Structural annotations (s-attributes) of a corpus capture metainformation for regions of tokens. The s_attributes()-method offers high-level access to the s-attributes present in a corpus or subcorpus, or the values of s-attributes in a corpus/partition.

s_attributes(.Object, ...)

# S4 method for class 'character'
s_attributes(.Object, s_attribute = NULL, unique = TRUE, regex = NULL, ...)

# S4 method for class 'corpus'
s_attributes(.Object, s_attribute = NULL, unique = TRUE, regex = NULL, ...)

# S4 method for class 'slice'
s_attributes(.Object, s_attribute = NULL, unique = TRUE, ...)

# S4 method for class 'partition'
s_attributes(.Object, s_attribute = NULL, unique = TRUE, ...)

# S4 method for class 'subcorpus'
s_attributes(.Object, s_attribute = NULL, unique = TRUE, ...)

# S4 method for class 'context'
s_attributes(.Object, s_attribute = NULL)

# S4 method for class 'partition_bundle'
s_attributes(.Object, s_attribute, unique = TRUE, ...)

# S4 method for class 'call'
s_attributes(.Object, corpus)

# S4 method for class 'quosure'
s_attributes(.Object, corpus)

# S4 method for class 'name'
s_attributes(.Object, corpus)

# S4 method for class 'remote_corpus'
s_attributes(.Object, ...)

# S4 method for class 'remote_partition'
s_attributes(.Object, ...)

# S4 method for class 'data.table'
s_attributes(.Object, corpus, s_attribute, registry)

Arguments

.Object

A corpus, subcorpus, partition object, or a call. A corpus can also be specified by a length-one character vector.

...

To maintain backward compatibility, if argument sAttribute (deprecated) is used. If .Object is a remote_corpus or remote_subcorpus 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_attribute

The name of a specific s-attribute.

unique

A logical value, whether to return unique values.

regex

A regular expression passed into grep to filter return value by applying a regex.

corpus

A corpus-object or a length one character vector denoting a corpus.

registry

The registry directory with the registry file defining corpus. If missing, the registry directory that can be derived using RcppCWB::corpus_registry_dir() is used.

Value

A character vector (s-attributes, or values of s-attributes).

Details

Importing XML into the Corpus Workbench (CWB) turns elements and element attributes into so-called "s-attributes". There are two basic uses of the s_attributes()-method: If the argument s_attribute is NULL (default), the return value is a character vector with all s-attributes present in a corpus.

If s_attribute denotes a specific s-attribute (a length-one character vector), the values of the s-attributes available in the corpus/partition are returned. if the s-attribute does not have values, NA is returned and a warning message is issued.

If argument unique is FALSE, the full sequence of the s_attributes is returned, which is a useful building block for decoding a corpus.

If argument s_attributes is a character providing several s-attributes, the method will return a data.table. If unique is TRUE, all unique combinations of the s-attributes will be reported by the data.table.

If the corpus is based on a nested XML structure, the order of items on the s_attribute vector matters. The method for corpus objects will take the first s-attribute as the benchmark and assume that further s-attributes are XML ancestors of the node.

If .Object is a context object, the s-attribute value for the first corpus position of every match is returned in a character vector. If the match is outside a region of the s-attribute, NA is returned.

If .Object is a call or a quosure (defined in the rlang package), the s_attributes-method will return a character vector with the s-attributes occurring in the call. This usage is relevant internally to implement the subset method to generate a subcorpus using non-standard evaluation. Usually it will not be relevant in an interactive session.

Examples

use("polmineR")
#> ✔ corpus loaded: GERMAPARLMINI (version: 0.1.0 | build date: 2023-04-16)

