re.compile versus r''

[Terrence Brannon]

Hello, I'm using a tool (PLY) which apparently expects the tokens to
be created using r''

But because one token is a rather complex regular expression, I want
to create the regular expression programmatically.

How can I generate a string and then create something of the same type
that the r'' function does?

Concretely, in the program below, consonant is not the same type as
t_NAME, but I assume that it needs to be for PLY to use it for
tokenizing:

import re

t_NAME = r'[a-zA-Z_][a-zA-Z0-9_]*'

guttural = 'kh?|gh?|\"n'
palatal = '(?:chh?|jh?|\~n)'
cerebral = '\.(?:th?|dh?|n)'
dental = '(?:th?|dh?|n)'
semivowel = '[yrlv]'
sibilant = '[\"\.]?s'
aspirant = 'h'

consonant = re.compile('|'.join([guttural , palatal , cerebral ,
dental , semivowel , sibilant , aspirant]))

print consonant
print t_NAME

 

[Terrence Brannon]

Oh my god, how embarrassing. the r'' notation is to create raw string
<http://www.swc.scipy.org/lec/glossary.html#gdef-raw_string>

I thought it was some form of blessing a string into a regular
expression class.

 

[Fredrik Lundh]

Hello, I'm using a tool (PLY) which apparently expects the tokens to
be created using r''

But because one token is a rather complex regular expression, I want
to create the regular expression programmatically.

How can I generate a string and then create something of the same type
that the r'' function does?

r'' is an alternative syntax for string literals that affects how escape
sequences are interpreted; there's no separate string type for strings
created by this syntax.

</F>

 

[Chris Rebert]

Hello, I'm using a tool (PLY) which apparently expects the tokens to
be created using r''

But because one token is a rather complex regular expression, I want
to create the regular expression programmatically.

How can I generate a string and then create something of the same type
that the r'' function does?

The "r" prefix isn't a function or a type, it's merely a special
literal syntax for strings that's handy when you're writing regexes
and therefore have to deal with another level of backslash escaping.
See the second to last paragraph of
http://docs.python.org/ref/strings.html for more info.

Regards,
Chris

Concretely, in the program below, consonant is not the same type as
t_NAME, but I assume that it needs to be for PLY to use it for
tokenizing:

import re

t_NAME = r'[a-zA-Z_][a-zA-Z0-9_]*'

guttural = 'kh?|gh?|\"n'
palatal = '(?:chh?|jh?|\~n)'
cerebral = '\.(?:th?|dh?|n)'
dental = '(?:th?|dh?|n)'
semivowel = '[yrlv]'
sibilant = '[\"\.]?s'
aspirant = 'h'

consonant = re.compile('|'.join([guttural , palatal , cerebral ,
dental , semivowel , sibilant , aspirant]))

print consonant
print t_NAME