lucene-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From yo...@apache.org
Subject svn commit: r1685342 - in /lucene/dev/branches/branch_5x: ./ solr/ solr/core/ solr/core/src/java/org/apache/solr/search/ solr/core/src/java/org/apache/solr/search/facet/ solr/core/src/test/org/apache/solr/search/facet/
Date Sat, 13 Jun 2015 22:34:46 GMT
Author: yonik
Date: Sat Jun 13 22:34:45 2015
New Revision: 1685342

URL: http://svn.apache.org/r1685342
Log:
SOLR-7676: nested docs facet domain switching

Added:
    lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/facet/BlockJoin.java
      - copied unchanged from r1685340, lucene/dev/trunk/solr/core/src/java/org/apache/solr/search/facet/BlockJoin.java
Modified:
    lucene/dev/branches/branch_5x/   (props changed)
    lucene/dev/branches/branch_5x/solr/   (props changed)
    lucene/dev/branches/branch_5x/solr/CHANGES.txt   (contents, props changed)
    lucene/dev/branches/branch_5x/solr/core/   (props changed)
    lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/QueryContext.java
    lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/SolrIndexSearcher.java
    lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/facet/FacetQuery.java
    lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/facet/FacetRequest.java
    lucene/dev/branches/branch_5x/solr/core/src/test/org/apache/solr/search/facet/TestJsonFacets.java

Modified: lucene/dev/branches/branch_5x/solr/CHANGES.txt
URL: http://svn.apache.org/viewvc/lucene/dev/branches/branch_5x/solr/CHANGES.txt?rev=1685342&r1=1685341&r2=1685342&view=diff
==============================================================================
--- lucene/dev/branches/branch_5x/solr/CHANGES.txt (original)
+++ lucene/dev/branches/branch_5x/solr/CHANGES.txt Sat Jun 13 22:34:45 2015
@@ -56,6 +56,13 @@ New Features
 
 * SOLR-7458: Expose HDFS Block Locality Metrics via JMX (Mike Drob via Mark Miller)
 
+* SOLR-7676: Faceting on nested objects / Block-join faceting with the new JSON Facet API.
+  Example: Assuming books with nested pages and an input domain of pages, the following
+  will switch the domain to books before faceting on the author field:
+    authors:{ type:terms, field:author, domain:{toParent:"type:book"} }
+  (yonik)
+
+
 Bug Fixes
 ----------------------
 

Modified: lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/QueryContext.java
URL: http://svn.apache.org/viewvc/lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/QueryContext.java?rev=1685342&r1=1685341&r2=1685342&view=diff
==============================================================================
--- lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/QueryContext.java
(original)
+++ lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/QueryContext.java
Sat Jun 13 22:34:45 2015
@@ -26,8 +26,10 @@ import org.apache.solr.common.SolrExcept
 import org.apache.solr.core.SolrCore;
 import org.apache.solr.request.SolrRequestInfo;
 
-/**
- * Bridge between old style context and a real class
+/*
+ * Bridge between old style context and a real class.
+ * This is currently slightly more heavy weight than necessary because of the need to inherit
from IdentityHashMap rather than
+ * instantiate it on demand (and the need to put "searcher" in the map)
  * @lucene.experimental
  */
 public class QueryContext extends IdentityHashMap implements Closeable {
@@ -45,7 +47,7 @@ public class QueryContext extends Identi
   public QueryContext(IndexSearcher searcher) {
     this.searcher = searcher instanceof SolrIndexSearcher ? (SolrIndexSearcher)searcher :
null;
     indexSearcher = searcher;
-    this.put("searcher", searcher); // see ValueSource.newContext()
+    this.put("searcher", searcher); // see ValueSource.newContext()  // TODO: move check
to "get"?
   }
 
 

Modified: lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/SolrIndexSearcher.java
URL: http://svn.apache.org/viewvc/lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/SolrIndexSearcher.java?rev=1685342&r1=1685341&r2=1685342&view=diff
==============================================================================
--- lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/SolrIndexSearcher.java
(original)
+++ lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/SolrIndexSearcher.java
Sat Jun 13 22:34:45 2015
@@ -872,6 +872,25 @@ public class SolrIndexSearcher extends I
     getDocSet(query);
   }
 
+  public BitDocSet getDocSetBits(Query q) throws IOException {
+    DocSet answer = getDocSet(q);
+    if (answer instanceof BitDocSet) {
+      return (BitDocSet)answer;
+    }
+
+    FixedBitSet bs = new FixedBitSet(maxDoc());
+    DocIterator iter = answer.iterator();
+    while (iter.hasNext()) {
+      bs.set(iter.nextDoc());
+    }
+
+    BitDocSet answerBits = new BitDocSet(bs , answer.size());
+    if (filterCache != null) {
+      filterCache.put(q, answerBits);
+    }
+    return answerBits;
+  }
+
   /**
    * Returns the set of document ids matching a query.
    * This method is cache-aware and attempts to retrieve the answer from the cache if possible.

