Skip to content

Commit ef1b62b

Browse files
committed
Move test_morpheme
1 parent e5c6fee commit ef1b62b

File tree

3 files changed

+3
-25
lines changed

3 files changed

+3
-25
lines changed

pythainlp/util/normalize.py

Lines changed: 2 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -277,8 +277,7 @@ def expand_maiyamok(sent: Union[str, List[str]]) -> List[str]:
277277
for j, token in enumerate(sent):
278278
if token.isspace() and "ๆ" in sent[j + 1]:
279279
continue
280-
# Replace any number of spaces followed by "ๆ" with "ๆ"
281-
token = re.sub(r'\s+ๆ', 'ๆ', token)
280+
token = re.sub(r"\s+ๆ", "ๆ", token)
282281
if "ๆ" == token:
283282
token = output_toks[i - 1]
284283
elif "ๆ" in token:
@@ -297,7 +296,7 @@ def maiyamok(sent: Union[str, List[str]]) -> List[str]:
297296
"""
298297
Expand Maiyamok.
299298
300-
Use expand_maiyamok() instead.
299+
Deprecated. Use expand_maiyamok() instead.
301300
302301
Maiyamok (ๆ) (Unicode U+0E46) is a Thai character indicating word
303302
repetition. This function preprocesses Thai text by replacing

tests/__init__.py

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -14,6 +14,7 @@
1414
"tests.test_ancient",
1515
# "tests.test_cli",
1616
# "tests.test_corpus",
17+
"tests.test_morpheme",
1718
"tests.test_soundex",
1819
"tests.test_spell",
1920
"tests.test_tokenize",

tests/test_util.py

Lines changed: 0 additions & 22 deletions
Original file line numberDiff line numberDiff line change
@@ -29,7 +29,6 @@
2929
expand_maiyamok,
3030
find_keyword,
3131
ipa_to_rtgs,
32-
is_native_thai,
3332
isthai,
3433
isthaichar,
3534
nectec_to_ipa,
@@ -617,27 +616,6 @@ def test_isthai(self):
617616
self.assertEqual(isthai("ต.ค.", ignore_chars=None), False)
618617
self.assertEqual(isthai("(ต.ค.)", ignore_chars=".()"), True)
619618

620-
def test_is_native_thai(self):
621-
self.assertEqual(is_native_thai(None), False)
622-
self.assertEqual(is_native_thai(""), False)
623-
self.assertEqual(is_native_thai("116"), False)
624-
self.assertEqual(is_native_thai("abc"), False)
625-
self.assertEqual(is_native_thai("ตา"), True)
626-
self.assertEqual(is_native_thai("ยา"), True)
627-
self.assertEqual(is_native_thai("ฆ่า"), True)
628-
self.assertEqual(is_native_thai("คน"), True)
629-
self.assertEqual(is_native_thai("กะ"), True)
630-
self.assertEqual(is_native_thai("มอ"), True)
631-
self.assertEqual(is_native_thai("กะ"), True)
632-
self.assertEqual(is_native_thai("กระ"), True)
633-
self.assertEqual(is_native_thai("ประท้วง"), True)
634-
self.assertEqual(is_native_thai("ศา"), False)
635-
self.assertEqual(is_native_thai("ลักษ์"), False)
636-
self.assertEqual(is_native_thai("มาร์ค"), False)
637-
self.assertEqual(is_native_thai("เลข"), False)
638-
self.assertEqual(is_native_thai("เทเวศน์"), False)
639-
self.assertEqual(is_native_thai("เทเวศร์"), False)
640-
641619
def test_display_thai_char(self):
642620
self.assertEqual(display_thai_char("้"), "_้")
643621
self.assertEqual(display_thai_char("ป"), "ป")

0 commit comments

Comments
 (0)