X-Git-Url: http://jsfdemo.indexdata.com/?a=blobdiff_plain;f=src%2Forg%2Fz3950%2Fzing%2Fcql%2FCQLParser.java;h=9ccb1398a89764fe857836332a6370a9654f2dd2;hb=0cf2302a18eacaa327f40743c4ac59cf6c630183;hp=13e505e9bb0f34f863a034005c2bb7f2bdda7958;hpb=c0cf61216e9a8d036a7aee17f1b24136a6630e65;p=cql-java-moved-to-github.git diff --git a/src/org/z3950/zing/cql/CQLParser.java b/src/org/z3950/zing/cql/CQLParser.java index 13e505e..9ccb139 100644 --- a/src/org/z3950/zing/cql/CQLParser.java +++ b/src/org/z3950/zing/cql/CQLParser.java @@ -1,4 +1,4 @@ -// $Id: CQLParser.java,v 1.23 2003-09-04 21:56:46 mike Exp $ +// $Id: CQLParser.java,v 1.38 2007-07-03 16:40:41 mike Exp $ package org.z3950.zing.cql; import java.io.IOException; @@ -12,15 +12,44 @@ import java.io.FileNotFoundException; /** * Compiles CQL strings into parse trees of CQLNode subtypes. * - * @version $Id: CQLParser.java,v 1.23 2003-09-04 21:56:46 mike Exp $ + * @version $Id: CQLParser.java,v 1.38 2007-07-03 16:40:41 mike Exp $ * @see http://zing.z3950.org/cql/index.html */ public class CQLParser { private CQLLexer lexer; + private int compat; // When false, implement CQL 1.2 + public static int V1POINT1 = 12368; + public static int V1POINT2 = 12369; + public static int V1POINT1SORT = 12370; + static private boolean DEBUG = false; static private boolean LEXDEBUG = false; + /** + * The new parser implements a dialect of CQL specified by the + * compat argument: + * + */ + CQLParser(int compat) { + this.compat = compat; + } + + /** + * The new parser implements CQL 1.2 + */ + CQLParser() { + this.compat = V1POINT2; + } + private static void debug(String str) { if (DEBUG) System.err.println("PARSEDEBUG: " + str); @@ -48,39 +77,66 @@ public class CQLParser { lexer.nextToken(); debug("about to parseQuery()"); - CQLNode root = parseQuery("srw.serverChoice", new CQLRelation("scr")); + CQLNode root = parseTopLevelPrefixes("cql.serverChoice", + new CQLRelation(compat == V1POINT2 ? "=" : "scr")); if (lexer.ttype != lexer.TT_EOF) throw new CQLParseException("junk after end: " + lexer.render()); return root; } - private CQLNode parseQuery(String qualifier, CQLRelation relation) + private CQLNode parseTopLevelPrefixes(String index, CQLRelation relation) + throws CQLParseException, IOException { + debug("top-level prefix mapping"); + + if (lexer.ttype == '>') { + return parsePrefix(index, relation, true); + } + + CQLNode node = parseQuery(index, relation); + if ((compat == V1POINT2 || compat == V1POINT1SORT) && + lexer.ttype == lexer.TT_SORTBY) { + match(lexer.ttype); + debug("sortspec"); + + CQLSortNode sortnode = new CQLSortNode(node); + while (lexer.ttype != lexer.TT_EOF) { + String sortindex = matchSymbol("sort index"); + ModifierSet ms = gatherModifiers(sortindex); + sortnode.addSortIndex(ms); + } + + if (sortnode.keys.size() == 0) { + throw new CQLParseException("no sort keys"); + } + + node = sortnode; + } + + return node; + } + + private CQLNode parseQuery(String index, CQLRelation relation) throws CQLParseException, IOException { debug("in parseQuery()"); - CQLNode term = parseTerm(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 = parseTerm(qualifier, relation); - term = new CQLAndNode(term, term2); - } else if (lexer.ttype == lexer.TT_OR) { - match(lexer.TT_OR); - CQLNode term2 = parseTerm(qualifier, relation); - term = new CQLOrNode(term, term2); - } else if (lexer.ttype == lexer.TT_NOT) { - match(lexer.TT_NOT); - CQLNode term2 = parseTerm(qualifier, 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 = parseTerm(qualifier, relation); - proxnode.addSecondSubterm(term2); - term = (CQLNode) proxnode; + lexer.ttype != ')' && + lexer.ttype != lexer.TT_SORTBY) { + if (lexer.ttype == lexer.TT_AND || + lexer.ttype == lexer.TT_OR || + lexer.ttype == lexer.TT_NOT || + lexer.ttype == lexer.TT_PROX) { + int type = lexer.ttype; + String val = lexer.sval; + match(type); + ModifierSet ms = gatherModifiers(val); + CQLNode term2 = parseTerm(index, relation); + term = ((type == lexer.TT_AND) ? new CQLAndNode(term, term2, ms) : + (type == lexer.TT_OR) ? new CQLOrNode (term, term2, ms) : + (type == lexer.TT_NOT) ? new CQLNotNode(term, term2, ms) : + new CQLProxNode(term, term2, ms)); } else { throw new CQLParseException("expected boolean, got " + lexer.render()); @@ -91,7 +147,34 @@ public class