this classifier of learning module is not optimized to run on this huge MNIST data

having this flag drastically reduces time to load & sanity check large image datasets for practical ML tasks

* Replaced mean with standard lib function introduced in 3.4

* #TODO Replaced every with all

* Proper names for variables introduced in 2to3 conversion.

* modifies Fig. to Figure all over the repository

* fixed a small type in logic.py

* removes obsolete Fig dictionary from utils.py

* modified rl notebook according to the new Figure conventions

* removes 'import Fig' and changed the names of Figures

* fixes a small error (by me) in executing the cells in rl notebook

*  Used isclose from utils to maintain compatibilty for 3.4

* Removed redundant distancesquared function

* Removed duplicate definition of clip in grid.py and used import from instead.

* Fixed spelling typo

* Made Log Usage Consistent in text.py

* Added Explicit Imports from utils

Remove count_if, find_if, every, some, Struct; replace them with modern
Python equivalents.

- Fix relative imports in base of is a package or not
- Fix not working notebooks
- Add notebooks to packages sources

Used tool 2to3 to finish the porting to Python 3 plus some fixes by
hand.

Move CountingProbDist to learning.py and use it in NaiveBayesLearner. Make NaiveBayesLearner use target-value frequencies too.

Don't compare strings with 'is not'. (Broke things in pypy.) Also show Struct members in sorted order so doctests work in pypy.