Better Table Parsing

[pjkaufman]

Fixes #652
Fixes #597

The regex table parsing logic was a little too complex and could cause performance issues. Thus it was replaced with a different parser that starts by identifying the table separator row and validates that it has no text earlier on the line, that there is an actual line before it, and that all cells in the separator row have at least one dash in it. Then it verifies that the separator row is not three dashes that has a line of text since that is either YAML or a header. Once that is done, it grabs any rows that may exist after the table separator row.

Changes Made:

• Updated the table parsing logic
• Added UTs for the table parsing
• Added a UT for the paragraph blank lines issue