nhlib-based classical hazard calculator core, initialization

With the imminent completion of bug # 994033, we have the ability to load a classical hazard job profile from an INI file into the DB. Existing engine code was adapted (see engine.py/engine2.py) to prepare all of the input necessary for running a classical hazard job. Now, we need to start building the pre-execution steps for the classical calculator.

Here is a rough outline of the steps required to run a calculation, in order of execution:

------
Engine
------
(The following 3 parts already exist, for the most part. Additional 'glue' code may be required.)
- parse job (from config file)
- validate job
- store job profile in database (this includes create initial records for input files: logic trees, site model, etc.; additional processing will be required--see below)

----------------------
initialize/pre_execute
----------------------
- parse & store site model (the input model should already be partially stored in the database; see `uiapi.input` and `uiapi.model_content` db tables)
- parse source model logic tree, store sources (from xml to db), create `src2ltsrc` links
- initialize calc progress counters
- record job stats

-------
execute
-------
- if geometry of interest is a region, discretize it into sites (based on th `region_grid_spacing`)
  - we can probably use nhlib geometry code for this; see `nhlib.geo.Polygon.discretize` (https://github.com/gem/nhlib/blob/master/nhlib/geo/polygon.py#L178)
- distribute tasks and wait

------------
post_execute
------------
- clear progress counters

---------------
post_processing
---------------
- mean curves (if requested)
- quantile curves (if requested)
- hazard maps (if requested)