Demand pattern library

[kaklise]

Summary

The following PR includes a demand pattern library with the following capabilities:

• Load a library of demand patterns from a JSON file (this contains patterns from existing INP files)
• Add a pattern to the library
• Create a pulse pattern (on/off sequence)
• Create a pattern that follows a gaussian or triangular distribution
• Combine patterns
• Create a copy of a pattern
• Modify the pattern timestep or start clocktime
• Add gaussian random noise to a pattern
• Normalize the pattern to have a mean of 1
• Filter patterns by category
• Convert the pattern to a WNTR Pattern object to it add to a WaterNetworkModel
• Convert the pattern to a Pandas Series (indexed by time in seconds) to easily analyze or further modify the pattern
• Plot patterns
• Save and load custom demand pattern libraries

The WNTR Pattern object now includes a flag (in time options) for pattern interpolation, this is used to extend and combine patterns.

Tests and documentation

Tests are located in test_library_demand.py
Documentation is in libraries.rst (the multispecies model library is now also included in this section of the documentation)

Acknowledgement

By contributing to this software project, I acknowledge that I have reviewed the software quality assurance guidelines and that my contributions are submitted under the Revised BSD License.

[kbonney]

The demand library feature is well-designed, documented, and tested. I have suggested some minor changes to the documentation (kaklise#81), but otherwise this PR is solid and ready to merge.

[coveralls]

coverage: 82.083% (+0.3%) from 81.788%
when pulling 14ab83d on kaklise:demand_library
into 102e2ad on USEPA:main.