On Aug 31, 9:29*am, rajmoha...@gmai l.com wrote:
Hi all,
* * Suppose I have a string which contains quotes inside quotes -
single and double quotes interchangeably -
*s = "a1' b1 " c1' d1 ' c2" b2 'a2"
* * *I need to start at b1 and end at b2 - i.e. I have to parse the
single quote strings from inside s.
Pyparsing defines a helper method called nestedExpr - typically it is
used to find nesting of ()'s, or []'s, etc., but I was interested to
see if I could use nestedExpr to match nested ()'s, []'s, AND {}'s all
in the same string (like we used to do in our algebra class to show
nesting of higher levels than parens - something like "{[a + 3*(b-c)]
+ 7}" - that is, ()'s nest within []'s, and []'s nest within {}'s).
This IS possible, but it uses some advanced pyparsing methods. I
adapted this example to map to your case - this was much simpler, as
""s nest within ''s, and ''s nest within ""s. I still keep a stack of
previous nesting, but I'm not sure this was absolutely necessary.
Here is the working code with your example:
from pyparsing import Forward, oneOf, NoMatch, Literal, CharsNotIn,
nestedExpr
# define special subclass of Forward, that saves previous contained
# expressions in a stack
class ForwardStack(Fo rward):
def __init__(self):
super(ForwardSt ack,self).__ini t__()
self.exprStack = []
self << NoMatch()
def __lshift__(self ,expr):
self.exprStack. append(self.exp r)
super(ForwardSt ack,self).__lsh ift__(expr)
return self
def pop(self):
self.expr = self.exprStack. pop()
# define the grammar
opening = ForwardStack()
closing = ForwardStack()
opening << oneOf(["'", '"'])
closing << NoMatch()
matchedNesting = nestedExpr(open ing, closing, CharsNotIn('\'" '),
ignoreExpr=None )
# define parse-time callbacks
alternate = {'"':"'", "'":'"'}
def pushAlternate(t ):
# closing expression should match the current opening quote char
closing << Literal( t[0] )
# if we find the other opening quote char, it is the beginning of
# a nested quote
opening << Literal( alternate[ t[0] ] )
def popClosing():
closing.pop()
opening.pop()
# when these expressions match, the parse action will be called
opening.setPars eAction(pushAlt ernate)
closing.setPars eAction(popClos ing)
# parse the test string
s = """ "a1' b1 " c1' d1 ' c2" b2 'a2" """
print matchedNesting. parseString(s)[0]
Prints:
['a1', [' b1 ', [' c1', [' d1 '], ' c2'], ' b2 '], 'a2']
-- Paul
