-// $Id: CQLParser.java,v 1.15 2002-11-06 00:05:58 mike Exp $
+// $Id: CQLParser.java,v 1.30 2007-06-28 00:24:48 mike Exp $
package org.z3950.zing.cql;
import java.io.IOException;
/**
* Compiles CQL strings into parse trees of CQLNode subtypes.
*
- * @version $Id: CQLParser.java,v 1.15 2002-11-06 00:05:58 mike Exp $
+ * @version $Id: CQLParser.java,v 1.30 2007-06-28 00:24:48 mike Exp $
* @see <A href="http://zing.z3950.org/cql/index.html"
* >http://zing.z3950.org/cql/index.html</A>
*/
* data structure) or, more often, simply rendered out in the
* desired form using one of the back-ends. <TT>toCQL()</TT>
* returns a decompiled CQL query equivalent to the one that was
- * compiled in the first place; and <TT>toXCQL()</TT> returns an
- * XML snippet representing the query.
+ * compiled in the first place; <TT>toXCQL()</TT> returns an
+ * XML snippet representing the query; and <TT>toPQF()</TT>
+ * returns the query rendered in Index Data's Prefix Query
+ * Format.
*
* @param cql The query
* @return A CQLNode object which is the root of a parse
lexer = new CQLLexer(cql, LEXDEBUG);
lexer.nextToken();
- debug("about to parse_query()");
- CQLNode root = parse_query("srw.serverChoice", new CQLRelation("="));
+ debug("about to parseQuery()");
+ CQLNode root = parseQuery("srw.serverChoice", new CQLRelation("scr"));
if (lexer.ttype != lexer.TT_EOF)
throw new CQLParseException("junk after end: " + lexer.render());
return root;
}
- private CQLNode parse_query(String qualifier, CQLRelation relation)
+ private CQLNode parseQuery(String index, CQLRelation relation)
throws CQLParseException, IOException {
- debug("in parse_query()");
+ debug("in parseQuery()");
- CQLNode term = parse_term(qualifier, relation);
+ CQLNode term = parseTerm(index, relation);
while (lexer.ttype != lexer.TT_EOF &&
lexer.ttype != ')') {
if (lexer.ttype == lexer.TT_AND) {
match(lexer.TT_AND);
- CQLNode term2 = parse_term(qualifier, relation);
+ CQLNode term2 = parseTerm(index, relation);
term = new CQLAndNode(term, term2);
} else if (lexer.ttype == lexer.TT_OR) {
match(lexer.TT_OR);
- CQLNode term2 = parse_term(qualifier, relation);
+ CQLNode term2 = parseTerm(index, relation);
term = new CQLOrNode(term, term2);
} else if (lexer.ttype == lexer.TT_NOT) {
match(lexer.TT_NOT);
- CQLNode term2 = parse_term(qualifier, relation);
+ CQLNode term2 = parseTerm(index, relation);
term = new CQLNotNode(term, term2);
} else if (lexer.ttype == lexer.TT_PROX) {
match(lexer.TT_PROX);
CQLProxNode proxnode = new CQLProxNode(term);
gatherProxParameters(proxnode);
- CQLNode term2 = parse_term(qualifier, relation);
+ CQLNode term2 = parseTerm(index, relation);
proxnode.addSecondSubterm(term2);
term = (CQLNode) proxnode;
} else {
return term;
}
- private CQLNode parse_term(String qualifier, CQLRelation relation)
+ private CQLNode parseTerm(String index, CQLRelation relation)
throws CQLParseException, IOException {
- debug("in parse_term()");
+ debug("in parseTerm()");
String word;
while (true) {
if (lexer.ttype == '(') {
debug("parenthesised term");
match('(');
- CQLNode expr = parse_query(qualifier, relation);
+ CQLNode expr = parseQuery(index, relation);
match(')');
return expr;
- } else if (lexer.ttype != lexer.TT_WORD &&
- lexer.ttype != lexer.TT_NUMBER &&
- lexer.ttype != '"') {
- throw new CQLParseException("expected qualifier or term, " +
- "got " + lexer.render());
+ } else if (lexer.ttype == '>') {
+ match('>');
+ return parsePrefix(index, relation);
}
debug("non-parenthesised term");
- if (lexer.ttype == lexer.TT_NUMBER) {
- word = lexer.render();
- } else {
- word = lexer.sval;
- }
- match(lexer.ttype);
- if (!isBaseRelation())
+ word = matchSymbol("index or term");
+ if (!isRelation() && lexer.ttype != lexer.TT_WORD)
break;
- qualifier = word;
- relation = new CQLRelation(lexer.render(lexer.ttype, false));
+ index = word;
+ relation = new CQLRelation(lexer.ttype == lexer.TT_WORD ?
