Morpheme.split method returns empty list

[mh-northlander]

Morpheme.split method seems returns empty morpheme list when the morpheme does not has split information.
It should return a list contains itself in that case.

reproduce:

tok = sudachipy.Dictionary().create()

ms = tok.tokenize("国会議事堂前駅で降りる")
[m.split(mode="a") for m in ms]

outputs:

[<MorphemeList[
  <Morpheme(国会, 0:2, (0, 364210))>,
  <Morpheme(議事, 2:4, (0, 686966))>,
  <Morpheme(堂, 4:5, (0, 368464))>,
  <Morpheme(前, 5:6, (0, 318425))>,
  <Morpheme(駅, 6:7, (0, 755333))>,
]>, <MorphemeList[
]>, <MorphemeList[
]>]

should be:

[<MorphemeList[
  <Morpheme(国会, 0:2, (0, 364210))>,
  <Morpheme(議事, 2:4, (0, 686966))>,
  <Morpheme(堂, 4:5, (0, 368464))>,
  <Morpheme(前, 5:6, (0, 318425))>,
  <Morpheme(駅, 6:7, (0, 755333))>,
]>, <MorphemeList[
  <Morpheme(で, ...)>,
]>, <MorphemeList[
  <Morpheme(降りる, ...)>,
]>]

[yanagisawa-yu-wap]

@mh-northlander

I reproduced #300 and confirmed the behavior is controlled by add_single.

REPL:

>>> import sudachipy
>>> tok = sudachipy.Dictionary().create()
>>> ms = tok.tokenize("国会議事堂前駅で降りる")

# current default (= add_single=False in my env): unsplittable morphemes become empty
>>> [m.split(mode="a") for m in ms]
[<MorphemeList[
  <Morpheme(国会, 0:2, (0, 364208))>,
  <Morpheme(議事, 2:4, (0, 686966))>,
  <Morpheme(堂, 4:5, (0, 368462))>,
  <Morpheme(前, 5:6, (0, 318423))>,
  <Morpheme(駅, 6:7, (0, 755334))>,
]>, <MorphemeList[
]>, <MorphemeList[
]>]
>>> [m.split(mode="a", add_single=False) for m in ms]
[<MorphemeList[
  <Morpheme(国会, 0:2, (0, 364208))>,
  <Morpheme(議事, 2:4, (0, 686966))>,
  <Morpheme(堂, 4:5, (0, 368462))>,
  <Morpheme(前, 5:6, (0, 318423))>,
  <Morpheme(駅, 6:7, (0, 755334))>,
]>, <MorphemeList[
]>, <MorphemeList[
]>]

expected semantics (“return self when not splittable”) can be achieved by add_single=True

>>> [m.split(mode="a", add_single=True) for m in ms]
[<MorphemeList[
  <Morpheme(国会, 0:2, (0, 364208))>,
  <Morpheme(議事, 2:4, (0, 686966))>,
  <Morpheme(堂, 4:5, (0, 368462))>,
  <Morpheme(前, 5:6, (0, 318423))>,
  <Morpheme(駅, 6:7, (0, 755334))>,
]>, <MorphemeList[
  <Morpheme(で, 7:8, (0, 101425))>,
]>, <MorphemeList[
  <Morpheme(降りる, 8:11, (0, 734940))>,
]>]

[hayashi-mas-wap]

I missed that option, thank you!