def test_pos_trigram_greek(self):
"""Test tagging Greek POS with trigram tagger."""
tagger = POSTag("grc")
tagged = tagger.tag_trigram(
"θεοὺς μὲν αἰτῶ τῶνδ᾽ ἀπαλλαγὴν πόνων φρουρᾶς ἐτείας μῆκος"
) # pylint: disable=line-too-long
self.assertTrue(tagged)
def test_pos_ngram123_tagger_greek(self):
"""Test tagging Greek POS with a 1-, 2-, and 3-gram backoff tagger."""
tagger = POSTag("grc")
tagged = tagger.tag_ngram_123_backoff(
"θεοὺς μὲν αἰτῶ τῶνδ᾽ ἀπαλλαγὴν πόνων φρουρᾶς ἐτείας μῆκος"
) # pylint: disable=line-too-long
self.assertTrue(tagged)
def test_pos_ngram123_tagger_old_english(self):
"""Test tagging Old English POS with a 1-, 2-, and 3-gram backoff tagger."""
tagger = POSTag("ang")
tagged = tagger.tag_ngram_123_backoff(
"Hwæt! We Gardena in geardagum, þeodcyninga, þrym gefrunon, hu ða æþelingas ellen fremedon."
) # pylint: disable=line-too-long
self.assertTrue(tagged)
def test_pos_perceptron_tagger_old_english(self):
"""Test tagging Old English POS with Perceptron tagger."""
tagger = POSTag("ang")
tagged = tagger.tag_perceptron(
"Hwæt! We Gardena in geardagum, þeodcyninga, þrym gefrunon, hu ða æþelingas ellen fremedon."
)
self.assertTrue(tagged)
def test_pos_ngram12_tagger_middle_low_german(self):
""" Test MOG POS 12-backoff tagger"""
tagger = POSTag("gml")
tagged = tagger.tag_ngram_12_backoff(
"Jck Johannes preister verwarer vnde voirs tender des Juncfrouwen kloisters to Mariendale"
)
self.assertTrue(tagged)
def test_pos_trigram_old_english(self):
"""Test tagging old_english POS with trigram tagger."""
tagger = POSTag("ang")
tagged = tagger.tag_trigram(
"Hwæt! We Gardena in geardagum, þeodcyninga, þrym gefrunon, hu ða æþelingas ellen fremedon."
)
self.assertTrue(tagged)
def test_pos_ngram123_tagger_latin(self):
"""Test tagging Latin POS with a 1-, 2-, and 3-gram backoff tagger."""
tagger = POSTag("lat")
tagged = tagger.tag_ngram_123_backoff(
"Gallia est omnis divisa in partes tres"
) # pylint: disable=line-too-long
self.assertTrue(tagged)
def _retrieve_tag(self, text):
"""Tag text with chosen tagger and clean tags.
Tag format: [('word', 'tag')]
:param text: string
:return: list of tuples, with each tuple containing the word and its pos tag
:rtype : list
"""
if (
self.tagger == "tag_ngram_123_backoff"
): # Data format: Perseus Style (see https://github.com/cltk/latin_treebank_perseus)
tags = POSTag("lat").tag_ngram_123_backoff(text.lower())
return [(tag[0], tag[1]) for tag in tags]
elif self.tagger == "tag_tnt":
tags = POSTag("lat").tag_tnt(text.lower())
return [(tag[0], tag[1]) for tag in tags]
elif self.tagger == "tag_crf":
tags = POSTag("lat").tag_crf(text.lower())
return [(tag[0], tag[1]) for tag in tags]
def test_pos_tnt_tagger_old_norse(self):
"""Test tagging Old Norse POS with TnT tagger."""
tagger = POSTag("non")
tagged = tagger.tag_tnt("Hlióðs bið ek allar.")
print(tagged)
self.assertTrue(tagged)
def test_pos_crf_tagger_latin(self):
"""Test tagging Latin POS with CRF tagger."""
tagger = POSTag("lat")
tagged = tagger.tag_crf("Gallia est omnis divisa in partes tres")
self.assertTrue(tagged)
def test_pos_trigram_latin(self):
"""Test tagging Latin POS with trigram tagger."""
tagger = POSTag("lat")
tagged = tagger.tag_trigram("Gallia est omnis divisa in partes tres")
self.assertTrue(tagged)
def __init__(self):
self.syllabifier = Syllabifier(language="old_norse_ipa")
self.tr = Transcriber(DIPHTHONGS_IPA, DIPHTHONGS_IPA_class, IPA_class,
old_norse_rules)
self.tagger = POSTag("old_norse")