-// $Id: CQLParser.java,v 1.6 2002-10-25 16:11:05 mike Exp $
+// $Id: CQLParser.java,v 1.10 2002-10-30 09:19:26 mike Exp $
package org.z3950.zing.cql;
-import java.util.Properties;
-import java.io.InputStream;
-import java.io.FileNotFoundException;
import java.io.IOException;
-import java.io.StringReader;
-import java.io.StreamTokenizer;
/**
- * Compiles a CQL string into a parse tree ...
+ * Compiles a CQL string into a parse tree.
* ###
*
- * @version $Id: CQLParser.java,v 1.6 2002-10-25 16:11:05 mike Exp $
+ * @version $Id: CQLParser.java,v 1.10 2002-10-30 09:19:26 mike Exp $
* @see <A href="http://zing.z3950.org/cql/index.html"
* >http://zing.z3950.org/cql/index.html</A>
*/
public class CQLParser {
- private String cql;
- private StreamTokenizer st;
+ private CQLLexer lexer;
+ static private boolean DEBUG = false;
+ static private boolean LEXDEBUG = false;
- private class CQLParseException extends Exception {
- CQLParseException(String s) { super(s); }
- }
-
- public CQLParser() {
- // Nothing to do: do we need this constructor, then?
+ private static void debug(String str) {
+ if (DEBUG)
+ System.err.println("PARSEDEBUG: " + str);
}
public CQLNode parse(String cql)
- throws FileNotFoundException, IOException {
- this.cql = cql;
- st = new StreamTokenizer(new StringReader(cql));
- // ### these settings are wrong
- st.wordChars('/', '/');
- st.wordChars('0', '9'); // ### but 1 is still recognised as TT_NUM
- st.wordChars('.', '.');
- st.wordChars('-', '-');
- st.ordinaryChar('=');
- st.ordinaryChar(',');
- st.ordinaryChar('(');
- st.ordinaryChar(')');
-
-// int token;
-// while ((token = st.nextToken()) != st.TT_EOF) {
-// System.out.println("token=" + token + ", " +
-// "nval=" + st.nval + ", " +
-// "sval=" + st.sval);
-// }
-
- st.nextToken();
- CQLNode root;
- try {
- root = parse_expression();
- } catch (CQLParseException ex) {
- System.err.println("### Oops: " + ex);
- return null;
- }
+ throws CQLParseException, IOException {
+ lexer = new CQLLexer(cql, LEXDEBUG);
- if (st.ttype != st.TT_EOF) {
- System.err.println("### Extra bits: " + render(st));
- return null;
- }
+ lexer.nextToken();
+ debug("about to parse_query()");
+ CQLNode root = parse_query("srw.serverChoice", new CQLRelation("="));
+ if (lexer.ttype != lexer.TT_EOF)
+ throw new CQLParseException("junk after end: " + lexer.render());
return root;
}
- private CQLNode parse_expression()
+ private CQLNode parse_query(String qualifier, CQLRelation relation)
throws CQLParseException, IOException {
- CQLNode term = parse_term();
-
- while (st.ttype == st.TT_WORD) {
- String op = st.sval.toLowerCase();
- if (st.sval.equals("and")) {
- match(st.TT_WORD);
- CQLNode term2 = parse_term();
+ debug("in parse_query()");
+
+ CQLNode term = parse_term(qualifier, 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);
term = new CQLAndNode(term, term2);
- } else if (st.sval.equals("or")) {
- match(st.TT_WORD);
- CQLNode term2 = parse_term();
+ } else if (lexer.ttype == lexer.TT_OR) {
+ match(lexer.TT_OR);
+ CQLNode term2 = parse_term(qualifier, relation);
term = new CQLOrNode(term, term2);
- } else if (st.sval.equals("not")) {
- match(st.TT_WORD);
- CQLNode term2 = parse_term();
+ } else if (lexer.ttype == lexer.TT_NOT) {
+ match(lexer.TT_NOT);
+ CQLNode term2 = parse_term(qualifier, relation);
term = new CQLNotNode(term, term2);
+ } else if (lexer.ttype == lexer.TT_PROX) {
+ // ### Handle "prox"
+ } else {
+ throw new CQLParseException("expected boolean, got " +
+ lexer.render());
}
}
+ debug("no more ops");
return term;
}
- private CQLNode parse_term()
+ private CQLNode parse_term(String qualifier, CQLRelation relation)
throws CQLParseException, IOException {
- if (st.ttype == '(') {
- match('(');
- CQLNode expr = parse_expression();
- match(')');
- return expr;
- }
+ debug("in parse_term()");
+
+ String word;
+ while (true) {
+ if (lexer.ttype == '(') {
+ debug("parenthesised term");
+ match('(');
+ CQLNode expr = parse_query(qualifier, relation);
+ match(')');
+ return expr;
