Regex for sentence boundary and for punctuation escaping.

structures.py

DEFAULT_SENTENCE_BOUNDARIES = ['(?<=[0-9]|[^0-9.])(\.)(?=[^0-9.]|[^0-9.]|[\s]|$)','\.{2,}','\!+','\:+','\?+']
"""
Breaking it down:
(?<=[0-9]|[^0-9.])(\.)(?=[^0-9.]|[^0-9.]|[\s]|$) -> looks for ant period that is not preceded or succeded by a digit or other period. 
This avoids the algorithm to split sentences at decimal numbers or reticences.
\.{2,} -> captures reticences.
\!+ -> captures series of exclamation points.
\:+ -> captures series of colons.
\?+ -> captures series of question marks.
"""

DEFAULT_PUNCTUATIONS = ['(?<=[0-9]|[^0-9.])(\.)(?=[^0-9.]|[^0-9.]|[\s]|$)','\.{2,}',
                        '\!+','\:+','\?+','\,+', r'\(|\)|\[|\]|\{|\}|\<|\>']

"""
The only difference here are in:
\,+ -> captures series of commas
\(|\)|\[|\]|\{|\}|\<|\> -> captures any parenthesis (the pipe | sign means 'or')
"""

# Another used regex is \s|\t|\n|\r -> it matches at any white char (\s), tab char (\t), new line char (\n) and carriage return char (\r)