parses config files

config.py

import os
import re
import json
import shlex
import types
import ConfigParser

_tokens = []
data    = {'global': {}}
regex   = {
    'section': re.compile('^\[[a-zA-Z0-9_]+\]$'),
    'string':  re.compile('^(\'.*\'|\".*\")$'),
    'key':     re.compile('^[a-zA-Z_]+:$'),
    'macro':   re.compile('^\$\{.*\}$')
}
    
def loadConfig(cfgfile):
    global _tokens, data, regex
    
    if not cfgfile:
        raise ValueError, 'You must specify a path+filename.'

    # We have a 3 pass parser here.
    # first pass, pull in the file and parse out comments
    _slurpFile(cfgfile)
    # second pass, actually parse the file into the data structure
    _parseFile()
    # and third pass, ensure all macro interpolation is handled
    tmp             = _interpolate(data)
    data       = tmp

    from pprint import pprint as pp
    pp(data)
        
def _interpolate(node):
    global _tokens, data, regex
        
    for key in node.keys():
        if type(node[key]) == types.DictType:
            t = _interpolate(node[key])
            node[key] = t
        else:
            if regex['macro'].match(node[key]):
                if node[key][2:-1].find(':') != -1:
                    k,v = node[key][2:-1].split(':')
                    node[key] = data[k][v]
                else:
                    node[key] = node[node[key][2:-1]]

    return node
                    

def _parseFile():
    global _tokens, data, regex
    
    section         = 'global'        
    for idx in range(0, len(_tokens)):
        if not regex['string'].match(_tokens[idx]):
            if regex['section'].match(_tokens[idx]):
                if _tokens[idx][1:-1] not in data.keys():
                    section = _tokens[idx][1:-1]
                    data[section] = {}
            elif regex['key'].match(_tokens[idx]):
                try:
                    if regex['string'].match(_tokens[idx+1]):
                        data[section][_tokens[idx][0:-1]] = json.loads(_tokens[idx+1][1:-1])
                    else:
                        data[section][_tokens[idx][0:-1]] = json.loads(_tokens[idx+1])
                except ValueError:
                    print section, idx, _tokens[idx]
                    data[section][_tokens[idx][0:-1]] = _tokens[idx+1]

def _slurpFile(cfgfile):
    global _tokens, data, regex
        
    if not os.path.isfile(cfgfile):
        raise ValueError, '%s: no such file' % cfgfile
        
    _cfgfp  = open(cfgfile)
    buff         = []
    tbuf         = _cfgfp.readline()

    # make sure we don't keep comments around
    # TODO: support end of line comments
    while tbuf:
        if tbuf[0] != '#':
            buff.append(tbuf)
        tbuf     = _cfgfp.readline()

    # Hand off a shlex parsed set of tokens for pass 2
    _tokens = shlex.split(''.join(buff))