Modified: lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/facet/FacetQuery.java
URL: http://svn.apache.org/viewvc/lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/facet/FacetQuery.java?rev=1685342&r1=1685341&r2=1685342&view=diff
==============================================================================
--- lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/facet/FacetQuery.java
(original)
+++ lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/facet/FacetQuery.java
Sat Jun 13 22:34:45 2015
@@ -59,3 +59,4 @@ class FacetQueryProcessor extends FacetP
 
 
 }
+

Modified: lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/facet/FacetRequest.java
URL: http://svn.apache.org/viewvc/lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/facet/FacetRequest.java?rev=1685342&r1=1685341&r2=1685342&view=diff
==============================================================================
--- lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/facet/FacetRequest.java
(original)
+++ lucene/dev/branches/branch_5x/solr/core/src/java/org/apache/solr/search/facet/FacetRequest.java
Sat Jun 13 22:34:45 2015
@@ -31,6 +31,7 @@ import java.util.Map;
 import org.apache.lucene.index.LeafReaderContext;
 import org.apache.lucene.search.BooleanClause;
 import org.apache.lucene.search.BooleanQuery;
+import org.apache.lucene.search.MatchAllDocsQuery;
 import org.apache.lucene.search.Query;
 import org.apache.solr.common.SolrException;
 import org.apache.solr.common.params.FacetParams;
@@ -41,6 +42,7 @@ import org.apache.solr.request.SolrQuery
 import org.apache.solr.request.SolrRequestInfo;
 import org.apache.solr.schema.IndexSchema;
 import org.apache.solr.schema.SchemaField;
+import org.apache.solr.search.BitDocSet;
 import org.apache.solr.search.DocIterator;
 import org.apache.solr.search.DocSet;
 import org.apache.solr.search.FunctionQParser;
@@ -54,8 +56,17 @@ import org.apache.solr.search.SyntaxErro
 public abstract class FacetRequest {
   protected Map<String,AggValueSource> facetStats;  // per-bucket statistics
   protected Map<String,FacetRequest> subFacets;     // list of facets
-  protected List<String> excludeTags;
+  protected List<String> filters;
   protected boolean processEmpty;
+  protected Domain domain;
+
+  // domain changes
+  public static class Domain {
+    public List<String> excludeTags;
+    public boolean toParent;
+    public boolean toChildren;
+    public String parents;
+  }
 
   public FacetRequest() {
     facetStats = new LinkedHashMap<>();
@@ -140,7 +151,42 @@ class FacetProcessor<FacetRequestT exten
   }
 
   protected void handleDomainChanges() throws IOException {
-    if (freq.excludeTags == null || freq.excludeTags.size() == 0) {
+    if (freq.domain == null) return;
+    handleFilterExclusions();
+    handleBlockJoin();
+  }
+
+  private void handleBlockJoin() throws IOException {
+    if (!(freq.domain.toChildren || freq.domain.toParent)) return;
+
+    // TODO: avoid query parsing per-bucket somehow...
+    String parentStr = freq.domain.parents;
+    Query parentQuery;
+    try {
+      QParser parser = QParser.getParser(parentStr, null, fcontext.req);
+      parentQuery = parser.getQuery();
+    } catch (SyntaxError err) {
+      throw new SolrException(SolrException.ErrorCode.BAD_REQUEST, "Error parsing block join
parent specification: " + parentStr);
+    }
+
+    BitDocSet parents = fcontext.searcher.getDocSetBits(parentQuery);
+    DocSet input = fcontext.base;
+    DocSet result;
+
+    if (freq.domain.toChildren) {
+      DocSet filt = fcontext.searcher.getDocSetBits( new MatchAllDocsQuery() );
+      result = BlockJoin.toChildren(input, parents, filt, fcontext.qcontext);
+    } else {
+      result = BlockJoin.toParents(input, parents, fcontext.qcontext);
+    }
+
+    fcontext.base = result;
+  }
+
+  private void handleFilterExclusions() throws IOException {
+    List<String> excludeTags = freq.domain.excludeTags;
+
+    if (excludeTags == null || excludeTags.size() == 0) {
       return;
     }
 
