Class: WLAPI::API

Inherits:
Object
  • Object
show all
Includes:
REXML
Defined in:
lib/wlapi/api.rb

Overview

This class represents an interface to the linguistic web services provided by the University of Leipzig.

See the project ‘Wortschatz Leipzig’ for more details.

Constant Summary collapse

ENDPOINT =

SOAP Services Endpoint.

'http://wortschatz.uni-leipzig.de/axis/services'
SERVICES =

The list of accessible services, the MARSService is excluded due to its internal authorization.

[
  :Baseform, :Cooccurrences, :CooccurrencesAll,
  :ExperimentalSynonyms, :Frequencies,
  :Kookurrenzschnitt, :Kreuzwortraetsel,
  :LeftCollocationFinder, :LeftNeighbours, :NGrams,
  :NGramReferences, :RightCollocationFinder,
  :RightNeighbours, :Sachgebiet, :Sentences,
  :Similarity, :Synonyms, :Thesaurus, :Wordforms
]

Instance Method Summary collapse

Constructor Details

#initialize(login = 'anonymous', pass = 'anonymous') ⇒ API

At the creation point clients for all services are being instantiated. You can also set the login and the password (it defaults to ‘anonymous’).

api = WLAPI::API.new



44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
 
# File 'lib/wlapi/api.rb', line 44

def initialize( = 'anonymous', pass = 'anonymous')

  # This hash contains the URLs to the single services.
  services = {}

  SERVICES.each { |service| services[service] = "#{ENDPOINT}/#{service}"}

  # cl short for client.
  # Dynamically create all the clients and set access credentials.
  # It can be a very bad idea to instantiate all the clients at once,
  # we should investigate the typical user behaviour.
  # If only one service is used in the separate session => rewrite the class!
  services.each do |key, val|
    cl_name = "@cl_#{key}"

    options = {:wsdl => val + '?wsdl',
               :namespaces => {'xmlns:dat' => 'http://datatypes.webservice.wortschatz.uni_leipzig.de',
                               'xmlns:urn' => val},
               :basic_auth => ['anonymous', 'anonymous'],
               :log => $DEBUG
              }

    client = Savon.client(options)

    eval("#{cl_name} = client")
  end

  # Savon creates very verbose logs, switching off.
  HTTPI.log = false unless $DEBUG
end

Instance Method Details

#baseform(word) ⇒ Array

Gets the baseform (whatever it is :), not lemma). Returns the lemmatized (base) form of the input word and the POS tag in an array:

api.baseform("Auto") => ["Auto", "N"]

Returns:

  • (Array) —

    a list



105
106
107
108
109
110
111
112
 
# File 'lib/wlapi/api.rb', line 105

def baseform(word)
  check_params(word)

  arg1 = ['Wort', word]
  answer = query(@cl_Baseform, arg1)

  get_answer(answer)
end

#check_params(*args) ⇒ Object (private) 



434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
 
# File 'lib/wlapi/api.rb', line 434

def check_params(*args)
  m = caller(1).first.match(/^.+`(.*)'$/)[1]
  num_of_args = self.method(m.to_sym).arity
  messages = []

  # Arity can be negativ => .abs.
  case num_of_args.abs
  when 1
    messages << msg(args[0], m, 'String') unless args[0].is_a?(String)
  when 2
    messages << msg(args[0], m, 'String') unless args[0].is_a?(String)
    messages << msg(args[1], m, 'Numeric') unless args[1].is_a?(Fixnum)
  when 3
    messages << msg(args[0], m, 'String') unless args[0].is_a?(String)
    unless args[1].is_a?(String) || args[1].is_a?(Fixnum)
      messages << msg(args[1], m, 'String or Numeric')
    end
    messages << msg(args[2], m, 'Numeric') unless args[2].is_a?(Fixnum)
  end

  if messages.any?
    fail WLAPI::UserError, messages.join("\n")
  end
end

#cooccurrences(word, sign, limit = 10) ⇒ Object

Returns statistically significant co-occurrences of the input word.



312
313
314
315
316
317
318
319
320
321
 
# File 'lib/wlapi/api.rb', line 312

def cooccurrences(word, sign, limit = 10)
  check_params(word, sign, limit)

  arg1 = ['Wort', word]
  arg2 = ['Mindestsignifikanz', sign]
  arg3 = ['Limit', limit]
  answer = query(@cl_Cooccurrences, arg1, arg2, arg3)

  get_answer(answer)
end

#cooccurrences_all(word, sign, limit = 10) ⇒ Object

Returns statistically significant co-occurrences of the input word. However, it searches in the unrestricted version of the co-occurrences table as in the Cooccurrences services, which means significantly longer wait times.



