/**
* 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;
}
/**
* 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(final String string) throws JSONException {
final JSONObject o = new JSONObject();
final XMLTokener x = new XMLTokener(string);
while (x.more() && x.skipPast("<")) {
parse(x, o, null);
}
return o;
}
/**
* 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.
* @param keepStrings Don't type-convert text nodes and attibute values
* @return A JSONArray if the value is the outermost tag, otherwise null.
* @throws JSONException
*/
private static Object parse(XMLTokener x, boolean arrayForm, JSONArray ja, boolean keepStrings)
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, keepStrings ? token : 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, keepStrings);
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
? keepStrings ? token : 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, JSONObject.stringToValue((String) token));
token = null;
} else {
jsonobject.accumulate(string, "");
}
} else if (token == SLASH) {
// Empty tag <.../>
if (x.nextToken() != GT) {
throw x.syntaxError("Misshaped tag");
}
if (jsonobject.length() > 0) {
context.accumulate(tagName, jsonobject);
} else {
context.accumulate(tagName, "");
}
return false;
} else if (token == GT) {
// Content, between <...> and </...>
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", JSONObject.stringToValue(string));
}
} else if (token == LT) {
// Nested element
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");
}
}
}
}
/* */ private static JSONArray parse(XMLTokener x, JSONArray ja)
/* */ throws JSONException
/* */ {
/* */ while (true)
/* */ {
/* 63 */ Object t = x.nextContent();
/* 64 */ if (t == XML.LT) {
/* 65 */ t = x.nextToken();
/* 66 */ if ((t instanceof Character))
/* */ {
/* 70 */ if (t == XML.BANG) {
/* 71 */ char c = x.next();
/* 72 */ if (c == '-') {
/* 73 */ if (x.next() == '-') {
/* 74 */ x.skipPast("-->");
/* */ }
/* 76 */ x.back();
continue;
/* 77 */ }
if (c == '[') {
/* 78 */ t = x.nextToken();
/* 79 */ if ((t.equals("CDATA")) && (x.next() == '[')) {
/* 80 */ x.nextCDATA();
continue;
/* */ }
/* 82 */ throw x.syntaxError("Expected 'CDATA['");
/* */ }
/* */
/* 85 */ int i = 1;
/* */ do {
/* 87 */ t = x.nextMeta();
/* 88 */ if (t == null) /* 89 */ throw x.syntaxError("Missing '>' after '<!'.");
/* 90 */ if (t == XML.LT) /* 91 */ i++;
/* 92 */ else if (t == XML.GT) /* 93 */ i--;
/* */ }
/* 95 */ while (i > 0);
continue;
/* */ }
/* 97 */ if (t == XML.QUEST)
/* */ {
/* 101 */ x.skipPast("?>");
continue;
/* 102 */ }
if (t == XML.SLASH)
/* */ {
/* 106 */ t = x.nextToken();
/* 107 */ if (ja == null) {
/* 108 */ throw x.syntaxError("Mismatched close tag '" + t + "'");
/* */ }
/* 110 */ if (!t.equals(ja.get(0))) {
/* 111 */ throw x.syntaxError("Mismatched '" + ja.get(0) + "' and '" + t + "'");
/* */ }
/* 113 */ if (x.nextToken() != XML.GT) {
/* 114 */ throw x.syntaxError("Misshaped close tag");
/* */ }
/* 116 */ return null;
/* */ }
/* 118 */ throw x.syntaxError("Misshaped tag");
/* */ }
/* */
/* 124 */ JSONArray newja = new JSONArray();
/* 125 */ JSONObject attributes = new JSONObject();
/* 126 */ if (ja != null) {
/* 127 */ ja.put(newja);
/* */ }
/* 129 */ newja.put(t);
/* 130 */ t = null;
/* */ while (true) {
/* 132 */ if (t == null) {
/* 133 */ t = x.nextToken();
/* */ }
/* 135 */ if (t == null) {
/* 136 */ throw x.syntaxError("Misshaped tag");
/* */ }
/* 138 */ if (!(t instanceof String))
/* */ {
/* */ break;
/* */ }
/* */
/* 144 */ String s = (String) t;
/* 145 */ t = x.nextToken();
/* 146 */ if (t == XML.EQ) {
/* 147 */ t = x.nextToken();
/* 148 */ if (!(t instanceof String)) {
/* 149 */ throw x.syntaxError("Missing value");
/* */ }
/* 151 */ attributes.accumulate(s, t);
/* 152 */ t = null;
continue;
/* */ }
/* 154 */ attributes.accumulate(s, "");
/* */ }
/* */
/* 157 */ if (attributes.length() > 0) {
/* 158 */ newja.put(attributes);
/* */ }
/* */
/* 163 */ if (t == XML.SLASH) {
/* 164 */ if (x.nextToken() != XML.GT) {
/* 165 */ throw x.syntaxError("Misshaped tag");
/* */ }
/* 167 */ if (ja == null) {
/* 168 */ return newja;
/* */ }
/* */
/* */ }
/* */
/* 173 */ if (t == XML.GT) {
/* 174 */ parse(x, newja);
/* 175 */ if (ja == null) {
/* 176 */ return newja;
/* */ }
/* */ }
/* 179 */ throw x.syntaxError("Misshaped tag");
/* */ }
/* */
/* 183 */ if (ja != null) /* 184 */ ja.put(t);
/* */ }
/* */ }
/**
* 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(final XMLTokener x, final JSONObject context, final 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, t);
t = null;
} else {
o.accumulate(s, "");
}
// Empty tag <.../>
} else if (t == SLASH) {
if (x.nextToken() != GT) {
throw x.syntaxError("Misshaped tag");
}
context.accumulate(n, o);
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", 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");
}
}
}
}
