From 63baa5cc4709694b15be165b597a2a3200b425c0 Mon Sep 17 00:00:00 2001
From: Patrick Schmitz <pschmitz@berkeley.edu>
Date: Sat, 3 Dec 2011 00:29:30 +0000
Subject: [PATCH] CSPACE-4302 Add support for booleans (AND, OR, NOT) and
 phrase search (latter waiting for update from nuxeo).

---
 .../query/nuxeo/QueryManagerNuxeoImpl.java    | 79 ++++++++++++++++++-
 1 file changed, 77 insertions(+), 2 deletions(-)

diff --git a/services/common/src/main/java/org/collectionspace/services/common/query/nuxeo/QueryManagerNuxeoImpl.java b/services/common/src/main/java/org/collectionspace/services/common/query/nuxeo/QueryManagerNuxeoImpl.java
index 242f715b3..151f745fb 100644
--- a/services/common/src/main/java/org/collectionspace/services/common/query/nuxeo/QueryManagerNuxeoImpl.java
+++ b/services/common/src/main/java/org/collectionspace/services/common/query/nuxeo/QueryManagerNuxeoImpl.java
@@ -29,6 +29,7 @@ package org.collectionspace.services.common.query.nuxeo;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
+import java.util.regex.Matcher;
 import java.util.regex.Pattern;
 
 import org.nuxeo.ecm.core.api.DocumentModel;
@@ -61,10 +62,10 @@ public class QueryManagerNuxeoImpl implements IQueryManager {
 	// words
 	private static Pattern nonWordChars = Pattern
 			.compile("[^\\p{L}\\p{M}\\p{N}_']");
+	private static Pattern kwdTokenizer = Pattern.compile("(?:(['\"])(.*?)(?<!\\\\)(?>\\\\\\\\)*\\1|([^ ]+))");
 	private static Pattern unescapedDblQuotes = Pattern.compile("(?<!\\\\)\"");
 	private static Pattern unescapedSingleQuote = Pattern.compile("(?<!\\\\)'");
-	private static Pattern kwdSearchProblemChars = Pattern
-			.compile("[\\:\\(\\)]");
+	private static Pattern kwdSearchProblemChars = Pattern.compile("[\\:\\(\\)]");
 	private static Pattern kwdSearchHyphen = Pattern.compile(" - ");
 
 	private static String getLikeForm() {
@@ -155,6 +156,80 @@ public class QueryManagerNuxeoImpl implements IQueryManager {
 	// in NXQL
 	@Override
 	public String createWhereClauseFromKeywords(String keywords) {
+		String result = null;
+		StringBuffer fullTextWhereClause = new StringBuffer(SEARCH_GROUP_OPEN);
+		// StringBuffer phraseWhereClause = new StringBuffer(SEARCH_GROUP_OPEN);
+		// Split on unescaped double quotes to handle phrases
+		Matcher regexMatcher = kwdTokenizer.matcher(keywords.trim());
+		boolean addNOT = false;
+		boolean newWordSet = true;
+		while (regexMatcher.find()) {
+			String phrase = regexMatcher.group();
+			// Not needed - already trimmed by split: 
+			// String trimmed = phrase.trim();
+			// Ignore empty strings from match, or goofy input
+			if (phrase.isEmpty())
+				continue;
+			// Note we let OR through as is
+			if("AND".equalsIgnoreCase(phrase)) {
+				continue;	// AND is default
+			} else if("NOT".equalsIgnoreCase(phrase)) {
+				addNOT = true;
+				continue;
+			}
+			if (newWordSet) {
+				fullTextWhereClause.append(ECM_FULLTEXT_LIKE + "'");
+				newWordSet = false;
+			} else {
+				fullTextWhereClause.append(SEARCH_TERM_SEPARATOR);
+			}
+			if(addNOT) {
+				fullTextWhereClause.append("-");	// Negate the next term
+				addNOT = false;
+			}
+			// Next comment block of questionable value...
+			
+			// ignore the special chars except single quote here - can't hurt
+			// TODO this should become a special function that strips things the
+			// fulltext will ignore, including non-word chars and too-short
+			// words,
+			// and escaping single quotes. Can return a boolean for anything
+			// stripped,
+			// which triggers the back-up search. We can think about whether
+			// stripping
+			// short words not in a quoted phrase should trigger the backup.
+			phrase = unescapedSingleQuote.matcher(phrase).replaceAll("\\\\'");
+			// If there are non-word chars in the phrase, we need to match the
+			// phrase exactly against the fulltext table for this object
+			// if(nonWordChars.matcher(trimmed).matches()) {
+			// }
+			// Replace problem chars with spaces. Patches CSPACE-4147,
+			// CSPACE-4106
+			phrase = kwdSearchProblemChars.matcher(phrase).replaceAll(" ");
+			phrase = kwdSearchHyphen.matcher(phrase).replaceAll(" ");
+
+			fullTextWhereClause.append(phrase);
+			
+			if (logger.isTraceEnabled() == true) {
+				logger.trace("Current built whereClause is: "
+						+ fullTextWhereClause.toString());
+			}
+		}
+		if (fullTextWhereClause.length()==0) {
+			throw new RuntimeException(
+					"No usable keywords specified in string:[" + keywords + "]");
+		}
+		fullTextWhereClause.append("'" + SEARCH_GROUP_CLOSE);
+
+		result = fullTextWhereClause.toString();
+		if (logger.isDebugEnabled()) {
+			logger.debug("Final built WHERE clause is: " + result);
+		}
+
+		return result;
+	}
+
+	public String createWhereClauseFromKeywordsOld(String keywords) {
 		String result = null;
 		StringBuffer fullTextWhereClause = new StringBuffer(SEARCH_GROUP_OPEN);
 		// StringBuffer phraseWhereClause = new StringBuffer(SEARCH_GROUP_OPEN);
-- 
2.47.3