Acorn

- Added the new `ipython.js` file to the package data for acorn so that it gets
distributed correctly.

This revision addresses many of the missing features supporting the
ipython/jupyter notebooks that are breaking features. Any notebook-level loops
are now intercepted and wrapped within a cell function so that they don't
generate millions of database entries; thumbnails get saved to the database (as
long as the plots use PNG and JPEG as display types, we'll support vega and svg
later). Also, diffing between cell executions is now implemented as well.

- Fixed issues 7, 20 and 25

*NOTE*: unit testing is only up to 93%. We are still missing good unit tests for
`matplotlib` and `sklearn` regressors auto-fit and auto-predict interception is
not tested yet. The ipython notebook support module is now quite large and
doesn't have any unit tests yet. Adding these unit tests will require using js
to emulate the notebooks, etc. which I don't have time for now.

- Fixed issue 14.
- Added a function to list all projects and tasks in the database directory.

- Fixed `pandas.get` and `pandas.__getitem__` issues.
- Fixed the broken decoration problem; it was caused by the `_cstack_call` depth being too high when an exception was handled by `acorn` or unhandled by `acorn`, but handled by a higher level calling method.
- Fixed the `pytest` internal errors; they were being caused by trying to print the arguments in the `xwrapper` exception handler; now we just print the types of the arguments.

- Added two extra ignores to `sklearn` to stop segfaulting problem.

- Fixed streamlining for the `pandas` data frame plots by updating the config file. Also added streamlining support for constructors.