327
328
329
330
331
332
333
334
335
336
 
# File 'lib/wlapi/api.rb', line 327

def cooccurrences_all(word, sign, limit = 10)
  check_params(word, sign, limit)

  arg1 = ['Wort', word]
  arg2 = ['Mindestsignifikanz', sign]
  arg3 = ['Limit', limit]
  answer = query(@cl_CooccurrencesAll, arg1, arg2, arg3)

  get_answer(answer)
end

#crossword(word, word_length, limit = 10) ⇒ Object

Attempts to find suitable words given a pattern as word parameter, a word length and the number of words to find at max (limit), e.g. API#crossword('%uto', 4) would return find 24 results and return them as an array: [Auto, Auto, ...]:

api.crossword('%uto') => ["Auto", "Auto", ...]

SQL like syntax is used for pattern (% for an arbitrary string, _ for a single character).

Note: Umlaute will count as one character

– Let’s keep all public method names in English: kreuzwortraetsel => crossword.



369
370
371
372
373
374
375
376
377
378
 
# File 'lib/wlapi/api.rb', line 369

def crossword(word, word_length, limit = 10)
  check_params(word, word_length, limit)

  arg1 = ['Wort', word ]
  arg2 = ['Wortlaenge', word_length]
  arg3 = ['Limit', limit]
  answer = query(@cl_Kreuzwortraetsel, arg1, arg2, arg3)

  get_answer(answer)
end

#domain(word) ⇒ Object

Returns categories for a given input word as an array:

api.domain("Michael") => ["Vorname", "Nachname", "Männername"]

– Is it a good name? all names are in English, but here.. let’s call it domain, not sachgebiet



119
120
121
122
123
124
125
126
 
# File 'lib/wlapi/api.rb', line 119

def domain(word)
  check_params(word)

  arg1 = ['Wort', word]
  answer = query(@cl_Sachgebiet, arg1)

  get_answer(answer)
end

#experimental_synonyms(word, limit = 10) ⇒ Object

This service delivers an experimental synonyms request for internal tests. – don’t know, if we have to include this service…



244
245
246
247
248
249
250
251
252
 
# File 'lib/wlapi/api.rb', line 244

def experimental_synonyms(word, limit = 10)
  check_params(word, limit)

  arg1 = ['Wort', word]
  arg2 = ['Limit', limit]
  answer = query(@cl_ExperimentalSynonyms, arg1, arg2)

  get_answer(answer)
end

#frequencies(word) ⇒ Array

Returns the frequency and frequency class of the input word. Frequency class is computed in relation to the most frequent word in the corpus. The higher the class, the rarer the word:

api.frequencies("Autos") => ["40614", "9"]

Returns:

  • (Array) —

    a list



91
92
93
94
95
96
97
98
 
# File 'lib/wlapi/api.rb', line 91

def frequencies(word)
  check_params(word)

  arg1 = ['Wort', word]
  answer = query(@cl_Frequencies, arg1)

  get_answer(answer)
end

#get_answer(doc, mod = '') ⇒ Object (private)

This method extracts valuable data from the XML structure of the soap response. It returns an array with extracted xml text nodes or nil, if the service provided no answer.



421
422
423
424
425
426
427
428
429
430
431
432
 
# File 'lib/wlapi/api.rb', line 421

def get_answer(doc, mod='')
  result = []
  # The path seems to be weird, because the namespaces change incrementally
  # in the output, so I don't want to treat it here.
  # A modifier needed because synonyms service provides duplicate values.
  XPath.each(doc, "//result/*/*#{mod}") do |el|
    warn(el.text) if $DEBUG
    result << el.text
  end

  result.any? ? result : nil
end

#intersection(word1, word2, limit = 10) ⇒ Object

Returns the intersection of the co-occurrences of the two given words. The result set is ordered according to the sum of the significances in descending order. Note that due to the join involved, this make take some time. – let’s call it intersection, not kookurrenzschnitt is being used INTERN, we need additional credentials



345
346
347
348
349
350
351
352
353
354
 
# File 'lib/wlapi/api.rb', line 345

def intersection(word1, word2, limit = 10)
  check_params(word1, word2, limit)

  arg1 = ['Wort 1', word1]
  arg2 = ['Wort 2', word2]
  arg3 = ['Limit', limit]
  answer = query(@cl_Kookurrenzschnitt, arg1, arg2, arg3)

  get_answer(answer)
end

#left_collocation_finder(word, pos, limit = 10) ⇒ Object

Attempts to find linguistic collocations that occur to the left of the given input word. The parameter ‘Wortart’ accepts four values ‘A, V, N, S’ which stand for adjective, verb, noun and stopword respectively. The parameter restricts the type of words found. It returns an array:

api.left_collocation_finder("Stuhl", "A", 10) =>
["apostolisch", "A", "Stuhl", ...]