CQLParser { return term; } - private CQLNode parseTerm(String qualifier, CQLRelation relation) + private ModifierSet gatherModifiers(String base) + throws CQLParseException, IOException { + debug("in gatherModifiers()"); + + ModifierSet ms = new ModifierSet(base); + while (lexer.ttype == '/') { + match('/'); + if (lexer.ttype != lexer.TT_WORD) + throw new CQLParseException("expected modifier, " + + "got " + lexer.render()); + String type = lexer.sval.toLowerCase(); + match(lexer.ttype); + if (!isRelation()) { + // It's a simple modifier consisting of type only + ms.addModifier(type); + } else { + // It's a complex modifier of the form type=value + String comparision = lexer.render(lexer.ttype, false); + match(lexer.ttype); + String value = matchSymbol("modifier value"); + ms.addModifier(type, comparision, value); + } + } + + return ms; + } + + private CQLNode parseTerm(String index, CQLRelation relation) throws CQLParseException, IOException { debug("in parseTerm()"); @@ -100,57 +183,40 @@ public class CQLParser { if (lexer.ttype == '(') { debug("parenthesised term"); match('('); - CQLNode expr = parseQuery(qualifier, relation); + CQLNode expr = parseQuery(index, relation); match(')'); return expr; } else if (lexer.ttype == '>') { - match('>'); - return parsePrefix(qualifier, relation); + return parsePrefix(index, relation, false); } debug("non-parenthesised term"); - word = matchSymbol("qualifier or term"); - if (!isBaseRelation()) + word = matchSymbol("index or term"); + if (!isRelation() && lexer.ttype != lexer.TT_WORD) break; - qualifier = word; - relation = new CQLRelation(lexer.ttype == lexer.TT_WORD ? - lexer.sval : - lexer.render(lexer.ttype, false)); + index = word; + String relstr = (lexer.ttype == lexer.TT_WORD ? + lexer.sval : lexer.render(lexer.ttype, false)); + relation = new CQLRelation(relstr); match(lexer.ttype); - - while (lexer.ttype == '/') { - match('/'); - if (lexer.ttype != lexer.TT_RELEVANT && - lexer.ttype != lexer.TT_FUZZY && - lexer.ttype != lexer.TT_STEM && - lexer.ttype != lexer.TT_PHONETIC && - lexer.ttype != lexer.TT_WORD) - throw new CQLParseException("expected relation modifier, " - + "got " + lexer.render()); - if (lexer.ttype == lexer.TT_WORD && - lexer.sval.indexOf('.') == -1) - throw new CQLParseException("unknown first-class " + - "relation modifier: " + - lexer.sval); - - relation.addModifier(lexer.sval.toLowerCase()); - match(lexer.ttype); - } - - debug("qualifier='" + qualifier + ", " + + ModifierSet ms = gatherModifiers(relstr); + relation.setModifiers(ms); + 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 qualifier, CQLRelation relation) + private CQLNode parsePrefix(String index, CQLRelation relation, + boolean topLevel) throws CQLParseException, IOException { debug("prefix mapping"); + match('>'); String name = null; String identifier = matchSymbol("prefix-name"); if (lexer.ttype == '=') { @@ -158,101 +224,24 @@ public class CQLParser { name = identifier; identifier = matchSymbol("prefix-identifer"); } - CQLNode term = parseQuery(qualifier, relation); - return new CQLPrefixNode(name, identifier, term); - } - - private void gatherProxParameters(CQLProxNode node) - throws CQLParseException, IOException { - for (int i = 0; i < 4; i++) { - if (lexer.ttype != '/') - return; // end of proximity parameters + CQLNode node = topLevel ? + parseTopLevelPrefixes(index, relation) : + parseQuery(index, relation); - match('/'); - if (lexer.ttype != '/') { - // not an omitted default - switch (i) { - // Order should be: relation/distance/unit/ordering - // For now, use MA's: unit/relation/distance/ordering - case 0: gatherProxRelation(node); break; - case 1: gatherProxDistance(node); break; - case 2: gatherProxUnit(node); break; - case 3: gatherProxOrdering(node); break; - } - } - } - } - - private void gatherProxRelation(CQLProxNode node) - throws CQLParseException, IOException { - if (!isProxRelation()) - throw new CQLParseException("expected proximity relation, got " + - lexer.render()); - node.addModifier("relation", lexer.render(lexer.ttype, false)); - match(lexer.ttype); - debug("gPR matched " + lexer.render(lexer.ttype, false)); - } - - private void gatherProxDistance(CQLProxNode node) - throws CQLParseException, IOException { - if (lexer.ttype != lexer.TT_NUMBER) - throw new CQLParseException("expected