def test_rule(self):
in_samples = [
clean_sexp("""< (?* [副詞 * * * * ((数量相対名詞修飾))] ) >"""),
clean_sexp("""< (?*) ((ダミー1 ^ダミー2)) >"""),
clean_sexp("""< (?* [* * * * * ((付属))]) >"""),
clean_sexp("""< (?* [* * * * * ((付属))]*) >"""),
]
gold_fc_strs = [
"None",
"""FC<SI=['ダミー1'], SNI=['ダミー2']>""",
"None",
"None",
]
gold_nums = [
2,
1,
2,
2,
]
for intext, gold_fc_str, gold_num in zip(in_samples, gold_fc_strs,
gold_nums):
p = sexpdata.loads(intext)
tr = TagRule(p)
self.assertEqual(gold_fc_str, str(tr.feature_constraints))
self.assertEqual(gold_num, len(tr.rules))
def test_rules(self):
tags_rule_text = clean_sexp(
"""( ?* ^< (?*) ((ダミーA)) > * < (?*) ((^ダミーB)) >? ?* )""")
p = sexpdata.loads(tags_rule_text)
for order in [True, False]:
tagr = TagsRule(p, order)
self.assertEqual(True, len(str(tagr)) > 0)
def test_any(self):
sample = clean_sexp(""" < (?* [副詞 * * * * ((数量相対名詞修飾))] ) > """)
in_parts = [""" ( ?* ) """, sample]
golds = [True, False]
for intext, gold in zip(in_parts, golds):
p = sexpdata.loads(intext)
self.assertEqual(gold, TagRule.is_any(p))
def __init__(self, text: str):
self.text = text
_s = sexpdata.loads(clean_sexp(text))
assert len(_s) >= 4
self._center_len = 1
self.prev_constraints: Optional[Any] = None
if not self.RuleClass.is_any(_s[0]):
self.prev_constraints = self.SeqRuleClass(_s[0], False)
self.next_constraints: Optional[Any] = None
if not self.RuleClass.is_any(_s[2]):
self.next_constraints = self.SeqRuleClass(_s[2], True)
self.constraints: Optional[Any] = None
if not self.RuleClass.is_any(_s[1]):
self.constraints = self.SeqRuleClass(_s[1], True)
self._center_len = len(self.constraints)
self.operations: List[Tuple[bool, str]] = []
for v in _s[3:]:
feature = sexpdata.dumps(v)
remove = False
if feature.startswith("^"):
remove = True
feature = feature[1:]
self.operations.append((remove, feature))
def test_match(self):
rule_text = clean_sexp("""< (?* [判定詞] ) ((ダミー1 ^ダミー2)) >""")
tr = TagRule(sexpdata.loads(rule_text))
mytest = "無停止だ"
mlist: MList = self.jumanpp.get(mytest)
tags = Tags(mlist, [0])
tags[0].fs.add("ダミー0")
self.assertEqual(False, tr.match(tags, 0))
tags[0].fs.add("ダミー1")
self.assertEqual(True, tr.match(tags, 0))
rule_text2 = clean_sexp("""< (?* [接頭辞] ?*) >""")
tr2 = TagRule(sexpdata.loads(rule_text2))
self.assertEqual(True, tr2.match(tags, 0))
rule_text4 = clean_sexp("""< (?* [接頭辞] ) ((ダミー9)) >""")
tr4 = TagRule(sexpdata.loads(rule_text4))
self.assertEqual(False, tr4.match(tags, 0))
rule_text3 = clean_sexp("""< (?* [非マッチ] ) >""")
tr3 = TagRule(sexpdata.loads(rule_text3))
self.assertEqual(False, tr3.match(tags, 0))