Python 2 not yet supported

[jayvdb]

Traceback:
test/test_on_file.py:5: in <module>
    from charset_normalizer import CharsetNormalizerMatches as CnM
charset_normalizer/__init__.py:2: in <module>
    from charset_normalizer.normalizer import CharsetNormalizerMatches, CharsetNormalizerMatch
charset_normalizer/normalizer.py:3: in <module>
    import statistics
E   ImportError: No module named statistics

[jayvdb]

Also seen on https://travis-ci.org/Ousret/charset_normalizer/jobs/577468690

There is a drop-in statistics backport which can be required only on py2. https://pypi.org/project/statistics

After that, the next py2 problem is ImportError: cannot import name lru_cache , which also has a backport, but it is in a different namespace so needs a try block https://pypi.org/project/backports.functools_lru_cache/

[jayvdb]

And then

AttributeError: 'str' object has no attribute 'isprintable'

and

File "../charset_normalizer/normalizer.py", line 239, in from_bytes
    str(sequences, encoding=p)
TypeError: str() takes at most 1 argument (2 given)

And test_should_be_accented failing for some reason.

[jayvdb]

'isprintable' is the hard one ... it will require a backport of py3 unicode to py2. There are a few of them - I used one in about 2015, but I suspect they may not be maintained very well as the usefulness and usage reduced.

[jayvdb]

https://pypi.org/project/unicodedata2/ is the one I was thinking about ; not sure it has isprintable . maybe need future for that?

[Ousret]

Hi,

I've tried the same and stopped at isprintable method support.
I'm looking/considering bulding my own isprintable method and add it to unicode helper class.

[jayvdb]

https://github.com/PythonCharmers/python-future/blob/b7fd8a50117a1eb9126ca8840e2b4c4a5c5f70bf/src/future/types/newstr.py#L414 raise NotImplementedError('fixme') :/

PythonCharmers/python-future#494

[Ousret]

See #5