@@ -153,7 +199,7 @@ class FacetProcessor<FacetRequestT exten
     }
 
     IdentityHashMap<Query,Boolean> excludeSet = new IdentityHashMap<>();
-    for (String excludeTag : freq.excludeTags) {
+    for (String excludeTag : excludeTags) {
       Object olst = tagMap.get(excludeTag);
       // tagMap has entries of List<String,List<QParser>>, but subject to change
in the future
       if (!(olst instanceof Collection)) continue;
@@ -487,7 +533,7 @@ abstract class FacetParser<FacetRequestT
     } else if ("query".equals(type)) {
       return parseQueryFacet(key, args);
     } else if ("range".equals(type)) {
-     return parseRangeFacet(key, args);
+      return parseRangeFacet(key, args);
     }
 
     return parseStat(key, type, args);
@@ -528,10 +574,41 @@ abstract class FacetParser<FacetRequestT
   }
 
 
+  private FacetRequest.Domain getDomain() {
+    if (facet.domain == null) {
+      facet.domain = new FacetRequest.Domain();
+    }
+    return facet.domain;
+  }
+
   protected void parseCommonParams(Object o) {
     if (o instanceof Map) {
       Map<String,Object> m = (Map<String,Object>)o;
-      facet.excludeTags = getStringList(m, "excludeTags");
+      List<String> excludeTags = getStringList(m, "excludeTags");
+      if (excludeTags != null) {
+        getDomain().excludeTags = excludeTags;
+      }
+
+      Map<String,Object> domainMap = (Map<String,Object>) m.get("domain");
+      if (domainMap != null) {
+        excludeTags = getStringList(m, "excludeTags");
+        if (excludeTags != null) {
+          getDomain().excludeTags = excludeTags;
+        }
+
+        String blockParent = (String)domainMap.get("blockParent");
+        String blockChildren = (String)domainMap.get("blockChildren");
+
+        if (blockParent != null) {
+          getDomain().toParent = true;
+          getDomain().parents = blockParent;
+        } else if (blockChildren != null) {
+          getDomain().toChildren = true;
+          getDomain().parents = blockChildren;
+        }
+
+      }
+
     }
   }
 
@@ -696,6 +773,34 @@ class FacetQueryParser extends FacetPars
   }
 }
 
+/*** not a separate type of parser for now...
+class FacetBlockParentParser extends FacetParser<FacetBlockParent> {
+  public FacetBlockParentParser(FacetParser parent, String key) {
+    super(parent, key);
+    facet = new FacetBlockParent();
+  }
+
+  @Override
+  public FacetBlockParent parse(Object arg) throws SyntaxError {
+    parseCommonParams(arg);
+
+    if (arg instanceof String) {
+      // just the field name...
+      facet.parents = (String)arg;
+
+    } else if (arg instanceof Map) {
+      Map<String, Object> m = (Map<String, Object>) arg;
+      facet.parents = getString(m, "parents", null);
+
+      parseSubs( m.get("facet") );
+    }
+
+    return facet;
+  }
+}
+***/
+
+
 class FacetFieldParser extends FacetParser<FacetField> {
   public FacetFieldParser(FacetParser parent, String key) {
     super(parent, key);

Modified: lucene/dev/branches/branch_5x/solr/core/src/test/org/apache/solr/search/facet/TestJsonFacets.java
URL: http://svn.apache.org/viewvc/lucene/dev/branches/branch_5x/solr/core/src/test/org/apache/solr/search/facet/TestJsonFacets.java?rev=1685342&r1=1685341&r2=1685342&view=diff
==============================================================================
--- lucene/dev/branches/branch_5x/solr/core/src/test/org/apache/solr/search/facet/TestJsonFacets.java
(original)
+++ lucene/dev/branches/branch_5x/solr/core/src/test/org/apache/solr/search/facet/TestJsonFacets.java
Sat Jun 13 22:34:45 2015
@@ -1072,6 +1072,82 @@ public class TestJsonFacets extends Solr
     );
   }
 
