Translating regex match groups

Question

I need to match name objects according to pdf file specification. However, names may contain hexadecimal digits (preceded by #) to specify special characters. I would like to translate these matches to corresponding characters. Is there a clever way to do it without re-parsing the match string?

import re

Name = re.compile(r'''
    (/                                        # Literal "/"
        (?:                                   #
            (?:\#[A-Fa-f0-9]{2})              # Hex numbers
            |                                 # 
            [^\x00-\x20 \x23 \x2f \x7e-\xff]  # Other
        )+                                    #
    )                                         #
    ''', re.VERBOSE)

#  some examples

names = """
    The following are examples of valid literal names:

    Raw string                       Translation

    1.  /Adobe#20Green            -> "Adobe Green"
    2.  /PANTONE#205757#20CV      -> "PANTONE 5757 CV"
    3.  /paired#28#29parentheses  -> "paired( )parentheses"
    4.  /The_Key_of_F#23_Minor    -> "The_Key_of_F#_Minor"
    5.  /A#42                     -> "AB"
    6.  /Name1
    7.  /ASomewhatLongerName
    8.  /A;Name_With-Various***Characters?
    9.  /1.2
    10. /$$
    11. /@pattern
    12. /.notdef
    """

Answer 1

Have a look at re.sub .

You can use this with a function to match hex '#[0-9A-F]{2}' numbers and translate these using a function.

Eg

def hexrepl(m):
    return chr(int(m.group(0)[1:3],16))

re.sub(r'#[0-9A-F]{2}', hexrepl, '/Adobe#20Green')

Will return '/Adobe Green'

Answer 2

I'd use finditer() with a wrapper generator:

import re
from functools import partial

def _hexrepl(match):
    return chr(int(match.group(1), 16))
unescape = partial(re.compile(r'#([0-9A-F]{2})').sub, _hexrepl)

def pdfnames(inputtext):
    for match in Name.finditer(inputtext):
        yield unescape(match.group(0))

Demo:

>>> for name in pdfnames(names):
...     print name
... 
/Adobe Green
/PANTONE 5757 CV
/paired()parentheses
/The_Key_of_F#_Minor
/AB
/Name1
/ASomewhatLongerName
/A;Name_With-Various***Characters?
/1.2
/$$
/@pattern
/.notdef

There is no more clever way that I know of; the re engine cannot otherwise combine substitution and matching.