CLN: created foundation for complex type annotations (GH14468)

This is mostly just a stub  file for now until a more clear picture develops

What has been noticed so far:

* numpy has no clear definition of what array_like means. all of these are valid:
   - Python scalars
   - tuples
   - lists
   - buffers
   - scalars in both python and numpy
   - more?
* similar story but not so extreme with dtypes
* python and numpy scalar helpers have been defined

Author: brmc

pandas/types/hinting.py

@@ -1,2 +1,9 @@
 #!/usr/bin/env python
 # -*- coding: utf-8 -*-
+import numpy as np
+
+from typing import TypeVar, AnyStr
+
+ArrayLike = TypeVar('ArrayLike', buffer, list, dict, np.array)
+Scalar = TypeVar('Scalar', int, float)
+PythonScalar = TypeVar('PythonScalar', int, float, AnyStr)