+
+
+  @Test
+  public void testBlockJoin() throws Exception {
+    doBlockJoin(Client.localClient());
+  }
+
+  public void doBlockJoin(Client client) throws Exception {
+    ModifiableSolrParams p = params("rows","0");
+
+    client.deleteByQuery("*:*", null);
+
+    SolrInputDocument parent;
+    parent = sdoc("id", "1", "type_s","book", "book_s","A", "v_t","q");
+    client.add(parent, null);
+
+    parent = sdoc("id", "2", "type_s","book", "book_s","B", "v_t","q w");
+    parent.addChildDocument( sdoc("id","2.1", "type_s","page", "page_s","a", "v_t","x y z")
 );
+    parent.addChildDocument( sdoc("id","2.2", "type_s","page", "page_s","b", "v_t","x y 
") );
+    parent.addChildDocument( sdoc("id","2.3", "type_s","page", "page_s","c", "v_t","  y z"
)  );
+    client.add(parent, null);
+
+    parent = sdoc("id", "3", "type_s","book", "book_s","C", "v_t","q w e");
+    parent.addChildDocument( sdoc("id","3.1", "type_s","page", "page_s","d", "v_t","x   
")  );
+    parent.addChildDocument( sdoc("id","3.2", "type_s","page", "page_s","e", "v_t","  y 
")  );
+    parent.addChildDocument( sdoc("id","3.3", "type_s","page", "page_s","f", "v_t","    z")
 );
+    client.add(parent, null);
+
+    parent = sdoc("id", "4", "type_s","book", "book_s","D", "v_t","e");
+    client.add(parent, null);
+
+    client.commit();
+
+    client.testJQ(params(p, "q", "*:*"
+            , "json.facet", "{ " +
+                "pages:{ type:query, domain:{blockChildren:'type_s:book'} , facet:{ x:{field:v_t}
} }" +
+                ",pages2:{type:terms, field:v_t, domain:{blockChildren:'type_s:book'} }"
+
+                ",books:{ type:query, domain:{blockParent:'type_s:book'}  , facet:{ x:{field:v_t}
} }" +
+                ",books2:{type:terms, field:v_t, domain:{blockParent:'type_s:book'} }" +
+                ",pageof3:{ type:query, q:'id:3', facet : { x : { type:terms, field:page_s,
domain:{blockChildren:'type_s:book'}}} }" +
+                ",bookof22:{ type:query, q:'id:2.2', facet : { x : { type:terms, field:book_s,
domain:{blockParent:'type_s:book'}}} }" +
+                ",missing_blockParent:{ type:query, domain:{blockParent:'type_s:does_not_exist'}
}" +
+                ",missing_blockChildren:{ type:query, domain:{blockChildren:'type_s:does_not_exist'}
}" +
+                "}"
+        )
+        , "facets=={ count:10" +
+            ", pages:{count:6 , x:{buckets:[ {val:y,count:4},{val:x,count:3},{val:z,count:3}
]}  }" +
+            ", pages2:{ buckets:[ {val:y,count:4},{val:x,count:3},{val:z,count:3} ] }" +
+            ", books:{count:4 , x:{buckets:[ {val:q,count:3},{val:e,count:2},{val:w,count:2}
]}  }" +
+            ", books2:{ buckets:[ {val:q,count:3},{val:e,count:2},{val:w,count:2} ] }" +
+            ", pageof3:{count:1 , x:{buckets:[ {val:d,count:1},{val:e,count:1},{val:f,count:1}
]}  }" +
+            ", bookof22:{count:1 , x:{buckets:[ {val:B,count:1} ]}  }" +
+            ", missing_blockParent:{count:0}" +
+            ", missing_blockChildren:{count:0}" +
+            "}"
+    );
+
+    // no matches in base query
+    client.testJQ(params("q", "no_match_s:NO_MATCHES"
+            , "json.facet", "{ processEmpty:true," +
+                "pages:{ type:query, domain:{blockChildren:'type_s:book'} }" +
+                ",books:{ type:query, domain:{blockParent:'type_s:book'} }" +
+                "}"
+        )
+        , "facets=={ count:0" +
+            ", pages:{count:0}" +
+            ", books:{count:0}" +
+            "}"
+    );
+
+
+  }
+
+
+
+
   public void XtestPercentiles() {
     AVLTreeDigest catA = new AVLTreeDigest(100);
     catA.add(4);



Mime
View raw message