/**
* Convert a well-formed (but not necessarily valid) XML string into a JSONObject. Some
* information may be lost in this transformation because JSON is a data format and XML is a
* document format. XML uses elements, attributes, and content text, while JSON uses unordered
* collections of name/value pairs and arrays of values. JSON does not does not like to
* distinguish between elements and attributes. Sequences of similar elements are represented as
* JSONArrays. Content text may be placed in a "content" member. Comments, prologs, DTDs, and
* <code><[ [ ]]></code> are ignored.
*
* @param string The source string.
* @return A JSONObject containing the structured data from the XML string.
* @throws JSONException
*/
public static JSONObject toJSONObject(String string) throws JSONException {
JSONObject jo = new JSONObject();
XMLTokener x = new XMLTokener(string);
while (x.more() && x.skipPast("<")) {
parse(x, jo, null);
}
return jo;
}
/**
* Parse XML values and store them in a JSONArray.
*
* @param x The XMLTokener containing the source string.
* @param arrayForm true if array form, false if object form.
* @param ja The JSONArray that is containing the current tag or null if we are at the outermost
* level.
* @return A JSONArray if the value is the outermost tag, otherwise null.
* @throws JSONException
*/
private static Object parse(XMLTokener x, boolean arrayForm, JSONArray ja) throws JSONException {
String attribute;
char c;
String closeTag = null;
int i;
JSONArray newja = null;
JSONObject newjo = null;
Object token;
String tagName = null;
// Test for and skip past these forms:
// <!-- ... -->
// <![ ... ]]>
// <! ... >
// <? ... ?>
while (true) {
if (!x.more()) {
throw x.syntaxError("Bad XML");
}
token = x.nextContent();
if (token == XML.LT) {
token = x.nextToken();
if (token instanceof Character) {
if (token == XML.SLASH) {
// Close tag </
token = x.nextToken();
if (!(token instanceof String)) {
throw new JSONException("Expected a closing name instead of '" + token + "'.");
}
if (x.nextToken() != XML.GT) {
throw x.syntaxError("Misshaped close tag");
}
return token;
} else if (token == XML.BANG) {
// <!
c = x.next();
if (c == '-') {
if (x.next() == '-') {
x.skipPast("-->");
} else {
x.back();
}
} else if (c == '[') {
token = x.nextToken();
if (token.equals("CDATA") && x.next() == '[') {
if (ja != null) {
ja.put(x.nextCDATA());
}
} else {
throw x.syntaxError("Expected 'CDATA['");
}
} else {
i = 1;
do {
token = x.nextMeta();
if (token == null) {
throw x.syntaxError("Missing '>' after '<!'.");
} else if (token == XML.LT) {
i += 1;
} else if (token == XML.GT) {
i -= 1;
}
} while (i > 0);
}
} else if (token == XML.QUEST) {
// <?
x.skipPast("?>");
} else {
throw x.syntaxError("Misshaped tag");
}
// Open tag <
} else {
if (!(token instanceof String)) {
throw x.syntaxError("Bad tagName '" + token + "'.");
}
tagName = (String) token;
newja = new JSONArray();
newjo = new JSONObject();
if (arrayForm) {
newja.put(tagName);
if (ja != null) {
ja.put(newja);
}
} else {
newjo.put("tagName", tagName);
if (ja != null) {
ja.put(newjo);
}
}
token = null;
for (; ; ) {
if (token == null) {
token = x.nextToken();
}
if (token == null) {
throw x.syntaxError("Misshaped tag");
}
if (!(token instanceof String)) {
break;
}
// attribute = value
attribute = (String) token;
if (!arrayForm && ("tagName".equals(attribute) || "childNode".equals(attribute))) {
throw x.syntaxError("Reserved attribute.");
}
token = x.nextToken();
if (token == XML.EQ) {
token = x.nextToken();
if (!(token instanceof String)) {
throw x.syntaxError("Missing value");
}
newjo.accumulate(attribute, XML.stringToValue((String) token));
token = null;
} else {
newjo.accumulate(attribute, "");
}
}
if (arrayForm && newjo.length() > 0) {
newja.put(newjo);
}
// Empty tag <.../>
if (token == XML.SLASH) {
if (x.nextToken() != XML.GT) {
throw x.syntaxError("Misshaped tag");
}
if (ja == null) {
if (arrayForm) {
return newja;
} else {
return newjo;
}
}
// Content, between <...> and </...>
} else {
if (token != XML.GT) {
throw x.syntaxError("Misshaped tag");
}
closeTag = (String) parse(x, arrayForm, newja);
if (closeTag != null) {
if (!closeTag.equals(tagName)) {
throw x.syntaxError("Mismatched '" + tagName + "' and '" + closeTag + "'");
}
tagName = null;
if (!arrayForm && newja.length() > 0) {
newjo.put("childNodes", newja);
}
if (ja == null) {
if (arrayForm) {
return newja;
} else {
return newjo;
}
}
}
}
}
} else {
if (ja != null) {
ja.put(token instanceof String ? XML.stringToValue((String) token) : token);
}
}
}
}
/**
* Scan the content following the named tag, attaching it to the context.
