Fix timestamp issue for python2 in s3

[akyrola]

Thanks to @romain-intel 's integration tests, he noticed that the #778 had a bug for python2: datetime had .timestamp() function only since python 3.3.

This PR replaces the .timestamp() with a get_timestamp() function that calculates the timestamp using datetime operations.

Test with both py2 and py3:

Python 2.7.15 | packaged by conda-forge | (default, Mar  5 2020, 14:58:04) 
[GCC Clang 9.0.1 ] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> from metaflow import S3
>>> s3 = S3()
>>> info = s3.info_many(['s3://aapo-mf-test/mnist/t10k-labels-idx1-ubyte.gz', 's3://aapo-mf-test/mnist/train-labels-idx1-ubyte.gz'])
>>> info[0].last_modified, info[1].last_modified
(1632312604.0, 1632312562.0)
>>> 


(py2) MacBook-Pro-5:metaflow akyrola$ conda activate metaflow_dev
(metaflow_dev) MacBook-Pro-5:metaflow akyrola$ aws-vault exec mfdeveloper --  python
Python 3.9.7 | packaged by conda-forge | (default, Sep 14 2021, 01:18:03) 
[Clang 11.1.0 ] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> from metaflow import S3
>>> s3 = S3()
>>> info = s3.info_many(['s3://aapo-mf-test/mnist/t10k-labels-idx1-ubyte.gz', 's3://aapo-mf-test/mnist/train-labels-idx1-ubyte.gz'])
>>> info[0].last_modified, info[1].last_modified