> -----Original Message-----
> From: Ralph Corderoy [mailto:ra...@in...]
> Sent: Wednesday, October 11, 2006 7:08 AM
> To: Paul McGuire
> Cc: pyp...@li...
> Subject: Re: [Pyparsing] Operator Precedence and Associativity.
>
>
> Hi Paul,
>
> > Well, why not?! Check this out!
> >
> > def operatorGrammar( baseExpr, opList ):
> > ret = Forward()
> > lastExpr = baseExpr | ( Suppress('(') + ret +
> Suppress(')') ) ...
>
> That's impressive and exactly what I was after. You may want
> to consider formalising it as part of pyparsing since
> recursive descent parsers often use operator precedence
> parsers for that part of things.
>
> Thanks!
>
>
> Ralph.
>
Ralph -
Note that this does not *quite* do the LR-style recursive descent you were
looking for. That is to say,
9 + 2 + 3
parses as:
[[9, '+', 2, '+', 3]]
and not as:
[[[9, '+', 2], '+', 3]]
which is what I think you were looking for originally.
On the whole, though, it seems quite interesting. It took me about 2
minutes to do the following boolean expression grammar:
boolExpr = operatorGrammar( Word(alphas),
[
("not",1,"R"),
("or",2,"L"),
("and",2,"L"),
])
test = ["p and not q",
"not not p",
"not(p and q)",
"q or not p and r",
"q or not (p and r)"
]
for t in test:
print t,'\n',boolExpr.parseString(t),'\n'
Giving:
p and not q
[['p', 'and', ['not', 'q']]]
not not p
[['not', ['not', 'p']]]
not(p and q)
[['not', ['p', 'and', 'q']]]
q or not p and r
[[['q', 'or', ['not', 'p']], 'and', 'r']]
q or not (p and r)
[['q', 'or', ['not', ['p', 'and', 'r']]]]
I need to give this a little more thought as far as including it into "core"
pyparsing (as opposed to say, just making it another example - the examples
directory is sort of becoming its own Pyparsing Cookbook). I want to make
sure I get the API right. For instance, I might want to support a 4'th
entry in each input tuple, representing a parse action to be attached to the
internally generated pyparsing expression.
-- Paul
|
