Mercurial Hosting > luan
annotate src/goodjava/lucene/queryparser/GoodQueryParser.java @ 1822:aeaf2b65c005
better parse tables
author | Franklin Schmidt <fschmidt@gmail.com> |
---|---|
date | Tue, 09 Jul 2024 12:46:27 -0600 |
parents | d1e7564a9ce5 |
children |
rev | line source |
---|---|
1458 | 1 package goodjava.lucene.queryparser; |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
2 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
3 import java.util.List; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
4 import java.util.ArrayList; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
5 import java.util.regex.Pattern; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
6 import org.apache.lucene.search.Query; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
7 import org.apache.lucene.search.MatchAllDocsQuery; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
8 import org.apache.lucene.search.BooleanClause; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
9 import org.apache.lucene.search.BooleanQuery; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
10 import org.apache.lucene.search.Sort; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
11 import org.apache.lucene.search.SortField; |
1402
27efb1fcbcb5
move luan.lib to goodjava
Franklin Schmidt <fschmidt@gmail.com>
parents:
1362
diff
changeset
|
12 import goodjava.parser.Parser; |
27efb1fcbcb5
move luan.lib to goodjava
Franklin Schmidt <fschmidt@gmail.com>
parents:
1362
diff
changeset
|
13 import goodjava.parser.ParseException; |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
14 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
15 |
1459 | 16 public class GoodQueryParser { |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
17 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
18 public static Query parseQuery(FieldParser fieldParser,String query) throws ParseException { |
1459 | 19 return new GoodQueryParser(fieldParser,query).parseQuery(); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
20 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
21 |
1622 | 22 public static String quote(String s) { |
23 s = s.replace("\\","\\\\"); | |
24 s = s.replace("\b","\\b"); | |
25 s = s.replace("\f","\\f"); | |
26 s = s.replace("\n","\\n"); | |
27 s = s.replace("\r","\\r"); | |
28 s = s.replace("\t","\\t"); | |
29 s = s.replace("\"","\\\""); | |
30 return '"'+s+'"'; | |
31 } | |
32 | |
1619 | 33 public static String escape(String s) { |
1458 | 34 s = s.replace("\\","\\\\"); |
35 s = s.replace("\b","\\b"); | |
36 s = s.replace("\f","\\f"); | |
37 s = s.replace("\n","\\n"); | |
38 s = s.replace("\r","\\r"); | |
39 s = s.replace("\t","\\t"); | |
40 s = s.replace("\"","\\\""); | |
1621 | 41 s = s.replace("-","\\-"); |
1619 | 42 return s; |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
43 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
44 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
45 public static Sort parseSort(FieldParser fieldParser,String sort) throws ParseException { |
1459 | 46 return new GoodQueryParser(fieldParser,sort).parseSort(); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
47 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
48 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
49 |
1749
d1e7564a9ce5
improve query parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
1622
diff
changeset
|
50 private static final String NOT_IN_RANGE = " \t\r\n\":[]{}^()"; |
d1e7564a9ce5
improve query parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
1622
diff
changeset
|
51 private static final String NOT_IN_TERM = NOT_IN_RANGE; |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
52 private static final String NOT_IN_FIELD = NOT_IN_TERM + ","; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
53 private final FieldParser fieldParser; |
1110
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
54 private final Parser parser; |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
55 |
1459 | 56 private GoodQueryParser(FieldParser fieldParser,String query) { |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
57 this.fieldParser = fieldParser; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
58 this.parser = new Parser(query); |
1110
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
59 parser.begin(); |
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
60 } |
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
61 |
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
62 ParseException exception(String msg) { |
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
63 parser.failure(); |
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
64 return new ParseException(parser,msg); |
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
65 } |
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
66 |
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
67 ParseException exception(Exception cause) { |
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
68 parser.failure(); |
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
69 return new ParseException(parser,cause); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
70 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
71 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
72 private Query parseQuery() throws ParseException { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
73 Spaces(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
74 BooleanQuery bq = new BooleanQuery(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
75 while( !parser.endOfInput() ) { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
76 bq.add( Term(null) ); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
77 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
78 BooleanClause[] clauses = bq.getClauses(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
79 switch( clauses.length ) { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
80 case 0: |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
81 return new MatchAllDocsQuery(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
82 case 1: |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
83 { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
84 BooleanClause bc = clauses[0]; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
85 if( bc.getOccur() != BooleanClause.Occur.MUST_NOT ) |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
86 return bc.getQuery(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
87 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
88 default: |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
89 return bq; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
90 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
91 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
92 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
93 private BooleanClause Term(String defaultField) throws ParseException { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
94 BooleanClause.Occur occur; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
95 if( parser.match('+') ) { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
96 occur = BooleanClause.Occur.MUST; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
97 Spaces(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
98 } else if( parser.match('-') ) { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
99 occur = BooleanClause.Occur.MUST_NOT; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
100 Spaces(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
101 } else { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
