java parsing of complex string? -
i getting string response server this..
[ anchor{anchorname='&loreal_lip_balm', clientname='loreal india', anchorprice=10, campaigns=[ campaign{ campaignid='loreal_camp1', question='how turn melted lipstick tinted lip balm', startdate=2015-08-04, enddate=2015-08-04, imageurl='null', regioncountry='all', rewardinfo='null', campaignprice=20, options=null } ]}, anchor{ anchorname='&loreal_total_repair_5', clientname='loreal india', anchorprice=125, campaigns=[ campaign{ campaignid='loreal_camp2', question='is product buy?', startdate=2015-08-04, enddate=2015-08-04, imageurl='null', regioncountry='all', rewardinfo='null', campaignprice=20, options=null } ] } ]. can tell me how parse this.it not json response.
i used hand coded parsers simple languages felt using http://www.antlr.org/ or https://javacc.java.net bit heavy. used same structure parsing json, css simple template. modified parse response. see whether helps.
package snippet; import java.io.ioexception; import java.io.pushbackreader; import java.io.reader; import java.io.stringreader; import java.text.parseexception; import java.text.simpledateformat; import java.util.arraylist; import java.util.arrays; import java.util.hashmap; import java.util.list; import java.util.map; import snippet.recursivedescentparser.token.tokentype; public class recursivedescentparser { // @formatter:off final static string text = "" + "[" + " anchor{anchorname='&loreal_lip_balm'," + " clientname='loreal india'," + "" + " anchorprice=10," + " campaigns=[" + " campaign{" + " campaignid='loreal_camp1'," + "" + " question='how turn melted lipstick tinted lip balm'," + "" + " startdate=2015-08-04," + " enddate=2015-08-04," + " imageurl='null'," + "" + " regioncountry='all'," + " rewardinfo='null'," + " campaignprice=20," + "" + " options=null" + " }" + " ]}," + " anchor{" + "" + " anchorname='&loreal_total_repair_5'," + " clientname='loreal india'," + " anchorprice=125," + "" + " campaigns=[" + " campaign{" + " campaignid='loreal_camp2'," + " question='is good" + " product buy?'," + " startdate=2015-08-04," + " enddate=2015-08-04," + "" + " imageurl='null'," + " regioncountry='all'," + " rewardinfo='null'," + "" + " campaignprice=20," + " options=null" + " }" + " ]" + " }" + " ]" + ""; // @formatter:on static class token { enum tokentype { open_bracket, close_bracket, open_brace, close_brace, string, name, comma, equals, integer, date, eof, null; } private string text; private tokentype type; public token(tokentype type) { this(type, null); } public token(tokentype type, string text) { this.type = type; this.text = text; } public tokentype gettype() { return type; } public string gettext() { return text; } @override public string tostring() { return "token [text=" + text + ", type=" + type + "]"; } } static class tokenreader { private pushbackreader reader; public tokenreader(reader reader) { this.reader = new pushbackreader(reader); } public token nexttoken() throws ioexception { token t = nexttokenx(); system.out.println("got: " + t); return t; } public token nexttokenx() throws ioexception { int c; while ((c = reader.read()) != -1 && character.iswhitespace((char) c)) ; if (c == -1) return new token(tokentype.eof); switch (c) { case '[': return new token(tokentype.open_bracket); case ']': return new token(tokentype.close_bracket); case '{': return new token(tokentype.open_brace); case '}': return new token(tokentype.close_brace); case ',': return new token(tokentype.comma); case '=': return new token(tokentype.equals); default: if (character.isdigit(c)) return readintegerordate(c); if (c == '\'') return readstring(c); if (character.isjavaidentifierstart(c)) return readname(c); throw new runtimeexception("invalid character '" + ((char) c) + "' in input"); } } private token readname(int c) throws ioexception { stringbuilder sb = new stringbuilder(); sb.append((char) c); while ((c = reader.read()) != -1 && character.isjavaidentifierpart(c)) sb.append((char) c); if (c != -1) reader.unread(c); if ("null".equals(sb.tostring())) return new token(tokentype.null); return new token(tokentype.name, sb.tostring()); } private