Index: branch-0.7/src/java/org/apache/nutch/searcher/LuceneQueryOptimizer.java =================================================================== --- branch-0.7/src/java/org/apache/nutch/searcher/LuceneQueryOptimizer.java (revision 354477) +++ branch-0.7/src/java/org/apache/nutch/searcher/LuceneQueryOptimizer.java (working copy) @@ -54,7 +54,7 @@ public TopDocs optimize(BooleanQuery original, Searcher searcher, int numHits, - String sortField, boolean reverse) + SortField[] sortFields) throws IOException { BooleanQuery query = new BooleanQuery(); @@ -124,11 +124,11 @@ } } - if (sortField == null && !reverse) { + if (sortFields == null) { return searcher.search(query, filter, numHits); } else { return searcher.search(query, filter, numHits, - new Sort(sortField, reverse)); + new Sort(sortFields)); } } } Index: branch-0.7/src/java/org/apache/nutch/searcher/NutchBean.java =================================================================== --- branch-0.7/src/java/org/apache/nutch/searcher/NutchBean.java (revision 354477) +++ branch-0.7/src/java/org/apache/nutch/searcher/NutchBean.java (working copy) @@ -27,6 +27,8 @@ import org.apache.nutch.parse.*; import org.apache.nutch.indexer.*; +import org.apache.lucene.search.SortField; + /** * One stop shopping for search-related functionality. * @version $Id: NutchBean.java,v 1.19 2005/02/07 19:10:08 cutting Exp $ @@ -142,6 +144,12 @@ return searcher.search(query, numHits, dedupField, sortField, reverse); } + public Hits search(Query query, int numHits, + String dedupField, SortField[] sortFields) + throws IOException { + return searcher.search(query, numHits, dedupField, sortFields); + } + private class DupHits extends ArrayList { private boolean maxSizeExceeded; } Index: branch-0.7/src/java/org/apache/nutch/searcher/DistributedSearch.java =================================================================== --- branch-0.7/src/java/org/apache/nutch/searcher/DistributedSearch.java (revision 354477) +++ branch-0.7/src/java/org/apache/nutch/searcher/DistributedSearch.java (working copy) @@ -28,6 +28,8 @@ import org.apache.nutch.io.*; import org.apache.nutch.ipc.RPC; +import org.apache.lucene.search.SortField; + /** Implements the search API over IPC connnections. */ public class DistributedSearch { public static final Logger LOG = @@ -180,7 +182,14 @@ segmentToAddress.keySet().toArray(new String[segmentToAddress.size()]); } - public Hits search(final Query query, final int numHits, + public Hits search(Query query, int numHits, + String dedupField, SortField[] sortFields) + throws IOException { + //oy, punting for now + return null; + } + + public Hits search(final Query query, final int numHits, final String dedupField, final String sortField, final boolean reverse) throws IOException { long totalHits = 0; Index: branch-0.7/src/java/org/apache/nutch/searcher/IndexSearcher.java =================================================================== --- branch-0.7/src/java/org/apache/nutch/searcher/IndexSearcher.java (revision 354477) +++ branch-0.7/src/java/org/apache/nutch/searcher/IndexSearcher.java (working copy) @@ -30,6 +30,7 @@ import org.apache.lucene.search.ScoreDoc; import org.apache.lucene.search.FieldDoc; import org.apache.lucene.search.FieldCache; +import org.apache.lucene.search.SortField; import org.apache.lucene.document.Document; import org.apache.lucene.document.Field; @@ -76,17 +77,27 @@ } public Hits search(Query query, int numHits, - String dedupField, String sortField, boolean reverse) - + String dedupField, SortField[] sortFields) throws IOException { org.apache.lucene.search.BooleanQuery luceneQuery = QueryFilters.filter(query); - return translateHits - (optimizer.optimize(luceneQuery, luceneSearcher, numHits, - sortField, reverse), - dedupField, sortField); + TopDocs topDocs = optimizer.optimize(luceneQuery, luceneSearcher, + numHits, sortFields); + + return