300
301
302
303
304
305
306
307
308
309
 
# File 'lib/wlapi/api.rb', line 300

def left_collocation_finder(word, pos, limit = 10)
  check_params(word, pos, limit)

  arg1 = ['Wort', word]
  arg2 = ['Wortart', pos]
  arg3 = ['Limit', limit]
  answer = query(@cl_LeftCollocationFinder, arg1, arg2, arg3)

  get_answer(answer)
end

#left_neighbours(word, limit = 10) ⇒ Object

TODO:

ok, but results should be filtered

For a given input word, returns statistically significant left neighbours (words co-occurring immediately to the left of the input word).

api.left_neighbours("Auto") => ["geparktes", "Auto", "561", ...]

–



198
199
200
201
202
203
204
205
206
 
# File 'lib/wlapi/api.rb', line 198

def left_neighbours(word, limit = 10)
  check_params(word, limit)

  arg1 = ['Wort', word]
  arg2 = ['Limit', limit]
  answer = query(@cl_LeftNeighbours, arg1, arg2)

  get_answer(answer)
end

#msg(arg, meth, cls) ⇒ Object (private) 



459
460
461
462
 
# File 'lib/wlapi/api.rb', line 459

def msg(arg, meth, cls)
  "Argument <#{arg}> for the method <#{meth}> should be a <#{cls}>, "\
  "not <#{arg.class}>!"
end

#ngram_references(pattern, limit = 10) ⇒ Object 



263
264
265
266
267
268
 
# File 'lib/wlapi/api.rb', line 263

def ngram_references(pattern, limit = 10)
  arg1 = ['Pattern', pattern]
  arg2 = ['Limit', limit]
  answer = query(@cl_NGramReferences, arg1, arg2)
# raise(NotImplementedError, 'This method will be implemented in the next release.')
end

#ngrams(pattern, limit = 10) ⇒ Object 



255
256
257
258
259
260
 
# File 'lib/wlapi/api.rb', line 255

def ngrams(pattern, limit = 10)
  arg1 = ['Pattern', pattern]
  arg2 = ['Limit', limit]
  answer = query(@cl_NGrams, arg1, arg2)
# raise(NotImplementedError, 'This method will be implemented in the next release.')
end

#query(cl, *args) ⇒ Object (private)

Main query method, it invokes the soap engine. It combines all the data to one SOAP request and gets the answer. <args> contains an array [[key1, value1], [key2, value2], [key3, value3]] with keys and values for the soap query.



386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
 
# File 'lib/wlapi/api.rb', line 386

def query(cl, *args)
  # WSDL is disabled since calling the server for wsdl can last too long.
  v = []
  body = {
    'urn:objRequestParameters' => {
      'urn:corpus' => 'de',
      'urn:parameters' => {
        'urn:dataVectors' => v
      }
    }
  }

  # _args_ is an Array of arrays with keys and values
  # Setting the first argument (usually 'Wort').
  # Setting the second argument (usually 'Limit').
  # Setting the third argument (no common value).
  args.each do |key, val|
    v << {'dat:dataRow' => [key, val]}
  end

  begin
    resp = cl.call(:execute, {:message => body})
  rescue => e
    raise(WLAPI::ExternalError, e)
  end

  doc = Document.new(resp.to_xml)
  warn(doc) if $DEBUG

  doc
end

#right_collocation_finder(word, pos, limit = 10) ⇒ Object

Attempts to find linguistic collocations that occur to the right of the given input word. The parameter ‘Wortart’ accepts four values ‘A, V, N, S’ which stand for adjective, verb, noun and stopword respectively. The parameter restricts the type of words found. It returns an array:

api.right_collocation_finder("Auto", "V", 10) =>
["Auto", "abfackeln", "V", ...]