*
* @param x The XMLTokener containing the source string.
* @param context The JSONObject that will include the new material.
* @param name The tag name.
* @return true if the close tag is processed.
* @throws JSONException
*/
private static boolean parse(XMLTokener x, JSONObject context, String name) throws JSONException {
char c;
int i;
JSONObject jsonobject = null;
String string;
String tagName;
Object token;
// Test for and skip past these forms:
// <!-- ... -->
// <! ... >
// <![ ... ]]>
// <? ... ?>
// Report errors for these forms:
// <>
// <=
// <<
token = x.nextToken();
// <!
if (token == BANG) {
c = x.next();
if (c == '-') {
if (x.next() == '-') {
x.skipPast("-->");
return false;
}
x.back();
} else if (c == '[') {
token = x.nextToken();
if ("CDATA".equals(token)) {
if (x.next() == '[') {
string = x.nextCDATA();
if (string.length() > 0) {
context.accumulate("content", string);
}
return false;
}
}
throw x.syntaxError("Expected 'CDATA['");
}
i = 1;
do {
token = x.nextMeta();
if (token == null) {
throw x.syntaxError("Missing '>' after '<!'.");
} else if (token == LT) {
i += 1;
} else if (token == GT) {
i -= 1;
}
} while (i > 0);
return false;
} else if (token == QUEST) {
// <?
x.skipPast("?>");
return false;
} else if (token == SLASH) {
// Close tag </
token = x.nextToken();
if (name == null) {
throw x.syntaxError("Mismatched close tag " + token);
}
if (!token.equals(name)) {
throw x.syntaxError("Mismatched " + name + " and " + token);
}
if (x.nextToken() != GT) {
throw x.syntaxError("Misshaped close tag");
}
return true;
} else if (token instanceof Character) {
throw x.syntaxError("Misshaped tag");
// Open tag <
} else {
tagName = (String) token;
token = null;
jsonobject = new JSONObject();
for (; ; ) {
if (token == null) {
token = x.nextToken();
}
// attribute = value
if (token instanceof String) {
string = (String) token;
token = x.nextToken();
if (token == EQ) {
token = x.nextToken();
if (!(token instanceof String)) {
throw x.syntaxError("Missing value");
}
jsonobject.accumulate(string, XML.stringToValue((String) token));
token = null;
} else {
jsonobject.accumulate(string, "");
}
// Empty tag <.../>
} else if (token == SLASH) {
if (x.nextToken() != GT) {
throw x.syntaxError("Misshaped tag");
}
if (jsonobject.length() > 0) {
context.accumulate(tagName, jsonobject);
} else {
context.accumulate(tagName, "");
}
return false;
// Content, between <...> and </...>
} else if (token == GT) {
for (; ; ) {
token = x.nextContent();
if (token == null) {
if (tagName != null) {
throw x.syntaxError("Unclosed tag " + tagName);
}
return false;
} else if (token instanceof String) {
string = (String) token;
if (string.length() > 0) {
jsonobject.accumulate("content", XML.stringToValue(string));
}
// Nested element
} else if (token == LT) {
if (parse(x, jsonobject, tagName)) {
if (jsonobject.length() == 0) {
context.accumulate(tagName, "");
} else if (jsonobject.length() == 1 && jsonobject.opt("content") != null) {
context.accumulate(tagName, jsonobject.opt("content"));
} else {
context.accumulate(tagName, jsonobject);
}
return false;
}
}
}
} else {
throw x.syntaxError("Misshaped tag");
}
}
}
}
/**
* Scan the content following the named tag, attaching it to the context.