s_attributes("GERMAPARLMINI")
#> [1] "interjection"  "date"          "party"         "speaker"      
#> [5] "protocol_date" "protocol_lp"   "role"         
s_attributes("GERMAPARLMINI", "date") # dates of plenary meetings
#> [1] "2009-10-27" "2009-10-28" "2009-11-10" "2009-11-11" "2009-11-12"
s_attributes("GERMAPARLMINI", s_attribute = c("date", "party"))  
#>           date           party
#>         <char>          <char>
#>  1: 2009-10-27              NA
#>  2: 2009-10-27         CDU_CSU
#>  3: 2009-10-27             SPD
#>  4: 2009-10-27             FDP
#>  5: 2009-10-27       DIE_LINKE
#>  6: 2009-10-27 B90_DIE_GRUENEN
#>  7: 2009-10-28              NA
#>  8: 2009-10-28         CDU_CSU
#>  9: 2009-10-28             FDP
#> 10: 2009-11-10              NA
#> 11: 2009-11-10         CDU_CSU
#> 12: 2009-11-10             SPD
#> 13: 2009-11-10             FDP
#> 14: 2009-11-10       DIE_LINKE
#> 15: 2009-11-10 B90_DIE_GRUENEN
#> 16: 2009-11-11              NA
#> 17: 2009-11-11             FDP
#> 18: 2009-11-11             SPD
#> 19: 2009-11-11         CDU_CSU
#> 20: 2009-11-11       DIE_LINKE
#> 21: 2009-11-11 B90_DIE_GRUENEN
#> 22: 2009-11-12              NA
#> 23: 2009-11-12             FDP
#> 24: 2009-11-12             SPD
#> 25: 2009-11-12         CDU_CSU
#> 26: 2009-11-12 B90_DIE_GRUENEN
#> 27: 2009-11-12       DIE_LINKE
#>           date           party
#>         <char>          <char>
s_attributes(corpus("GERMAPARLMINI"))
#> [1] "interjection"  "date"          "party"         "speaker"      
#> [5] "protocol_date" "protocol_lp"   "role"         
p <- partition("GERMAPARLMINI", date = "2009-11-10")
#> ... get cpos and strucs
s_attributes(p)
#> [1] "interjection"  "date"          "party"         "speaker"      
#> [5] "protocol_date" "protocol_lp"   "role"         
s_attributes(p, "speaker") # get names of speakers
#>  [1] "Norbert Lammert"             "Angela Dorothea Merkel"     
#>  [3] "Frank-Walter Steinmeier"     "Birgit Homburger"           
#>  [5] "Wolfgang Thierse"            "Oskar Lafontaine"           
#>  [7] "Jürgen Trittin"              "Volker Kauder"              
#>  [9] "Joachim Poß"                 "Hans-Peter Friedrich"       
#> [11] "Agnes Krumwiede"             "Gerda Hasselfeldt"          
#> [13] "Arnold Vaatz"                "Guido Westerwelle"          
#> [15] "Gernot Erler"                "Andreas Schockenhoff"       
#> [17] "Jan  van Aken"               "Frithjof Schmidt"           
#> [19] "Dirk Niebel"                 "Sascha Raabe"               
#> [21] "Angelica Schwall-Düren"      "Karl-Theodor  zu Guttenberg"
#> [23] "Wolfgang Gehrcke"            "Omid Nouripour"             
#> [25] "Petra Pau"                   "Ilse Aigner"                
#> [27] "Waltraud Wolff"              "Hans-Michael Goldmann"      
#> [29] "Katrin Göring-Eckardt"       "Kirsten Tackmann"           
#> [31] "Ulrike Höfken"               "Peter Bleser"               
#> [33] "Elvira Drobinski-Weiß"       "Christel Happach-Kasan"     
#> [35] "Caren Lay"                   "Johannes Röring"            
#> [37] "Wilhelm Priesmeier"         

# Get s-attributes occurring in a call
s_attributes(quote(grep("Merkel", speaker)), corpus = "GERMAPARLMINI")
#> character 
#> "speaker" 
s_attributes(quote(speaker == "Angela Merkel"), corpus = "GERMAPARLMINI")
#> character 
#> "speaker" 
s_attributes(quote(speaker != "Angela Merkel"), corpus = "GERMAPARLMINI")
#> character 
#> "speaker" 
s_attributes(
  quote(speaker == "Angela Merkel" & date == "2009-10-28"),
  corpus = "GERMAPARLMINI"
)
#> character character 
#> "speaker"    "date" 

# Get s-attributes from quosure
s_attributes(
  rlang::new_quosure(quote(grep("Merkel", speaker))),
  corpus = "GERMAPARLMINI"
)
#> character 
#> "speaker"