Add Copyright Holder information for Packages

[pombredanne]

This is to support nexB/scancode.io#631
This feature consist in:

• adding a new field to the package model named "holder" (or "primary holder" or "copyright holder"
• add code that will process the package "copyright" when present and extract the holder from it. There is code for this likely in

  scancode-toolkit/src/summarycode/copyright_tallies.py

  Line 225 in 22109c4

  def tally_persons(texts):

  that's used to computed summary holders and/or in cluecode.py to run a copyright detection on the package value and get the holder out of it. It should be canonicalized IMHO