give proper credit to its contributors.

Citing Multi-CAST

Data from the Multi-CAST collection should be cited as:

If for some reason you need to cite this package specifically, please refer to citation(multicastR).

References

See Also

mc_index, mc_metadata, mc_referents, mc_clauses

Examples

## Not run: 
  # retrieve and print the most recent version of the
  # Multi-CAST annotations
  multicast()

  # retrieve the version of the annotation data published
  # in January 2021
  multicast(2021)

## End(Not run)

multicastR: A companion to the Multi-CAST collection.

Description

The multicastR package provides a basic interface for accessing the annotated corpus data in the Multi-CAST collection (edited by Geoffrey Haig and Stefan Schnell), a database of spoken natural language texts that draws from a diverse set of languages. The corpus data are downloaded on command from the servers of the University of Bamberg via the multicast method. Details on the Multi-CAST project and a list of publications can be found online at https://multicast.aspra.uni-bamberg.de/.

Licensing

The Multi-CAST annotation data accessed by this package are published under a Create Commons Attribution 4.0 International (CC-BY 4.0) licence (https://creativecommons.org/licenses/by-sa/4.0/). Please refer to the Multi-CAST website for information on how to give proper credit to its contributors.

Citing Multi-CAST

Data from the Multi-CAST collection should be cited as:

If for some reason you need to cite this package specifically, please refer to citation(multicastR).

See Also

multicast, mc_index, mc_metadata, mc_referents, mc_clauses