+ lexer.sval :
+ lexer.render(lexer.ttype, false));
match(lexer.ttype);
while (lexer.ttype == '/') {
match('/');
- if (lexer.ttype != lexer.TT_RELEVANT &&
- lexer.ttype != lexer.TT_FUZZY &&
- lexer.ttype != lexer.TT_STEM)
+ if (lexer.ttype != lexer.TT_WORD)
throw new CQLParseException("expected relation modifier, "
+ "got " + lexer.render());
- relation.addModifier(lexer.sval);
+ String type = lexer.sval.toLowerCase();
match(lexer.ttype);
+ if (!isRelation()) {
+ // It's a simple modifier consisting of type only
+ relation.addModifier(type);
+ } else {
+ // It's a complex modifier of the form type=value
+ String comparision = lexer.render(lexer.ttype, false);
+ match(lexer.ttype);
+
+ // Yuck
+ String value = lexer.ttype == lexer.TT_WORD ? lexer.sval :
+ (double) lexer.nval == (int) lexer.nval ?
+ new Integer((int) lexer.nval).toString() :
+ new Double((double) lexer.nval).toString();
+
+ matchSymbol("relation-modifier value");
+ relation.addModifier(type, comparision, value);
+ }
}
- debug("qualifier='" + qualifier + ", " +
+ debug("index='" + index + ", " +
"relation='" + relation.toCQL() + "'");
}
- CQLTermNode node = new CQLTermNode(qualifier, relation, word);
+ CQLTermNode node = new CQLTermNode(index, relation, word);
debug("made term node " + node.toCQL());
return node;
}
+ private CQLNode parsePrefix(String index, CQLRelation relation)
+ throws CQLParseException, IOException {
+ debug("prefix mapping");
+
+ String name = null;
+ String identifier = matchSymbol("prefix-name");
+ if (lexer.ttype == '=') {
+ match('=');
+ name = identifier;
+ identifier = matchSymbol("prefix-identifer");
+ }
+ CQLNode term = parseQuery(index, relation);
+ return new CQLPrefixNode(name, identifier, term);
+ }
+
private void gatherProxParameters(CQLProxNode node)
throws CQLParseException, IOException {
for (int i = 0; i < 4; i++) {
private void gatherProxRelation(CQLProxNode node)
throws CQLParseException, IOException {
- if (!isProxRelation())
+ if (!isRelation())
throw new CQLParseException("expected proximity relation, got " +
lexer.render());
- node.addModifier("relation", lexer.render(lexer.ttype, false));
+ node.addModifier("relation", null, lexer.render(lexer.ttype, false));
match(lexer.ttype);
debug("gPR matched " + lexer.render(lexer.ttype, false));
}
if (lexer.ttype != lexer.TT_NUMBER)
throw new CQLParseException("expected proximity distance, got " +
lexer.render());
- node.addModifier("distance", lexer.render(lexer.ttype, false));
+ node.addModifier("distance", null, lexer.render(lexer.ttype, false));
match(lexer.ttype);
debug("gPD matched " + lexer.render(lexer.ttype, false));
}
lexer.ttype != lexer.TT_ELEMENT)
throw new CQLParseException("expected proximity unit, got " +
lexer.render());
- node.addModifier("unit", lexer.render());
+ node.addModifier("unit", null, lexer.render());
match(lexer.ttype);
}
lexer.ttype != lexer.TT_UNORDERED)
throw new CQLParseException("expected proximity ordering, got " +
lexer.render());
- node.addModifier("ordering", lexer.render());
+ node.addModifier("ordering", null, lexer.render());
match(lexer.ttype);
}
- private boolean isBaseRelation() {
- debug("isBaseRelation: checking ttype=" + lexer.ttype +
- " (" + lexer.render() + ")");
- return (isProxRelation() ||
- lexer.ttype == lexer.TT_ANY ||
- lexer.ttype == lexer.TT_ALL ||
- lexer.ttype == lexer.TT_EXACT);
- }
-
- private boolean isProxRelation() {
- debug("isProxRelation: checking ttype=" + lexer.ttype +
+ // Checks for a relation that may be used inside a prox operator
+ private boolean isRelation() {
+ debug("isRelation: checking ttype=" + lexer.ttype +
" (" + lexer.render() + ")");
return (lexer.ttype == '<' ||
lexer.ttype == '>' ||
" (tmp=" + tmp + ")");
}
+ private String matchSymbol(String expected)
+ throws CQLParseException, IOException {
+
+ debug("in matchSymbol()");
+ if (lexer.ttype == lexer.TT_WORD ||
+ lexer.ttype == lexer.TT_NUMBER ||
+ lexer.ttype == '"' ||
+ // The following is a complete list of keywords. Because
+ // they're listed here, they can be used unquoted as
+ // indexes, terms, prefix names and prefix identifiers.
