
hachoir-regex is a Python library for regular expression (regex or regexp) manupulation. You can use a|b (or) and a+b (and) operators. Expressions are optimized during the construction: merge ranges, simplify repetitions, etc. It also contains a class for pattern matching allowing to search multiple strings and regex at the same time.
Website: http://hachoir.org/wiki/hachoir-regex
Regex examples ==============
Different methods to create regex:
from hachoir_regex import parse, createRange, createString createString("bike") + createString("motor")
<RegexString 'bikemotor'>
parse('(foo|fooo|foot|football)')
<RegexAnd 'foo(|[ot]|tball)'>
regex = createString("1") | createString("3"); regex
<RegexRange '[13]'>
regex |= createRange("2", "4"); regex
<RegexRange '[1-4]'>
As you can see, you can use classic "a|b" (or) and "a+b" (and) Python operators, and expressions are optimized for fast pattern matching.
Regex using repetition:
parse("(a{2,}){3,4}")
<RegexRepeat 'a{6,}'>
parse("(a*|b)*")
<RegexRepeat '[ab]*'>
parse("(a*|b|){4,5}")
<RegexRepeat '(a+|b){0,5}'>
Compute minimum and maximum length of matched pattern:
r=parse('(cat|horse)') r.minLength(), r.maxLength()
(3, 5)
r=parse('(a{2,}|b+)') r.minLength(), r.maxLength()
(1, None)
Pattern maching ===============
Use PatternMaching if you would like to match multiple strings and regex at the same time:
>>> from hachoir_regex import PatternMatching >>> p = PatternMatching() >>> p.addString("a") >>> p.addString("b") >>> p.addRegex("[cd]") >>> for start, end, item in p.search("a b c d"): ... print "%s..%s: %s" % (start, end, item) ... 0..1: a 2..3: b 4..5: [cd] 6..7: [cd]
You can also attach user data to a pattern:
>>> p = PatternMatching() >>> p.addString("un", 1) >>> p.addString("deux", 2) >>> for start, end, item in p.search("un deux"): ... print "%r at %s: userdata=%r" % (item, start, item.user) ... <StringPattern 'un'> at 0: userdata=1 <StringPattern 'deux'> at 3: userdata=2
Download hachoir-regex on Python cheeseshop, it's distributed under GNU GPL license.
Victor Stinner aka haypo http://hachoir.org/
participants (1)
-
haypo