translateHits(topDocs, dedupField, sortFields); + } + + public Hits search(Query query, int numHits, + String dedupField, String sortField, boolean reverse) + throws IOException { + + SortField tmp; + tmp = (sortField == null) ? SortField.FIELD_SCORE : + new SortField(sortField, reverse); + SortField[] sortFields = {tmp}; + return search(query, numHits, dedupField, sortFields); } public String getExplanation(Query query, Hit hit) throws IOException { @@ -118,10 +129,63 @@ return results; } + private WritableComparable getSortValue(FieldDoc doc, int fieldNum) { + WritableComparable sortValue; + + Object raw = doc.fields[fieldNum]; + if (raw instanceof Integer) { + sortValue = new IntWritable(((Integer)raw).intValue()); + } else if (raw instanceof Float) { + sortValue = new FloatWritable(((Float)raw).floatValue()); + } else if (raw instanceof String) { + sortValue = new UTF8((String)raw); + } else { + throw new RuntimeException("Unknown sort value type!"); + } + return sortValue; + } + + private WritableComparable getSortValue(FieldDoc doc, + SortField[] sortFields) { + WritableComparable sortValue; + if (sortFields.length == 1) { + if (sortFields[0].getType() == SortField.SCORE) + sortValue = new FloatWritable(doc.score); + else + sortValue = getSortValue(doc, 0); + } + else { + SortValueWritable sortValueWriteable; + sortValueWriteable = new SortValueWritable(sortFields.length); + for (int i = 0; i < sortFields.length; i++) { + if (sortFields[i].getType() == SortField.SCORE) + sortValueWriteable.put(new FloatWritable(doc.score), i); + else if (sortFields[i].getType() == SortField.DOC) + sortValueWriteable.put(new IntWritable(doc.doc), i); + else + sortValueWriteable.put(getSortValue(doc, i), i); + } + sortValue = sortValueWriteable; + } + return sortValue; + } + private Hits translateHits(TopDocs topDocs, String dedupField, String sortField) throws IOException { + SortField[] sortFields = new SortField[1]; + if (sortField == null) + sortFields[0] = SortField.FIELD_SCORE; + else + sortFields[0] = new SortField(sortField); + return translateHits(topDocs, dedupField, sortFields); + } + + private Hits translateHits(TopDocs topDocs, + String dedupField, SortField[] sortFields) + throws IOException { + String[] dedupValues = null; if (dedupField != null) dedupValues = FieldCache.DEFAULT.getStrings(reader, dedupField); @@ -134,20 +198,7 @@ int doc = scoreDocs[i].doc; WritableComparable sortValue; // convert value to writable - if (sortField == null) { - sortValue = new FloatWritable(scoreDocs[i].score); - } else { - Object raw = ((FieldDoc)scoreDocs[i]).fields[0]; - if (raw instanceof Integer) { - sortValue = new IntWritable(((Integer)raw).intValue()); - } else if (raw instanceof Float) { - sortValue = new FloatWritable(((Float)raw).floatValue()); - } else if (raw instanceof String) { - sortValue = new UTF8((String)raw); - } else { - throw new RuntimeException("Unknown sort value type!"); - } - } + sortValue = getSortValue((FieldDoc)scoreDocs[i], sortFields); String dedupValue = dedupValues == null ? null : dedupValues[doc]; Index: branch-0.7/src/java/org/apache/nutch/searcher/Searcher.java =================================================================== --- branch-0.7/src/java/org/apache/nutch/searcher/Searcher.java (revision 354477) +++ branch-0.7/src/java/org/apache/nutch/searcher/Searcher.java (working copy) @@ -16,6 +16,8 @@ package org.apache.nutch.searcher; +import org.apache.lucene.search.SortField; + import java.io.IOException; /** Service that searches. */ @@ -25,6 +27,11 @@ String dedupField, String sortField, boolean reverse) throws IOException; + /** Return the top-scoring hits for a query. */ + Hits search(Query query, int numHits, + String dedupField, SortField[] sortFields) + throws IOException; + /** Return an HTML-formatted explanation of how a query scored. */ String getExplanation(Query query, Hit hit) throws IOException; }