102 occur = BooleanClause.Occur.SHOULD; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
103 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
104 String field = QueryField(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
105 if( field == null ) |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
106 field = defaultField; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
107 Query query = NestedTerm(field); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
108 if( query == null ) |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
109 query = RangeTerm(field); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
110 if( query == null ) { |
1572
45e42f403fb3
minor - lucene query parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
1564
diff
changeset
|
111 //parser.begin(); |
1362
77f2d091f17f
minor - SaneQueryParser
Franklin Schmidt <fschmidt@gmail.com>
parents:
1344
diff
changeset
|
112 String match = SimpleTerm(NOT_IN_TERM); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
113 query = fieldParser.getQuery(this,field,match); |
1572
45e42f403fb3
minor - lucene query parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
1564
diff
changeset
|
114 //parser.success(); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
115 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
116 if( parser.match('^') ) { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
117 Spaces(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
118 int start = parser.begin(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
119 try { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
120 while( parser.anyOf("0123456789.") ); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
121 String match = parser.textFrom(start); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
122 float boost = Float.parseFloat(match); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
123 query.setBoost(boost); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
124 } catch(NumberFormatException e) { |
1110
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
125 throw exception(e); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
126 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
127 parser.success(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
128 Spaces(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
129 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
130 BooleanClause bc = new BooleanClause(query,occur); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
131 return bc; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
132 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
133 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
134 private Query NestedTerm(String field) throws ParseException { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
135 parser.begin(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
136 if( !parser.match('(') ) |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
137 return parser.failure(null); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
138 BooleanQuery bq = new BooleanQuery(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
139 while( !parser.match(')') ) { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
140 if( parser.endOfInput() ) |
1110
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
141 throw exception("unclosed parentheses"); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
142 bq.add( Term(field) ); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
143 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
144 Spaces(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
145 BooleanClause[] clauses = bq.getClauses(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
146 switch( clauses.length ) { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
147 case 0: |
1110
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
148 throw exception("empty parentheses"); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
149 case 1: |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
150 { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
151 BooleanClause bc = clauses[0]; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
152 if( bc.getOccur() != BooleanClause.Occur.MUST_NOT ) |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
153 return parser.success(bc.getQuery()); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
154 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
155 default: |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
156 return parser.success(bq); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
157 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
158 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
159 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
160 private Query RangeTerm(String field) throws ParseException { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
161 parser.begin(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
162 if( !parser.anyOf("[{") ) |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
163 return parser.failure(null); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
164 boolean includeMin = parser.lastChar() == '['; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
165 Spaces(); |
1362
77f2d091f17f
minor - SaneQueryParser
Franklin Schmidt <fschmidt@gmail.com>
parents:
1344
diff
changeset
|
166 String minQuery = SimpleTerm(NOT_IN_RANGE); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
167 TO(); |
1362
77f2d091f17f
minor - SaneQueryParser
Franklin Schmidt <fschmidt@gmail.com>
parents:
1344
diff
changeset
|
168 String maxQuery = SimpleTerm(NOT_IN_RANGE); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
169 if( !parser.anyOf("]}") ) |
1110
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
170 throw exception("unclosed range"); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
171 boolean includeMax = parser.lastChar() == ']'; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
172 Spaces(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
173 Query query = fieldParser.getRangeQuery(this,field,minQuery,maxQuery,includeMin,includeMax); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
174 return parser.success(query); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
175 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
176 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
177 private void TO() throws ParseException { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
178 parser.begin(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
179 if( !(parser.match("TO") && Space()) ) |
1110
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
180 throw exception("'TO' expected"); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
181 Spaces(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
182 parser.success(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
183 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
184 |
1362
77f2d091f17f
minor - SaneQueryParser
Franklin Schmidt <fschmidt@gmail.com>
parents:
1344
diff
changeset
|
185 private String SimpleTerm(String exclude) throws ParseException { |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
186 parser.begin(); |
1422 | 187 String match = Quoted(); |
188 if( match==null ) | |
1362
77f2d091f17f
minor - SaneQueryParser
Franklin Schmidt <fschmidt@gmail.com>
parents:
1344
diff
changeset
|
189 match = Unquoted(exclude); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
190 if( match.length() == 0 ) |
1110
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
191 throw exception("invalid input"); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