281
282
283
284
285
286
287
288
289
290
 
# File 'lib/wlapi/api.rb', line 281

def right_collocation_finder(word, pos, limit = 10)
  check_params(word, pos, limit)

  arg1 = ['Wort', word]
  arg2 = ['Wortart', pos]
  arg3 = ['Limit', limit]
  answer = query(@cl_RightCollocationFinder, arg1, arg2, arg3)

  get_answer(answer)
end

#right_neighbours(word, limit = 10) ⇒ Object

TODO:

ok, but results should be filtered

For a given input word, returns statistically significant right neighbours (words co-occurring immediately to the right of the input word).

api.right_neighbours("Auto") => ["Auto", "erfaßt", "575", ...]

–



213
214
215
216
217
218
219
220
221
 
# File 'lib/wlapi/api.rb', line 213

def right_neighbours(word, limit = 10)
  check_params(word, limit)

  arg1 = ['Wort', word]
  arg2 = ['Limit', limit]
  answer = query(@cl_RightNeighbours, arg1, arg2)

  get_answer(answer)
end

#sentences(word, limit = 10) ⇒ Object

TODO:

ok, but results should be filtered

Returns sample sentences containing the input word. The return value is an array:

api.sentences("Auto") => ["40808144", "Zweitens der freche,
frische Klang der Hupe
und drittens die hinreißend gestylten 16-Zoll-Felgen,
die es leider nur für dieses Auto gibt.", ...]

–



183
184
185
186
187
188
189
190
191
 
# File 'lib/wlapi/api.rb', line 183

def sentences(word, limit = 10)
  check_params(word, limit)

  arg1 = ['Wort', word]
  arg2 = ['Limit', limit]
  answer = query(@cl_Sentences, arg1, arg2)

  get_answer(answer)
end

#similarity(word, limit = 10) ⇒ Object

Returns automatically computed contextually similar words of the input word. Such similar words may be antonyms, hyperonyms, synonyms, cohyponyms or other. Note that due to the huge amount of data any query to this services may take a long time.

api.similarity("Auto") => ["Auto", "Wagen", "26", ...]



231
232
233
234
235
236
237
238
239
 
# File 'lib/wlapi/api.rb', line 231

def similarity(word, limit = 10)
  check_params(word, limit)

  arg1 = ['Wort', word]
  arg2 = ['Limit', limit]
  answer = query(@cl_Similarity, arg1, arg2)

  get_answer(answer)
end

#synonyms(word, limit = 10) ⇒ Array

This method searches for synonyms. Returns synonyms of the input word. In other words, this is a thesaurus.

api.synonyms("Auto") => ["Kraftwagen", "Automobil", "Benzinkutsche", ...]

Returns:

  • (Array) —

    a list



164
165
166
167
168
169
170
171
172
173
 
# File 'lib/wlapi/api.rb', line 164

def synonyms(word, limit = 10)
  check_params(word, limit)

  arg1 = ['Wort', word]
  arg2 = ['Limit', limit]
  answer = query(@cl_Synonyms, arg1, arg2)

  # Synonym service provides multiple values, so we take only odd.
  get_answer(answer, '[position() mod 2 = 1 ]')
end

#thesaurus(word, limit = 10) ⇒ Array

As the Synonyms service returns synonyms of the given input word. However, this first builds a lemma of the input word and thus returns more synonyms:

api.thesaurus("Auto") => ["Auto", "Bahn", "Wagen", "Zug", "Schiff", ...]

Returns:

  • (Array) —

    a list



150
151
152
153
154
155
156
157
158
 
# File 'lib/wlapi/api.rb', line 150

def thesaurus(word, limit = 10)
  check_params(word, limit)

  arg1 = ['Wort', word]
  arg2 = ['Limit', limit]
  answer = query(@cl_Thesaurus, arg1, arg2)

  get_answer(answer)
end

#wordforms(word, limit = 10) ⇒ Array

Returns all other word forms of the same lemma for a given word form.

api.wordforms("Auto") => ["Auto", "Autos"]

Returns:

  • (Array) —

    a list



134
135
136
137
138
139
140
141
142
143
 
# File 'lib/wlapi/api.rb', line 134

def wordforms(word, limit = 10)
  check_params(word, limit)

  # note, it is the only service which requires 'Word', not 'Wort'
  arg1 = ['Word', word]
  arg2 = ['Limit', limit]
  answer = query(@cl_Wordforms, arg1, arg2)

  get_answer(answer)
end