proximity distance, got " + - lexer.render()); - node.addModifier("distance", lexer.render(lexer.ttype, false)); - match(lexer.ttype); - debug("gPD matched " + lexer.render(lexer.ttype, false)); + return new CQLPrefixNode(name, identifier, node); } - private void gatherProxUnit(CQLProxNode node) - throws CQLParseException, IOException { - if (lexer.ttype != lexer.TT_pWORD && - lexer.ttype != lexer.TT_SENTENCE && - lexer.ttype != lexer.TT_PARAGRAPH && - lexer.ttype != lexer.TT_ELEMENT) - throw new CQLParseException("expected proximity unit, got " + - lexer.render()); - node.addModifier("unit", lexer.render()); - match(lexer.ttype); - } - - private void gatherProxOrdering(CQLProxNode node) - throws CQLParseException, IOException { - if (lexer.ttype != lexer.TT_ORDERED && - lexer.ttype != lexer.TT_UNORDERED) - throw new CQLParseException("expected proximity ordering, got " + - lexer.render()); - node.addModifier("ordering", lexer.render()); - match(lexer.ttype); - } - - private boolean isBaseRelation() - throws CQLParseException { - debug("isBaseRelation: checking ttype=" + lexer.ttype + - " (" + lexer.render() + ")"); - - if (lexer.ttype == lexer.TT_WORD && - lexer.sval.indexOf('.') == -1) - throw new CQLParseException("unknown first-class relation: " + - lexer.sval); - - return (isProxRelation() || - lexer.ttype == lexer.TT_ANY || - lexer.ttype == lexer.TT_ALL || - lexer.ttype == lexer.TT_EXACT || - lexer.ttype == lexer.TT_SCR || - lexer.ttype == lexer.TT_WORD); - } - - // Checks for a relation that may be used inside a prox operator - private boolean isProxRelation() { - debug("isProxRelation: checking ttype=" + lexer.ttype + + // Checks for a relation + private boolean isRelation() { + debug("isRelation: 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_NE || + lexer.ttype == lexer.TT_EQEQ); } private void match(int token) @@ -277,27 +266,14 @@ public class CQLParser { lexer.ttype == '"' || // The following is a complete list of keywords. Because // they're listed here, they can be used unquoted as - // qualifiers, terms, prefix names and prefix identifiers. + // 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_ANY || - lexer.ttype == lexer.TT_ALL || - lexer.ttype == lexer.TT_EXACT || - 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 || - lexer.ttype == lexer.TT_RELEVANT || - lexer.ttype == lexer.TT_FUZZY || - lexer.ttype == lexer.TT_STEM || - lexer.ttype == lexer.TT_SCR || - lexer.ttype == lexer.TT_PHONETIC) { + lexer.ttype == lexer.TT_SORTBY) { String symbol = (lexer.ttype == lexer.TT_NUMBER) ? lexer.render() : lexer.sval; match(lexer.ttype); @@ -356,6 +332,10 @@ public class CQLParser { * </triple> * *

+ * @param -1 + * CQL version 1.1 (default version 1.2) + * @param -d + * Debug mode: extra output written to stderr. * @param -c * Causes the output to be written in CQL rather than XCQL - that * is, a query equivalent to that which was input, is output. In @@ -368,11 +348,22 @@ public class CQLParser { char mode = 'x'; // x=XCQL, c=CQL, p=PQF String pfile = null; - Vector argv = new Vector(); + Vector argv = new Vector(); for (int i = 0; i < args.length; i++) { argv.add(args[i]); } + int compat = V1POINT2; + if (argv.size() > 0 && argv.get(0).equals("-1")) { + compat = V1POINT1; + argv.remove(0); + } + + 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); @@ -384,8 +375,8 @@ public class CQLParser { } if (argv.size() > 1) { - System.err.println( - "Usage: CQLParser [-c] [-p []"); + System.err.println("Usage: CQLParser [-1] [-d] [-c] " + + "[-p []"); System.err.println("If unspecified, query is read from stdin"); System.exit(1); } @@ -405,7 +396,7 @@ public class CQLParser { cql = new String(bytes); } - CQLParser parser = new CQLParser(); + CQLParser parser = new CQLParser(compat); CQLNode root = null; try { root = parser.parse(cql); @@ -435,8 +426,8 @@ public class CQLParser { } catch (IOException ex) { System.err.println("Can't render query: " + ex.getMessage()); System.exit(5); - } catch (UnknownQualifierException ex) { - System.err.println("Unknown qualifier: " + ex.getMessage()); + } catch (UnknownIndexException ex) { + System.err.println("Unknown index: " + ex.getMessage()); System.exit(6); } catch (UnknownRelationException ex) { System.err.println("Unknown relation: " + ex.getMessage());