Convert XML to JSON using Python 2.7.x

convert.py

#!/usr/bin/python

import xml.etree.ElementTree as ET;
import json, os, fnmatch, time;


def coerceStringToType(val):
    if not val:
        return None;

    try:
        f = float(val);
        i = int(f);
        if f == i:
            return i;
        else:
            return f;
    except (ValueError, TypeError) as e:
        if val.lower() == "true":
            return True;
        if val.lower() == "false":
            return False;
        return val;


def coerceArrayOfStrings(parsed):
    ret = [];
    hasString = False;
    for elt in parsed:
        val = coerceStringToType(elt.strip());

        # This is the case where the string is "...,foo,,bar,..."
        if val == None:
            print "Bad string cooersion:", parsed;
            sys.exit(1);

        # If there are any strings, treat all elements as strings
        if isinstance(val, str) and not hasString:
            hasString = True;
            for i in range(0, len(ret)):
                ret[i] = str(ret[i]);

        if hasString:
            ret.append(str(val));
        else:
            ret.append(val);

    if len(ret) == 1:
        return ret[0];

    return ret;


def coerceUnknownValue(val):
    # Don't think this can happen, but just in case
    if not isinstance(val, str):
        return val;

    val = val.strip();

    # No empty strings
    if not val:
        return None;

    return coerceArrayOfStrings(val.split(','));


def coerceKeyName(name):
    lowerFirst = lambda s: s[:1].lower() + s[1:] if s else '';

    if name.lower() == "id":
        return name.lower();

    return lowerFirst(name);


def parseNode(root):
    out = {};

    for child in root:
        newName = coerceKeyName(child.tag);

        if newName in out:
            if not isinstance(out[newName], list):
                out[newName] = list(out[newName]);
            out[newName].append(child);
        else:
            out[newName] = child;

    for key in out:
        if isinstance(out[key], list):
            for i in range(0, len(out[key])):
                out[key][i] = parseNode(out[key][i]);
        else:
            out[key] = parseNode(out[key]);

    for attr in root.attrib:
        out[coerceKeyName(attr)] = coerceUnknownValue(root.attrib[attr]);

    if root.text != None and root.text.strip():
        val = coerceUnknownValue(root.text);
        if len(out) < 1:
            out = val; 
        else:
            out["value"] = val;

    if isinstance(out, dict) or isinstance(out, list):
        if len(out) < 1:
            out = None;

    return out;


def parseFile(fileName):
    tree = ET.parse(fileName);
    return parseNode(tree.getroot());


if __name__ == "__main__":
    start = time.clock();

    count = 0;

    for root, dirs, files in os.walk('.'):
        for file in files:
            if fnmatch.fnmatch(file, '*.xml'):
                out = parseFile(root + '/' + file);
                f = open(root + '/' + file.split('.')[0] + '.json', 'w');
                json.dump(out, f, indent=2, sort_keys=True);
                f.close();
                count = count + 1;

    end = time.clock();

    print "Processed", count, "files in", (end - start), "seconds.";