How to parse F5 bigip.conf using pyparsing

Question

am trying to figure out how to use this nifty lib to parse BigIP config files... the grammar should,be something like this:

stanza :: name { content }
name   :: several words, might contain alphas nums dot dash underscore or slash
content:: stanza OR ZeroOrMore(printable characters)

To make things slightly more complicated, one exception:

If name starts with "rule ", then content cannot be "stanza"

I started with this:

from pyparsing import *
def parse(config):
    def BNF():
        """
        Example:
        ...
        ltm virtual /Common/vdi.uis.test.com_80_vs {
            destination /Common/1.2.3.4:80
            http-class {
                /Common/http2https
            }
            ip-protocol tcp
            mask 255.255.255.255
            profiles {
                /Common/http { }
                /Common/tcp { }
            }
            vlans-disabled
        }
        ...
        """        
        lcb, rcb, slash, dot, underscore, dash = [c for c in '{}/._-']
        name_word = Word(alphas + nums + dot + underscore + slash + dash)
        name = OneOrMore(name_word).setResultsName("name")
        stanza = Forward()
        content = OneOrMore(stanza | ZeroOrMore(OneOrMore(Word(printables)))).setResultsName("content")
        stanza << Group(name + lcb + content + rcb).setResultsName("stanza")
        return stanza


    return [x for x in BNF().scanString(config)]

The code above seems to lock up in some infinite loop. It is also missing my requirement for excluding looking for 'stanza" if "name" starts with "rule ".

Can you provide some example input to test with? What exception were you getting? — Michael0x2a, Apr 08 '13 at 21:23

Daniel 'Dang' Griffith · Answer 1 · 2013-04-09T16:43:43.263

OneOrMore(ZeroOrMore(OneOrMore(Word(printables))) will always match, thus leading to the infinite loop.

Also, printables includes a closing curly bracket, which gets consumed by the content term, and is no longer available for the stanza. (If your content can including a closing bracket, you need to define something to escape it, to distinguish a content bracket from a stanza bracket.)

To address the name rule, you need another content definition, one that doesn't include stanza, and a "rule rule".

def parse(config):
    def BNF():
        lcb, rcb, slash, dot, underscore, dash = [c for c in '{}/._-']
        printables_no_rcb = Word(printables, excludeChars=rcb)
        name_word = Word(alphas + nums + dot + underscore + slash + dash)
        name = OneOrMore(name_word).setResultsName("name")
        rule = Group(Literal('rule') + name).setResultsName("name")
        rule_content = OneOrMore(printables_no_rcb).setResultsName("content")
        stanza = Forward()
        content = OneOrMore(stanza | OneOrMore(printables_no_rcb)).setResultsName("content")
        stanza << Group(rule + lcb + rule_content + rcb | name + lcb + content + rcb).setResultsName("stanza")
        return stanza
    return [x for x in BNF().scanString(config)]

How to parse F5 bigip.conf using pyparsing

1 Answers1