RegEx for matching brackets

[NevilleDNZ]

Below is a (flawed) one line RegEx that checks curly brackets (from
awk/c/python input) are being matched. Is there a one liner for doing
this in python?

ThanX
N

re_open_close="(((\{))[^{}]*((?(0)\})))+"
re_open_close=re.compile(re_open_close)
tests="""
{ this is a test BAD
{ this is a test } OK
{ this is a test } { this is a test } OK
{ this is a test } { this { this is a test } is a test } OK
{ this is a test { this { this is a test } is a test } missing
close BAD
""".splitlines()[1:]
for test in tests:
if bool(re_open_close.search(test)) == bool(re.search("OK",test)):
print "DETECTED:",test
else:
print "MISSED:",test
[linux]$ python ./re_matching.py
DETECTED: { this is a test BAD
DETECTED: { this is a test } OK
DETECTED: { this is a test } { this is a test } OK
DETECTED: { this is a test } { this { this is a test } is a test }
OK
MISSED: { this is a test { this { this is a test } is a test }
missing close BAD

[John Machin]

On May 2, 9:44 am, NevilleDNZ <[e-mail]:

Below is a (flawed) one line RegEx that checks curly brackets (from
awk/c/python input) are being matched. Is there a one liner for doing
this in python?

ThanX
N

re_open_close="(((\{))[^{}]*((?(0)\})))+"
re_open_close=re.compile(re_open_close)
tests="""
{ this is a test BAD
{ this is a test } OK
{ this is a test } { this is a test } OK
{ this is a test } { this { this is a test } is a test } OK
{ this is a test { this { this is a test } is a test } missing
close BAD
""".splitlines()[1:]
for test in tests:
if bool(re_open_close.search(test)) == bool(re.search("OK",test)):
print "DETECTED:",test
else:
print "MISSED:",test
[linux]$ python ./re_matching.py
DETECTED: { this is a test BAD
DETECTED: { this is a test } OK
DETECTED: { this is a test } { this is a test } OK
DETECTED: { this is a test } { this { this is a test } is a test }
OK
MISSED: { this is a test { this { this is a test } is a test }
missing close BAD

Regexes can't count.

To check for balanced braces, you need to write code e.g.

# untested
def balanced_braces(s):
n = 0
for c in s:
if c == '{':
n += 1
elif c == '}':
n -= 1
if n < 0:
return False
return n == 0

HTH,
John

[George Sakkis]

On May 1, 7:44 pm, NevilleDNZ <[e-mail]:

Below is a (flawed) one line RegEx that checks curly brackets (from
awk/c/python input) are being matched. Is there a one liner for doing
this in python?

There is not even a 1000-liner regular expression for this; it's a
context-free language [1], not a regular one [2]. Either do it
manually or use a parser generator [3].

George

[1] http://en.wikipedia.org/wiki/Context-free_language
[2] http://en.wikipedia.org/wiki/Regular_language
[3] http://wiki.python.org/moin/LanguageParsing

[NevilleDNZ]

On May 2, 11:13 am, George Sakkis <[e-mail]:

[1]http://en.wikipedia.org/wiki/Context-free_language
[2]http://en.wikipedia.org/wiki/Regular_language
[3]http://wiki.python.org/moin/LanguageParsing

Thanx for the link to these parsers. ANTLR looks interesting.
Yoyo: http://www-users.cs.york.ac.uk/~fish...yoyovwg/readme

I figured out a way to do it in python.

#!/usr/bin/env python2
# -*- coding: utf-8 -*-
import re
tests="""
{ a test BAD
{ a test } OK
{ a test } { a test } OK
{ a test } { this { a test } is a test } OK
{ a test { this { a test } is a test } missing close BAD
a test } { this { a test } is a test } BAD
{ a test } this { a test } is a test } BAD
""".splitlines()[1:]

def referee(str):
return bool(re.search("OK",test))

def check_open_close(str):
try:
eval("".join({"{":"[","}":"],"}[c] for c in re.findall( "([{}])|(?:
[^{}]+)", str) if c))
except SyntaxError:
return False
else:
return True

for test in tests:
if check_open_close(test) == referee(test):
print "DETECTED:",test
else:
print "MISSED:",test
[linux]$ ./matchall_open_close.py
DETECTED: { a test BAD
DETECTED: { a test } OK
DETECTED: { a test } { a test } OK
DETECTED: { a test } { this { a test } is a test } OK
DETECTED: { a test { this { a test } is a test } missing close BAD
DETECTED: a test } { this { a test } is a test } BAD
DETECTED: { a test } this { a test } is a test } BAD

It's not exactly what I planned, but as a python one/two-liner it
works fine.
eval("".join({"{":"[","}":"],"}[c] for c in re.findall( "([{}])|(?:
[^{}]+)", str) if c))

ThanX again
N

[NevilleDNZ]

To check a complete python expression use:

def check_open_close(expr):
try:
eval(expr)
except SyntaxError:
return False
else:
return True

This also ignores brackets in quotes, and checks <= & >= operators are
syntatically correct etc...
But is may have side effects... ;-)
eg.
check_open_close('__import__("os").system("echo rm -rf /") # OK')

c.f http://en.wikipedia.org/wiki/Scope_creep

N