def realData():
p = MyParser('../train.wtag')
words = p.getAllPairTagsCombinations()
f104 = F104Builder(words)
firstSent = [w for (w, t) in p.splitted[0]]
history = History("RB", "VBG", firstSent, 3)
assert f104.getFeatureVector(history, "bla").size == 0
assert f104.getFeatureVector(history, "RP").size == 1
def __init__(self, parser: MyParser, offset) -> None:
self.parser = parser
self.f100 = F100Builder(parser.getWordsWithTag(), 0)
self.f103 = F103Builder(parser.getAllThreeTagsCombinations(),
self.f100.size)
self.f104 = F104Builder(parser.getAllPairTagsCombinations(),
self.f100.size + self.f103.size)
super().__init__(self.f100.size + self.f103.size + self.f104.size,
offset)
def __init__(self, train_parser: MyParser, isTraining) -> None:
self.parser = train_parser
self.isTraining = isTraining
vecSize = 0
self.f100 = F100Builder(train_parser.getWordsWithTag(), vecSize)
vecSize = self.f100.size
print("F100 size", self.f100.size)
self.f103 = F103Builder(train_parser.getAllThreeTagsCombinations(),
vecSize)
vecSize = vecSize + self.f103.size
print("F103 size", self.f103.size)
self.f104 = F104Builder(train_parser.getAllPairTagsCombinations(),
vecSize)
vecSize = vecSize + self.f104.size
print("F104 size", self.f104.size)
self.f106 = F106Builder(train_parser.getUniqueTags(), vecSize)
vecSize = vecSize + self.f106.size
print("F106 size", self.f106.size)
self.fSuf = SuffixFeatureBuilder(train_parser, vecSize)
vecSize = vecSize + self.fSuf.size
print("Suffix size", self.fSuf.size)
self.fPref = PrefixFeatureBuilder(train_parser, vecSize)
vecSize = vecSize + self.fPref.size
print("Prefix size", self.fPref.size)
self.fDigNum = DigitNumberFeatureBuilder(train_parser, vecSize)
vecSize = vecSize + self.fDigNum.size
print("DigitNum size", self.fDigNum.size)
self.fLetNum = DigitWordFeatureBuilder(train_parser, vecSize)
vecSize = vecSize + self.fLetNum.size
print("DigitLetter size", self.fLetNum.size)
self.fCaps = CapsFeatureBuilder(train_parser, vecSize)
vecSize = vecSize + self.fCaps.size
print("Caps size", self.fCaps.size)
self.fPrevNext = PrevNextWordFeatureBuilder(
train_parser.getAllPrevWordTagCombinations(),
train_parser.getAllNextWordTagCombinations(), vecSize)
vecSize = vecSize + self.fPrevNext.size
print("PrevNext size", self.fPrevNext.size)
super().__init__(vecSize, 0)
from MyParser import MyParser
p = MyParser("../train.wtag")
words = p.getWordsWithTag()
tag3 = p.getAllThreeTagsCombinations()
tag2 = p.getAllPairTagsCombinations()
tag = p.getUniqueTags()
# print(tag3)
# print(tag2)
print(tag)
