def test_build_lexeme():
template = {
'language_item_id':
'Q1860',
'language_code':
'en',
'lexical_category_item_id':
'Q1084',
'forms': [
{
'grammatical_features_item_ids': ['Q110786'],
},
{
'grammatical_features_item_ids': ['Q146786'],
},
],
}
form_data = werkzeug.datastructures.ImmutableMultiDict([
('form_representation', 'noun'), ('form_representation', 'nouns')
])
lexeme_data = lexeme_forms.build_lexeme(template, form_data)
assert lexeme_data == {
'type':
'lexeme',
'forms': [
{
'add': '',
'representations': {
'en': {
'language': 'en',
'value': 'noun'
}
},
'grammaticalFeatures': ['Q110786'],
'claims': {},
},
{
'add': '',
'representations': {
'en': {
'language': 'en',
'value': 'nouns'
}
},
'grammaticalFeatures': ['Q146786'],
'claims': {},
},
],
'lemmas': {
'en': {
'language': 'en',
'value': 'noun'
}
},
'language':
'Q1860',
'lexicalCategory':
'Q1084',
'claims': {},
}
def test_build_lexeme_blank_form(monkeypatch):
monkeypatch.setattr(lexeme_forms, 'get_lexeme_data', lambda lexeme_id, wiki: {'language': 'Q1860', 'lexicalCategory': 'Q1084'})
template = {
'language_item_id': 'Q1860',
'language_code': 'en',
'lexical_category_item_id': 'Q1084',
'forms': [
{
'grammatical_features_item_ids': ['Q110786'],
},
{
'grammatical_features_item_ids': ['Q146786'],
},
],
}
form_data = werkzeug.datastructures.ImmutableMultiDict([('lexeme_id', 'L123'), ('form_representation', ''), ('form_representation', 'nouns')])
lexeme_data = lexeme_forms.build_lexeme(template, form_data)
assert lexeme_data == {
'type': 'lexeme',
'forms': [
{
'add': '',
'representations': {'en': {'language': 'en', 'value': 'nouns'}},
'grammaticalFeatures': ['Q146786'],
'claims': {},
},
],
'id': 'L123',
'claims': {},
}
def test_integration_edit(template_name):
"""Create a lexeme from a template, then match it against the same template.
If this fails, then there might be a bug when creating or matching lexemes,
or the template might have ambiguous forms."""
template = templates[template_name]
form_data = werkzeug.datastructures.ImmutableMultiDict([
('form_representation', str(index))
for index in range(0, len(template['forms']))
])
lexeme_data = lexeme_forms.build_lexeme(template, form_data)
matched_template = matching.match_lexeme_forms_to_template(
lexeme_data['forms'], template)
assert matched_template.get('ambiguous_lexeme_forms') is None
assert matched_template.get('unmatched_lexeme_forms') is None
for index, form in enumerate(matched_template['forms']):
assert len(form['lexeme_forms']) == 1
assert form['lexeme_forms'][0]['representations'][
template['language_code']]['value'] == str(index)
def test_build_lexeme_with_statements_for_existing_lexeme(monkeypatch):
p5185 = [{
'mainsnak': {
'snaktype': 'value',
'property': 'P5185',
'datatype': 'wikibase-item',
'datavalue': {
'type': 'wikibase-entityid',
'value': {
'entity-type': 'item',
'id': 'Q499327',
},
},
},
'type': 'statement',
'rank': 'normal',
}]
p31 = [{
'mainsnak': {
'snaktype': 'value',
'property': 'P31',
'datatype': 'wikibase-item',
'datavalue': {
'type': 'wikibase-entityid',
'value': {
'entity-type': 'item',
'id': 'Q604984',
},
},
},
'type': 'statement',
'rank': 'normal',
}]
monkeypatch.setattr(
lexeme_forms, 'get_lexeme_data', lambda lexeme_id, wiki: {
'language': 'Q1860',
'lexicalCategory': 'Q1084',
'claims': {
'P5185': p5185,
},
})
template = {
'language_item_id':
'Q1860',
'language_code':
'en',
'lexical_category_item_id':
'Q1084',
'forms': [
{
'grammatical_features_item_ids': ['Q110786'],
'statements': 'singular test statements',
},
{
'grammatical_features_item_ids': ['Q146786'],
'statements': 'plural test statements',
},
],
'statements': {
'P5185': p5185,
'P31': p31,
},
}
form_data = werkzeug.datastructures.ImmutableMultiDict([
('lexeme_id', 'L123'), ('form_representation', 'noun'),
('form_representation', 'nouns')
])
lexeme_data = lexeme_forms.build_lexeme(template, form_data)
assert lexeme_data == {
'id':
'L123',
'type':
'lexeme',
'forms': [
{
'add': '',
'representations': {
'en': {
'language': 'en',
'value': 'noun'
}
},
'grammaticalFeatures': ['Q110786'],
'claims': 'singular test statements',
},
{
'add': '',
'representations': {
'en': {
'language': 'en',
'value': 'nouns'
}
},
'grammaticalFeatures': ['Q146786'],
'claims': 'plural test statements',
},
],
'claims': {
'P31': p31,
},
}
