Use ED error estimates, Power law fit with xmin

[spencerw]

Added three new features to the code:

• The FFD function can now accept an array of luminosities, for constructing FFDs with multiple stars
• The FFD function can now accept equivalent duration error estimates, so you don't have to estimate the x errors with the flare durations and flux error.
• Added a new function 'FFD_powerlaw', which fits a powerlaw to the FFD data. It can also automatically find xmin by minimizing the KS distance between the model and the trimmed data (following the description in Clauset et al. 2007