+ // ### Instead, we should ask the lexer whether what we
+ // have is a keyword, and let the knowledge reside there.
+ lexer.ttype == lexer.TT_AND ||
+ lexer.ttype == lexer.TT_OR ||
+ lexer.ttype == lexer.TT_NOT ||
+ lexer.ttype == lexer.TT_PROX ||
+ lexer.ttype == lexer.TT_pWORD ||
+ lexer.ttype == lexer.TT_SENTENCE ||
+ lexer.ttype == lexer.TT_PARAGRAPH ||
+ lexer.ttype == lexer.TT_ELEMENT ||
+ lexer.ttype == lexer.TT_ORDERED ||
+ lexer.ttype == lexer.TT_UNORDERED) {
+ String symbol = (lexer.ttype == lexer.TT_NUMBER) ?
+ lexer.render() : lexer.sval;
+ match(lexer.ttype);
+ return symbol;
+ }
+
+ throw new CQLParseException("expected " + expected + ", " +
+ "got " + lexer.render());
+ }
+
/**
* Simple test-harness for the CQLParser class.
char mode = 'x'; // x=XCQL, c=CQL, p=PQF
String pfile = null;
- Vector argv = new Vector();
+ Vector<String> argv = new Vector<String>();
for (int i = 0; i < args.length; i++) {
argv.add(args[i]);
}
+ if (argv.size() > 0 && argv.get(0).equals("-d")) {
+ DEBUG = true;
+ argv.remove(0);
+ }
+
if (argv.size() > 0 && argv.get(0).equals("-c")) {
mode = 'c';
argv.remove(0);
}
if (argv.size() > 1) {
- System.err.println(
- "Usage: CQLParser [-c] [-p <pqf-properties> [<CQL-query>]");
+ System.err.println("Usage: CQLParser [-d] [-c] [-p <pqf-properties> [<CQL-query>]");
System.err.println("If unspecified, query is read from stdin");
System.exit(1);
}
try {
// Read in the whole of standard input in one go
int nbytes = System.in.read(bytes);
- } catch (java.io.IOException ex) {
+ } catch (IOException ex) {
System.err.println("Can't read query: " + ex.getMessage());
System.exit(2);
}
} catch (CQLParseException ex) {
System.err.println("Syntax error: " + ex.getMessage());
System.exit(3);
- } catch (java.io.IOException ex) {
+ } catch (IOException ex) {
System.err.println("Can't compile query: " + ex.getMessage());
System.exit(4);
}
} else {
System.out.print(root.toXCQL(0));
}
- } catch (java.io.IOException ex) {
+ } catch (IOException ex) {
System.err.println("Can't render query: " + ex.getMessage());
System.exit(5);
+ } catch (UnknownIndexException ex) {
+ System.err.println("Unknown index: " + ex.getMessage());
+ System.exit(6);
+ } catch (UnknownRelationException ex) {
+ System.err.println("Unknown relation: " + ex.getMessage());
+ System.exit(7);
+ } catch (UnknownRelationModifierException ex) {
+ System.err.println("Unknown relation modifier: " +
+ ex.getMessage());
+ System.exit(8);
+ } catch (UnknownPositionException ex) {
+ System.err.println("Unknown position: " + ex.getMessage());
+ System.exit(9);
+ } catch (PQFTranslationException ex) {
+ // We catch all of this class's subclasses, so --
+ throw new Error("can't get a PQFTranslationException");
}
}
}