PySheaf draft complete!

This is the first major revision of PySheaf! While some of the functionality developed on the v0.x releases is not present, it is much more flexible and usable overall. If you need that functionality, please let me know as we'll be bringing capabilities forward.

Important changes:

1. The CellComplex class has been removed, as base spaces are assumed to be Alexandrov spaces built on partially ordered sets. The partial order's Hasse diagram is stored as a NetworkX DiGraph instance
2. The main functionalities supported are:
   a. Computing consistency radius
   b. Optimizing consistency radius by interpolating/extrapolating values of an assignment
   c. Identifying star open sets that are consistent to a specified level
3. Assignments are housed internal to the Sheaf structure
4. Optimizers for minimizing consistency radius and performing data fusion can now be supplied by the user
5. Cohomology is not presently supported.
   a. While the cochain complex exists mathematically for sheaves of abelian groups on partial orders, it is not necessarily efficient to compute!
   b. We anticipate adding this capability in the future

PySheaf for Python 3.6 initial revision

In addition to working with Python 3.6, a number of improvements to the computation of consistency filtrations have been made.

PySheaf for Python 2.7 incremental

Added Sheaf.consistentStarCollection

PySheaf for Python 2.7 (legacy)

This version is the legacy version of PySheaf for Python 2.7. Bugfixes will generally be pulled back from newer versions, but only at my convenience.

Final draft

This release is the final draft version of PySheaf for Python 2.7. Work on this is planned to be frozen, due to a number of substantial changes to notational conventions and ultimately a transition to Python 3.6.

Going forward, many of the Sheaf and CellComplex subclasses will be removed until they can be carefully tested and reintroduced as needed.

Initial Release

v0.1

Somehow lost this...