+ } else if (lexer.ttype != lexer.TT_WORD && lexer.ttype != '"') {
+ throw new CQLParseException("expected qualifier or term, " +
+ "got " + lexer.render());
+ }
- String word = st.sval;
- return new CQLTermNode("x", "=", word);
- }
+ debug("non-parenthesised term");
+ word = lexer.sval;
+ match(lexer.ttype);
+ if (!isBaseRelation())
+ break;
+
+ qualifier = word;
+ relation = new CQLRelation(lexer.render(lexer.ttype, false));
+ match(lexer.ttype);
+
+ while (lexer.ttype == '/') {
+ match('/');
+ // ### could insist on known modifiers only
+ if (lexer.ttype != lexer.TT_WORD)
+ throw new CQLParseException("expected relation modifier, "
+ + "got " + lexer.render());
+ relation.addModifier(lexer.sval);
+ match(lexer.TT_WORD);
+ }
- private void match(int token)
- throws CQLParseException, IOException {
- if (st.ttype != token)
- throw new CQLParseException("expected " + render(st, token, null) +
- ", " + "got " + render(st));
- st.nextToken();
- }
+ debug("qualifier='" + qualifier + ", " +
+ "relation='" + relation.toCQL() + "'");
+ }
- // ### This utility should surely be a method of the StreamTokenizer class
- private static String render(StreamTokenizer st) {
- return render(st, st.ttype, null);
+ CQLTermNode node = new CQLTermNode(qualifier, relation, word);
+ debug("made term node " + node.toCQL());
+ return node;
}
- private static String render(StreamTokenizer st, int token, String str) {
- String ret;
-
- if (token == st.TT_EOF) {
- return "EOF";
- } else if (token == st.TT_EOL) {
- return "EOL";
- } else if (token == st.TT_NUMBER) {
- return "number";
- } else if (token == st.TT_WORD) {
- return "word";
- } else if (token == '"' && token == '\'') {
- return "string";
- }
+ boolean isBaseRelation() {
+ debug("isBaseRelation: checking ttype=" + lexer.ttype +
+ " (" + lexer.render() + ")");
+ return (lexer.ttype == '<' ||
+ lexer.ttype == '>' ||
+ lexer.ttype == '=' ||
+ lexer.ttype == lexer.TT_LE ||
+ lexer.ttype == lexer.TT_GE ||
+ lexer.ttype == lexer.TT_NE ||
+ lexer.ttype == lexer.TT_ANY ||
+ lexer.ttype == lexer.TT_ALL ||
+ lexer.ttype == lexer.TT_EXACT);
+ }
- return "'" + String.valueOf((char) token) + "'";
+ private void match(int token)
+ throws CQLParseException, IOException {
+ debug("in match(" + lexer.render(token, true) + ")");
+ if (lexer.ttype != token)
+ throw new CQLParseException("expected " +
+ lexer.render(token, true) +
+ ", " + "got " + lexer.render());
+ int tmp = lexer.nextToken();
+ debug("match() got token=" + lexer.ttype + ", " +
+ "nval=" + lexer.nval + ", sval='" + lexer.sval + "'" +
+ " (tmp=" + tmp + ")");
}
// e.g. echo '(au=Kerninghan or au=Ritchie) and ti=Unix' |
// java org.z3950.zing.cql.CQLParser
// yields:
- // ###
+ // <triple>
+ // <boolean>and</boolean>
+ // <triple>
+ // <boolean>or</boolean>
+ // <searchClause>
+ // <index>au<index>
+ // <relation>=<relation>
+ // <term>Kerninghan<term>
+ // </searchClause>
+ // <searchClause>
+ // <index>au<index>
+ // <relation>=<relation>
+ // <term>Ritchie<term>
+ // </searchClause>
+ // </triple>
+ // <searchClause>
+ // <index>ti<index>
+ // <relation>=<relation>
+ // <term>Unix<term>
+ // </searchClause>
+ // </triple>
//
public static void main (String[] args) {
if (args.length != 0) {
System.exit(1);
}
- byte[] bytes = new byte[1000];
+ byte[] bytes = new byte[10000];
try {
+ // Read in the whole of standard input in one go
int nbytes = System.in.read(bytes);
} catch (java.io.IOException ex) {
- System.err.println("Can't read query: " + ex);
+ System.err.println("Can't read query: " + ex.getMessage());
System.exit(2);
}
String cql = new String(bytes);
CQLNode root;
try {
root = parser.parse(cql);
+ debug("root='" + root + "'");
System.out.println(root.toXCQL(0));
- } catch (java.io.IOException ex) {
- System.err.println("Can't compile query: " + ex);
+ } catch (CQLParseException ex) {
+ System.err.println("Syntax error: " + ex.getMessage());
System.exit(3);
+ } catch (java.io.IOException ex) {
+ System.err.println("Can't compile query: " + ex.getMessage());
+ System.exit(4);
}
}
}