python - How to parse tuples that has no delimiters? -

i'm using pyparser , want parse files contain tree structure values stored after equals sign no actual delimiters otherwise.

i've done parsing except rare cases when data stored in multiple lines, have rewrite parser not stop grabbing value equals sign end of line equals sign equals sign (or end) ignoring word preceeds (or ignoring end).

example of data:

(   itemname = foo   someotherstuff = bar   foo1   foo2   astring1 = itemname   someotherstuff  ) 

code:

equals = suppress("=") token = word(alphanums + "-,./_:*+=#[];") decimal = regex(r'-?0|[1-9]\d*').setparseaction(lambda t:int(t[0])) stringtemplate = token | decimal sexplist = group(suppress("(") + zeroormore(sexp) + suppress(")"))  sexp = forward() 

this doesn't work obviously

astring = group(stringtemplate + equals + stringtemplate)

so i've tried these:

multilinestring = group(token + equals + oneormore(stringtemplate) + ~followedby(stringtemplate + equals)) multilinestring = group(token + equals + oneormore(stringtemplate) + notany(stringtemplate + equals)) multilinestring = group(token + equals + oneormore(stringtemplate) + ~(stringtemplate + equals)) 

but didn't work. got either errors or parser gobbling of data 1 record.

you on right track ~followedby, these expressions have part of oneormore repetitive expression. like:

multilinestring = group(token + equals + oneormore(stringtemplate + ~followedby(equals))) match in multilinestring.searchstring(test):     match.pprint() 

some other comments:

• look @ definitions of token, decimal, , stringtemplate. stands now, never match decimal expression. avoid mistaking integer token, suggest change token word(alphas, alphanums + "-,./_:*+=#[];")
• you may want reconsider whether want allow '=' part of token. if omits space between token , delimiting '=', meant delimiter sucked token key.
• consider adding results names access key , value bits of each key-value pair: multilinestring = group(token("key") + equals + oneormore(stringtemplate + ~followedby(equals))("value")). can access them match.key , match.value - of course can use list indexing key match[0], find named access easier work with.