def test_word_segmentation_space_at_end():
s = [("a", 1, 1, 1, 1), ("b", 2, 1, 2, 1), (" ", 3, 1, 5, 1)]
x = Segmentation.word(s, " ", include_spaces=True)
e = [("ab", 1, 1, 2, 1), (" ", 3, 1, 5, 1)]
assert x == e
x = Segmentation.word(s, " ", include_spaces=False)
e = [("ab", 1, 1, 2, 1)]
assert x == e
def test_word_segmentation_space_at_beginning():
s = [(" ", 1, 1, 2, 1), ("b", 3, 1, 3, 1), ("c", 4, 1, 5, 1)]
x = Segmentation.word(s, " ", include_spaces=True)
e = [(" ", 1, 1, 2, 1), ("bc", 3, 1, 5, 1)]
assert x == e
x = Segmentation.word(s, " ", include_spaces=False)
e = [("bc", 3, 1, 5, 1)]
assert x == e
def test_word_segmentation_without_spaces():
s = [
("a", 1, 1, 2, 10),
("b", 3, 1, 3, 10),
(" ", 4, 1, 5, 10),
("c", 6, 1, 800, 10),
]
x = Segmentation.word(s, " ", include_spaces=False)
e = [("ab", 1, 1, 3, 10), ("c", 6, 1, 800, 10)]
assert x == e
def test_char_segmentation_empty():
with pytest.raises(AssertionError):
Segmentation.char([""], [], 1)
def test_word_segmentation_one_element():
x = Segmentation.word([("a", 1, 1, 10, 10)], " ", include_spaces=True)
assert x == [("a", 1, 1, 10, 10)]
def test_word_segmentation_raises(s):
with pytest.raises(AssertionError):
Segmentation.word(s, " ")
def test_word_segmentation_empty():
x = Segmentation.word([], " ", include_spaces=True)
assert x == []
def test_char_segmentation_scaling_error():
with pytest.raises(AssertionError):
Segmentation.char(["a"], [0, 1, 100], 1, width=50)
def test_char_segmentation_scaling():
txt = ["a", "b", "c"]
seg = [0, 3, 5, 7, 10]
x = Segmentation.char(txt, seg, 1, width=100)
e = [("a", 1, 1, 29, 1), ("b", 30, 1, 49, 1), ("c", 50, 1, 69, 1)]
assert x == e
def test_char_segmentation():
txt = ["a", "b", "c"]
seg = [0, 3, 5, 7, 10]
x = Segmentation.char(txt, seg, 1)
e = [("a", 1, 1, 2, 1), ("b", 3, 1, 4, 1), ("c", 5, 1, 6, 1)]
assert x == e