IO error related to the dataset compressed with .Z

[yzong12138]

nocr/trec-robust-2004 dataset, cannot ready the document from the file successfully A clear and concise description of what the bug is.

Affected dataset(s) nocr/trec-robust-2004

To Reproduce Steps to reproduce the behavior:

1. First get the access of the Disk 4 and 5 from the individual agreement
2. Copy the FBIS, FR94, etc which contains the data compressed at the format .Z, .0Z to the path ~/.ir_datasets/disks45/corpus/NEWS_data
3. Run the code by

   import ir_datasets
   robust_dataset = ir_datasets.load("disks45/nocr/trec-robust-2004")
   print([query for query, _ in zip(robust_dataset.docs_iter(), range(2))])

Expected behavior

>>> [query for query, _ in zip(robust_dataset.docs_iter(), range(2))]
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<stdin>", line 1, in <listcomp>
  File "/home/zong/.conda/envs/xp/lib/python3.10/site-packages/ir_datasets/util/__init__.py", line 147, in __next__
    return next(self.it)
  File "/home/zong/.conda/envs/xp/lib/python3.10/site-packages/ir_datasets/formats/trec.py", line 127, in docs_iter
    yield from self._docs_iter(path)
  File "/home/zong/.conda/envs/xp/lib/python3.10/site-packages/ir_datasets/formats/trec.py", line 162, in _docs_iter
    with io.BytesIO(unlzw3.unlzw(path)) as f:
  File "/home/zong/.conda/envs/xp/lib/python3.10/site-packages/unlzw3/__init__.py", line 70, in unlzw
    ba_in = bytearray(data)
TypeError: string argument without an encoding
>>> 

[yzong12138]

The bug should comes from the line 162: It should change from

with io.BytesIO(unlzw3.unlzw(path)) as f:

to

with io.BytesIO(unlzw3.unlzw(Path(path))) as f:

Before it will pass a str to the unlzw to decompress but now the input for the unlzw() is a Path so it will try to load the data from the path first.

[seanmacavaney]

closed with #247