192 return parser.success(match); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
193 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
194 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
195 private String QueryField() throws ParseException { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
196 parser.begin(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
197 String match = Field(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
198 if( match==null || !parser.match(':') ) |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
199 return parser.failure((String)null); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
200 Spaces(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
201 return parser.success(match); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
202 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
203 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
204 private String Field() throws ParseException { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
205 parser.begin(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
206 String match = Unquoted(NOT_IN_FIELD); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
207 if( match.length()==0 ) |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
208 return parser.failure((String)null); |
1619 | 209 match = StringFieldParser.unescape(this,match); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
210 return parser.success(match); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
211 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
212 |
1422 | 213 private String Quoted() throws ParseException { |
214 parser.begin(); | |
215 if( !parser.match('"') ) | |
216 return parser.failure(null); | |
217 StringBuilder sb = new StringBuilder(); | |
218 while( parser.anyChar() ) { | |
219 char c = parser.lastChar(); | |
220 switch(c) { | |
221 case '"': | |
1564
b1195cfe8712
lucene/queryparser fix
Franklin Schmidt <fschmidt@gmail.com>
parents:
1459
diff
changeset
|
222 Spaces(); |
1422 | 223 return parser.success(sb.toString()); |
224 case '\\': | |
225 if( parser.anyChar() ) { | |
226 c = parser.lastChar(); | |
227 switch(c) { | |
228 case '"': | |
229 case '\\': | |
230 sb.append(c); | |
231 continue; | |
232 case 'b': | |
233 sb.append('\b'); | |
234 continue; | |
235 case 'f': | |
236 sb.append('\f'); | |
237 continue; | |
238 case 'n': | |
239 sb.append('\n'); | |
240 continue; | |
241 case 'r': | |
242 sb.append('\r'); | |
243 continue; | |
244 case 't': | |
245 sb.append('\t'); | |
246 continue; | |
247 case 'u': | |
248 int n = 0; | |
249 for( int i=0; i<4; i++ ) { | |
250 int d; | |
251 if( parser.inCharRange('0','9') ) { | |
252 d = parser.lastChar() - '0'; | |
253 } else if( parser.inCharRange('a','f') ) { | |
254 d = parser.lastChar() - 'a' + 10; | |
255 } else if( parser.inCharRange('A','F') ) { | |
256 d = parser.lastChar() - 'A' + 10; | |
257 } else { | |
258 throw exception("invalid hex digit"); | |
259 } | |
260 n = 16*n + d; | |
261 } | |
262 sb.append((char)n); | |
263 continue; | |
264 } | |
265 } | |
266 throw exception("invalid escape char"); | |
267 default: | |
268 sb.append(c); | |
269 } | |
270 } | |
271 parser.failure(); | |
272 throw exception("unclosed string"); | |
273 } | |
274 | |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
275 private String Unquoted(String exclude) throws ParseException { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
276 int start = parser.begin(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
277 while( parser.noneOf(exclude) ) { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
278 checkEscape(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
279 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
280 String match = parser.textFrom(start); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
281 Spaces(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
282 return parser.success(match); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
283 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
284 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
285 private void checkEscape() { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
286 if( parser.lastChar() == '\\' ) |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
287 parser.anyChar(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
288 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
289 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
290 private void Spaces() { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
291 while( Space() ); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
292 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
293 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
294 private boolean Space() { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
295 return parser.anyOf(" \t\r\n"); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
296 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
297 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
298 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
299 // sort |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
300 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
301 private Sort parseSort() throws ParseException { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
302 Spaces(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
303 if( parser.endOfInput() ) |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
304 return null; |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
305 List<SortField> list = new ArrayList<SortField>(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
306 list.add( SortField() ); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
307 while( !parser.endOfInput() ) { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
308 parser.begin(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
309 if( !parser.match(',') ) |
1110
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
310 throw exception("',' expected"); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
311 Spaces(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
312 parser.success(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
313 list.add( SortField() ); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
314 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
315 return new Sort(list.toArray(new SortField[0])); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
316 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
317 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
318 private SortField SortField() throws ParseException { |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
319 parser.begin(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
320 String field = Field(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
321 if( field==null ) |
1110
38a42f437fd2
queryparser now uses parsers.Parser
Franklin Schmidt <fschmidt@gmail.com>
parents:
775
diff
changeset
|
322 throw exception("invalid input"); |
730
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
323 boolean reverse = !parser.matchIgnoreCase("asc") && parser.matchIgnoreCase("desc"); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
324 Spaces(); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
325 SortField sf = fieldParser.getSortField(this,field,reverse); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
326 return parser.success(sf); |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
327 } |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
328 |
01e68da6983b
add sane-lucene-queryparser source to luan
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff
changeset
|
329 } |