Dependency to six library

Imagehash 4.3.0 tries to import six, but it is not declared as a dependency for Python 3.10:

https://github.com/osmlab/editor-layer-index/actions/runs/3120554190/jobs/5061223939 https://github.com/osmlab/editor-layer-index/blob/gh-pages/requirements.txt

Traceback (most recent call last):
  File "/home/runner/work/editor-layer-index/editor-layer-index/scripts/sync_wms.py", line 28, in <module>
    import imagehash
  File "/opt/hostedtoolcache/Python/3.10.6/x64/lib/python3.10/site-packages/imagehash/__init__.py", line 155, in <module>
    from six.moves.collections_abc import Callable
ModuleNotFoundError: No module named 'six'
Error: Process completed with exit code 1.

There seems to be a fix already in master, but I’m not sure if this works in all cases. I would suggest something like this (not tested):

# not sure when typing.Callable was first introduced. Maybe 3.5?, But I'm not sure if such old python version are still supported by ImageHash
if sys.version_info >= (3, 6):
    # https://stackoverflow.com/questions/65858528/is-collections-abc-callable-bugged-in-python-3-9-1
    if sys.version_info > (3, 9, 1):
        from collections.abc import Callable
    else:
        from typing import Callable
    try:
        MeanFunc = Callable[[NDArray], float]
        HashFunc = Callable[[Image.Image], ImageHash]
    except TypeError:
        MeanFunc = Callable  # type: ignore
        HashFunc = Callable  # type: ignore

[JohannesBuchner]

I made a release, 4.3.1. Can you please test whether it works for you?

[JohannesBuchner]

I made a release, 4.3.1. Can you please test whether it works for you?

@JohannesBuchner Thanks for creating a new release that fast! The github action that was affected will trigger automatically on Sunday. I will report back if it solved the issue or not. But I’m very positive that it will.

OK, please reopen if it is still an issue.

Github Actions test these versions: 2.7, 3.5, 3.6, 3.9.1, 3 (latest, 3.10.4 at the moment)

I appreciate that you are willing to support that many old Python versions. In case you are not aware, Python 2.7 reached its end of life on January 1, 2020. A lot of Python packages follow roughly the numpy deprecation policy of supporting Python versions for 42 months: https://numpy.org/neps/nep-0029-deprecation_policy.html

I am aware. I might drop it once things really start to break, but it’s not that bad atm.