Compression not handled for GCS files

[gelioz]

Problem description

When reading gzip-encoded blob from GCS, file object returns compressed binary data instead of decompressed text.

Steps/code to reproduce the problem

In [30]: smart_open.open('gs://test/file.json.gz').read()
---------------------------------------------------------------------------
UnicodeDecodeError                        Traceback (most recent call last)
<ipython-input-30-4e7f4d55e81d> in <module>
----> 1 smart_open.open('gs://test/file.json.gz').read()

/usr/local/Cellar/python/3.7.7/Frameworks/Python.framework/Versions/3.7/lib/python3.7/codecs.py in read(self, size, chars, firstline)
    502                 break
    503             try:
--> 504                 newchars, decodedbytes = self.decode(data, self.errors)
    505             except UnicodeDecodeError as exc:
    506                 if firstline:

UnicodeDecodeError: 'utf-8' codec can't decode byte 0x8b in position 1: invalid start byte

Also, I was able to find root cause - gcs.Reader class not implements .name attribute which cause compression.compression_wrapper function to skip decompressing (because os.path.splitext(file_obj.name) returns ('unknown', '')).

Versions

Darwin-19.3.0-x86_64-i386-64bit
Python 3.7.7 (default, Mar 10 2020, 15:43:33)
[Clang 11.0.0 (clang-1100.0.33.17)]
smart_open 2.0.0

[mpenkov]

Hi @gelioz, thank you for reporting this.

We have already fixed the issue in the develop branch (see #506). Are you able to confirm?

[gelioz]

Thank you, @mpenkov.

Yes, I can confirm it works for me on develop branch. Is there any estimate on when 2.0.1 will be released?

[bradmurray]

Does this also fix the same issue for writing? I can currently read from a .gz file on GCS with no problem, but when I write it is writing uncompressed even thought the filename ends in .gz.

[gelioz]

Yes, it does - no issues with file compression on writing.

[mpenkov]

https://pypi.org/project/smart-open/2.1.0/