token readstring(int end) throws ioexception { stringbuilder sb = new stringbuilder(); int c; while ((c = reader.read()) != -1 && c != end) sb.append((char) c); return new token(tokentype.string, sb.tostring()); } private token readintegerordate(int c) throws ioexception { stringbuilder sb = new stringbuilder(); sb.append((char) c); while ((c = reader.read()) != -1 && character.isdigit((char) c)) sb.append((char) c); if (c == '-') { sb.append((char) c); return readdate(sb); } if (c != -1) reader.unread(c); return new token(tokentype.integer, sb.tostring()); } private token readdate(stringbuilder sb) throws ioexception { int c; while ((c = reader.read()) != -1 && character.isdigit((char) c)) sb.append((char) c); if (c == -1) throw new runtimeexception("eof while reading date"); if (c != '-') throw new runtimeexception("invalid character '" + (char) c + "' while reading date"); sb.append((char) c); while ((c = reader.read()) != -1 && character.isdigit((char) c)) sb.append((char) c); if (c != -1) reader.unread(c); return new token(tokentype.date, sb.tostring()); } } static class lexer { private tokenreader reader; private token current; public lexer(reader reader) { this.reader = new tokenreader(reader); } public token expect(tokentype... tt) throws ioexception { if (current == null) current = reader.nexttoken(); (tokentype tokentype : tt) { if (current.gettype() == tokentype) { token r = current; current = null; return r; } } throw new runtimeexception("expecting 1 of " + arrays.aslist(tt) + " got " + current); } public token expect1or0(tokentype... tt) throws ioexception { if (current == null) current = reader.nexttoken(); (tokentype tokentype : tt) { if (current.gettype() == tokentype) { token r = current; current = null; return r; } } return null; } } public object parse(string text) throws ioexception { return parse(new stringreader(text)); } private object parse(reader reader) throws ioexception { lexer lexer = new lexer(reader); return parse(lexer); } private object parse(lexer lexer) throws ioexception { token t = lexer.expect1or0(tokentype.open_brace, tokentype.open_bracket, tokentype.eof); if (t == null || t.gettype() == tokentype.eof) return null; else if (t.gettype() == tokentype.open_bracket) { return parselist(lexer); } else { return parsemap(null, lexer); } } private list<object> parselist(lexer lexer) throws ioexception { arraylist<object> result = new arraylist<object>(); token tname = lexer.expect1or0(tokentype.name); while (tname != null) { lexer.expect(tokentype.open_brace); result.add(parsemap(tname.gettext(), lexer)); if (lexer.expect1or0(tokentype.comma) != null) tname = lexer.expect(tokentype.name); else tname = null; } lexer.expect(tokentype.close_bracket); return result; } private object parsemap(string oname, lexer lexer) throws ioexception { map<string, object> result = new hashmap<string, object>(); if (oname != null) result.put("objectname", oname); token tname = lexer.expect1or0(tokentype.name); while (tname != null) { string name = tname.gettext(); lexer.expect(tokentype.equals); token next = lexer.expect(tokentype.string, tokentype.date, tokentype.integer, tokentype.open_bracket, tokentype.open_brace, tokentype.null); tokentype tt = next.gettype(); if (tt == tokentype.string) { result.put(name, next.gettext()); } else if (tt == tokentype.date) { simpledateformat sdf = new simpledateformat("yyyy-mm-dd"); try { result.put(name, sdf.parse(next.gettext())); } catch (parseexception e) { return new runtimeexception(e); } } else if (tt == tokentype.integer) { result.put(name, integer.valueof(next.gettext())); } else if (tt == tokentype.open_bracket) { result.put(name, parselist(lexer)); } else if (tt == tokentype.open_brace) { result.put(name, parsemap(null, lexer)); } else { result.put(name, null); } if (lexer.expect1or0(tokentype.comma) != null) tname = lexer.expect(tokentype.name); else tname = null; } lexer.expect(tokentype.close_brace); return result; } public static void main(string[] args) throws ioexception { recursivedescentparser parser = new recursivedescentparser(); object o = parser.parse(text); system.out.println(o); } }
Comments
Post a Comment