*
* @param x The XMLTokener containing the source string.
* @param context The JSONObject that will include the new material.
* @param name The tag name.
* @return true if the close tag is processed.
* @throws JSONException
*/
private static boolean parse(XMLTokener x, JSONObject context, String name) throws JSONException {
char c;
int i;
String n;
JSONObject o = null;
String s;
Object t;
// Test for and skip past these forms:
// <!-- ... -->
// <! ... >
// <![ ... ]]>
// <? ... ?>
// Report errors for these forms:
// <>
// <=
// <<
t = x.nextToken();
// <!
if (t == BANG) {
c = x.next();
if (c == '-') {
if (x.next() == '-') {
x.skipPast("-->");
return false;
}
x.back();
} else if (c == '[') {
t = x.nextToken();
if (t.equals("CDATA")) {
if (x.next() == '[') {
s = x.nextCDATA();
if (s.length() > 0) {
context.accumulate("content", s);
}
return false;
}
}
throw x.syntaxError("Expected 'CDATA['");
}
i = 1;
do {
t = x.nextMeta();
if (t == null) {
throw x.syntaxError("Missing '>' after '<!'.");
} else if (t == LT) {
i += 1;
} else if (t == GT) {
i -= 1;
}
} while (i > 0);
return false;
} else if (t == QUEST) {
// <?
x.skipPast("?>");
return false;
} else if (t == SLASH) {
// Close tag </
t = x.nextToken();
if (name == null) {
throw x.syntaxError("Mismatched close tag" + t);
}
if (!t.equals(name)) {
throw x.syntaxError("Mismatched " + name + " and " + t);
}
if (x.nextToken() != GT) {
throw x.syntaxError("Misshaped close tag");
}
return true;
} else if (t instanceof Character) {
throw x.syntaxError("Misshaped tag");
// Open tag <
} else {
n = (String) t;
t = null;
o = new JSONObject();
for (; ; ) {
if (t == null) {
t = x.nextToken();
}
// attribute = value
if (t instanceof String) {
s = (String) t;
t = x.nextToken();
if (t == EQ) {
t = x.nextToken();
if (!(t instanceof String)) {
throw x.syntaxError("Missing value");
}
o.accumulate(s, JSONObject.stringToValue((String) t));
t = null;
} else {
o.accumulate(s, "");
}
// Empty tag <.../>
} else if (t == SLASH) {
if (x.nextToken() != GT) {
throw x.syntaxError("Misshaped tag");
}
if (o.length() > 0) {
context.accumulate(n, o);
} else {
context.accumulate(n, "");
}
return false;
// Content, between <...> and </...>
} else if (t == GT) {
for (; ; ) {
t = x.nextContent();
if (t == null) {
if (n != null) {
throw x.syntaxError("Unclosed tag " + n);
}
return false;
} else if (t instanceof String) {
s = (String) t;
if (s.length() > 0) {
o.accumulate("content", JSONObject.stringToValue(s));
}
// Nested element
} else if (t == LT) {
if (parse(x, o, n)) {
if (o.length() == 0) {
context.accumulate(n, "");
} else if (o.length() == 1 && o.opt("content") != null) {
context.accumulate(n, o.opt("content"));
} else {
context.accumulate(n, o);
}
return false;
}
}
}
} else {
throw x.syntaxError("Misshaped tag");
}
}
}
}
