datadog 0.37.0 wheel import error for Python < 3.5

**Describe the bug**

After installing the latest datadog 0.37.0 wheel release using a Python version < 3.5, a `typing` package import error is raised when importing `datadog`.



**To Reproduce**

Steps to reproduce the behavior:



1. Install a version of Python < 3.5 (for example, Python 3.4 or Python 2.7).

2. Install the wheel release of datadog 0.37.0 by running `pip install datadog`.

3. Try importing `datadog` in an interactive interpreter to see the import error:



```

>>> import datadog

Traceback (most recent call last):

  File "<stdin>", line 1, in <module>

  File "datadog/__init__.py", line 17, in <module>

    from typing import Any, List, Optional

ImportError: No module named typing

```



**Expected behavior**

The `typing` package should be included in the wheel's install requirements if Python < 3.5 is being used.



In `setup.py`, the `typing` package is added to `install_requires` if running a Python version < 3.5 (otherwise, the package is already included in the standard library for newer Python versions). This optional `typing` dependency is correctly installed when running `python setup.py install` or pip-installing from source. The wheel release that is installed via `pip install datadog` doesn't include the `typing` dependency (probably because the wheel release isn't Python version specific).



A couple of solutions come to mind:



- (Probably easiest) Always include the `typing` package in `install_requires`, regardless of Python version. According to the [`typing` package docs](https://pypi.org/project/typing/), if `typing` is already part of the standard library, installing the package has no effect.

- Create separate wheel releases for Python < 3.5 and Python >= 3.5.



**Environment and Versions:**

Version: datadog 0.37.0

Tested with Python 2.7.15 on macOS 10.15.5.



**Additional context**

The `typing` package import was added in PR #565. Thanks @justinsousa for pointing out this issue!

[]