Re: [Pyparsing] How can I print syntax tree for debug?

[Paul M. <pa...@al...>]

Have you looked at the dump() method on ParseResults?  dump() lists out a
nested list of the ParseResults (the same as you get from the asList()
method), then lists a hierarchical tree of the named results.

Here is how your program's parsed tokens look using dump():

import pyparsing as pp
Name = pp.Word(pp.alphas).setResultsName('Name')
S = pp.delimitedList(Name, delim=pp.White())
results = S.parseString("a b c")
print results.dump()

Prints:
['a', 'b', 'c']
- Name: c

You're probably wondering why 'Name' only gives 'c' as the value.  In the
default use of setResultsName, results names work like a Python dict.  While
parsing progresses, 'Name' gets successively set to 'a', then 'b', and then
finally 'c'.

If you want to see all the 'Name' results, then use a different form of
setResultsName, passing listAllResults as True:

Name = pp.Word(pp.alphas).setResultsName('Name',listAllMatches=True)
S = pp.delimitedList(Name, delim=pp.White())
results = S.parseString("a b c")
print results.dump()

Prints:
['a', 'b', 'c']
- Name: ['a', 'b', 'c']


A couple of other comments:
- Pyparsing skips over whitespace by default, so there is no need to define
a sequence of Name words as a delimitedList(Name,delim=White()).  You can
just OneOrMore(Name) to get this same result.
- setResultsName can be used in the manner you have done, and you should use
listAllMatches=True so that you don't just get the latest Name.  When I
first added setResultsName, my intention was that a single expression might
be used several times within a larger grammar, with different results names
depending on where in the grammar the expression occurred.  For instance,
here is an expression for a two-digit number:

num2d = pp.Word(pp.nums,exact=2)

I could label these with:

num2d.setResultsName("twoDigitNumber")

But if I were parsing a timestamp, this would be much more meaningful:

timestamp = num2d.setResultsName("year") + "/" + \
		num2d.setResultsName("month") + "/" + \
		num2d.setResultsName("day") + \
		num2d.setResultsName("hours") + ":" + \
		num2d.setResultsName("minutes") + ":" + \
		num2d.setResultsName("seconds")

Parsing the string "07/06/23 12:34:56" and calling dump() gives this output:

['07', '/', '06', '/', '23', '12', ':', '34', ':', '56']
- day: 23
- hours: 12
- minutes: 34
- month: 06
- seconds: 56
- year: 07

I am glad you're using results names, I think they are an important feature
for all but the most trivial parsers.  In the next release of pyparsing,
this syntax will be simplfied to:

timestamp = num2d("year") + "/" + \
		num2d("month") + "/" + \
		num2d("day") + \
		num2d("hours") + ":" + \
		num2d("minutes") + ":" + \
		num2d("seconds")

Probably more than you were bargaining for, but I hope this helps.

-- Paul



> -----Original Message-----
> From: pyp...@li... 
> [mailto:pyp...@li...] On 
> Behalf Of sam lee
> Sent: Saturday, June 23, 2007 11:24 PM
> To: pyp...@li...
> Subject: [Pyparsing] How can I print syntax tree for debug?
> 
> Hi.
> 
> I want to print a syntax tree or something similar for 
> debugging purpose.
> 
> import pyparsing as pp
> Name = pp.Word(pp.alphas).setResultsName('Name')
> S = pp.delimitedList(Name, delim=pp.White())
> print S.parseString("a b c")
> ['a', 'b', 'c']
> 
> I want to print something like:
> ['Name:a', 'Name:b', 'Name:c'] #prefixing each token with 
> results name of Name.
> 
> What I am currently doing is:
> Name.setParseAction(lambda s,loc,toks: ['Name:'+x for x in toks])
> 
> But, I have to write 'Name:' part manually. Probably what I want is:
>     def display( self, s, loc, toks ):
>         return [self.getName() + x for x in toks]
>     pp.ParseResults.debugDisplay = display
>     Name.setParseAction(pp.ParseResults.debugDisplay)
> 
> Above doesn't work.
> 
> I think I am going in the wrong direction. Is there a nice way of
> printing debugging information? Printing side-way syntax tree would be
> superb!!
> 
> Something like:
>     Expr:
>          Name: 'apple'
>          Op: '+'
>          Literal:
>                Number: 233
> 
> 
> Thank you.
> 
> --------------------------------------------------------------
> -----------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> Pyparsing-users mailing list
> Pyp...@li...
> https://lists.sourceforge.net/lists/listinfo/pyparsing-users
> 
>