Basic usage

Synset and Variants

To create a new (empty) synset:

>>> import eurown
>>> a = eurown.Synset()
>>> print a
<eurown.Synset object at 0x80ab10c>
>>> a.polarisText
u'0 WORD_MEANING'
>>> print a.polarisText
0 WORD_MEANING
>>>

Property polarisText returns (unicode) string of Synset in Polaris import-export format.

Synset has part of speech property, that can be one of ‘a’,’b’,’v’,’n’, or pre-defined as ‘pn’ if we have WORD_INSTANCE instead of WORD_MEANING:

>>> a.pos = 'n'
>>> print a.polarisText
0 WORD_MEANING
  1 PART_OF_SPEECH "n"
>>> b = eurown.WordInstance()
>>> print b.polarisText
0 WORD_INSTANCE
  1 PART_OF_SPEECH "pn"

To make some new variants (literal and sense number, gloss for var3 as well):

>>> var1 = eurown.Variant(literal='test',sense=1)
>>> var2 = eurown.Variant(literal='trial',sense=1)
>>> var3 = eurown.Variant(literal='test',sense=2)
>>> var3.gloss = u'This is test'
>>> var4 = eurown.Variant(literal='exam',sense=1)

Let’s assign variants var1 and var2 to synset a:

>>> a.variants = eurown.Variants([var, var2])
>>> print a.polarisText
0 WORD_MEANING
  1 PART_OF_SPEECH "n"
  1 VARIANTS
    2 LITERAL "test"
      3 SENSE 1
    2 LITERAL "trial"
      3 SENSE 1

and make a new synset and assign to it variants var3 and var4:

>>> snset2 = eurown.Synset(pos='n')
>>> snset2.variants = eurown.Variants([var3, var4])
>>> print var3.polarisText
2 LITERAL "test"
  3 SENSE 2
  3 DEFINITION "this is test"

pluss vaiant var5 to append directly to snset2.variants:

>>> snset2.variants.append(eurown.Variant(literal='examination',sense=1))

Now we should have a synset (snset2) with three variants:

>>> print snset2.polarisText
0 WORD_MEANING
  1 PART_OF_SPEECH "n"
  1 VARIANTS
    2 LITERAL "test"
      3 SENSE 2
      3 DEFINITION "this is test"
    2 LITERAL "exam"
      3 SENSE 1
    2 LITERAL "examination"
      3 SENSE 1

Internal Relations

Relation consists of relation name and target concept. Let’s make a synset that would have a relation to our snset2:

>>> snset3 = eurown.Synset(pos='n')
>>> var6 = eurown.Variant(literal="communication",sense=1)
>>> var7 = eurown.Variant(literal="communicating",sense=1)
>>> snset3.variants = eurown.Variants([var6,var7])
>>> print snset3.polarisText
0 WORD_MEANING
  1 PART_OF_SPEECH "n"
    1 VARIANTS
      2 LITERAL "communication"
        3 SENSE 1
      2 LITERAL "communicating"
        3 SENSE 1

Now we can link it to our snset2 via “has_hyperonym” relation:

>>> rel = eurown.Relation(name='has_hyperonym',target_concept=snset3)
>>> snset2.addRelation(rel)
>>> print snset2.polarisText
0 WORD_MEANING
  1 PART_OF_SPEECH "n"
  1 VARIANTS
    2 LITERAL "test"
      3 SENSE 2
      3 DEFINITION "This is test"
    2 LITERAL "exam"
      3 SENSE 1
    2 LITERAL "examination"
      3 SENSE 1
  1 INTERNAL_LINKS
    2 RELATION "has_hyperonym"
      3 TARGET_CONCEPT
        4 PART_OF_SPEECH "n"
        4 LITERAL "communication"
          5 SENSE 1

The same result will give the addRelation() function.

Interlingual Equivalence Links

Interlingual Equivalence link may have

Parsing IO File

Parsing Polaris IO file is done by Parser. At first, we should create an instance of a parser:

>>> p = eurown.Parser()

Parser can get file name:

>>> p.fileName = 'kb59-utf_8.txt'

We can parse one line, one synset, or even one wordnet file at a time.

WordNet

The module can deal with more than one wordnet at a time. While instantiating a wordnet, we should give file name and make all necessary indexes. Making indexes may take time:

>>> wn = eurown.WordNet(name='et', ioFileName='kb59-utf_8.txt')
>>> wn.make_indexes()

Examples

A script that will ask user for a word to find and prints out some basic information (literal, sense, gloss and examples) for each synset:

import eurown

wn = eurown.WordNet(name='et',
                    ioFileName='kb59-utf_8.txt')

wn.make_indexes()

def test_by_literal(literal):
    if literal in wn.literalIndex:
        snset_offsets = wn.literalIndex[literal]
        for i in snset_offsets:
            print i
            p = eurown.Parser(fileName='kb59-utf_8.txt')
            synset = p.parse_synset(offset=i)
            print 5*'='
            for j in synset.variants:
                print '%s_%d' % (j.literal, j.sense)
                print j.gloss
                print j.examples

def show_synsets():
    while 1:
        a = raw_input('otsi: ')
        test_by_literal(a)

show_synsets()