Describe the bug
I did a phrase search via a Japanese tokenizer(sudachi). But phrase search doesn't work as intended.
In my opinion, if include an unindexed term in a phrase query, the term will be ignored in phrase query.

example code is below.
https://github.com/po3rin/python_playground/blob/master/try-pyterrier/sudachi.ipynb

dependencies

[tool.poetry.dependencies]
python = ">=3.8,<3.11"
pandas = "^1.4.0"
SudachiDict-core = ">=20210802"
SudachiPy = ">=0.6.2,<0.7.0"

I referred to the code below when searching Japanese.
https://colab.research.google.com/github/terrier-org/pyterrier/blob/master/examples/notebooks/non_en_retrieval.ipynb

Expected behavior
If it does not contain a term that is not indexed, we expect the search to end without a hit.

Documentation and Issues

• [ x ] I have checked the PyTerrier documentation for relevant content
• [ x ] I have checked for previous relevant PyTerrier issues