/**
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.solr.handler.component;
import java.io.IOException;
import java.net.URL;
import java.util.*;
import org.apache.solr.common.params.CommonParams;
import org.apache.solr.common.params.FacetParams;
import org.apache.solr.common.params.SolrParams;
import org.apache.solr.common.params.ModifiableSolrParams;
import org.apache.solr.common.util.NamedList;
import org.apache.solr.common.util.SimpleOrderedMap;
import org.apache.solr.common.util.StrUtils;
import org.apache.solr.common.SolrException;
import org.apache.solr.request.SimpleFacets;
import org.apache.lucene.util.OpenBitSet;
import org.apache.solr.search.QueryParsing;
import org.apache.solr.schema.FieldType;
import org.apache.lucene.queryParser.ParseException;
/**
* TODO!
*
* @version $Id: FacetComponent.java 1094218 2011-04-18 01:21:23Z yonik $
* @since solr 1.3
*/
public class FacetComponent extends SearchComponent
{
public static final String COMPONENT_NAME = "facet";
@Override
public void prepare(ResponseBuilder rb) throws IOException
{
if (rb.req.getParams().getBool(FacetParams.FACET,false)) {
rb.setNeedDocSet( true );
rb.doFacets = true;
}
}
/**
* Actually run the query
* @param rb
*/
@Override
public void process(ResponseBuilder rb) throws IOException
{
if (rb.doFacets) {
SolrParams params = rb.req.getParams();
SimpleFacets f = new SimpleFacets(rb.req,
rb.getResults().docSet,
params,
rb );
// TODO ???? add this directly to the response, or to the builder?
rb.rsp.add( "facet_counts", f.getFacetCounts() );
}
}
private static final String commandPrefix = "{!" + CommonParams.TERMS + "=$";
@Override
public int distributedProcess(ResponseBuilder rb) throws IOException {
if (!rb.doFacets) {
return ResponseBuilder.STAGE_DONE;
}
if (rb.stage == ResponseBuilder.STAGE_GET_FIELDS) {
// overlap facet refinement requests (those shards that we need a count for
// particular facet values from), where possible, with
// the requests to get fields (because we know that is the
// only other required phase).
// We do this in distributedProcess so we can look at all of the
// requests in the outgoing queue at once.
for (int shardNum=0; shardNum<rb.shards.length; shardNum++) {
List<String> refinements = null;
for (DistribFieldFacet dff : rb._facetInfo.facets.values()) {
if (!dff.needRefinements) continue;
List<String> refList = dff._toRefine[shardNum];
if (refList == null || refList.size()==0) continue;
String key = dff.getKey(); // reuse the same key that was used for the main facet
String termsKey = key + "__terms";
String termsVal = StrUtils.join(refList, ',');
String facetCommand;
// add terms into the original facet.field command
// do it via parameter reference to avoid another layer of encoding.
String termsKeyEncoded = QueryParsing.encodeLocalParamVal(termsKey);
if (dff.localParams != null) {
facetCommand = commandPrefix+termsKeyEncoded + " " + dff.facetStr.substring(2);
} else {
facetCommand = commandPrefix+termsKeyEncoded+'}'+dff.field;
}
if (refinements == null) {
refinements = new ArrayList<String>();
}
refinements.add(facetCommand);
refinements.add(termsKey);
refinements.add(termsVal);
}
if (refinements == null) continue;
String shard = rb.shards[shardNum];
ShardRequest refine = null;
boolean newRequest = false;
// try to find a request that is already going out to that shard.
// If nshards becomes to great, we way want to move to hashing for better
// scalability.
for (ShardRequest sreq : rb.outgoing) {
if ((sreq.purpose & ShardRequest.PURPOSE_GET_FIELDS)!=0
&& sreq.shards != null
&& sreq.shards.length==1
&& sreq.shards[0].equals(shard))
{
refine = sreq;
break;
}
}
if (refine == null) {
// we didn't find any other suitable requests going out to that shard, so
// create one ourselves.
newRequest = true;
refine = new ShardRequest();
refine.shards = new String[]{rb.shards[shardNum]};
refine.params = new ModifiableSolrParams(rb.req.getParams());
// don't request any documents
refine.params.remove(CommonParams.START);
refine.params.set(CommonParams.ROWS,"0");
}
refine.purpose |= ShardRequest.PURPOSE_REFINE_FACETS;
refine.params.set(FacetParams.FACET, "true");
refine.params.remove(FacetParams.FACET_FIELD);
refine.params.remove(FacetParams.FACET_QUERY);
for (int i=0; i<refinements.size();) {
String facetCommand=refinements.get(i++);
String termsKey=refinements.get(i++);
String termsVal=refinements.get(i++);
refine.params.add(FacetParams.FACET_FIELD, facetCommand);
refine.params.set(termsKey, termsVal);
}
if (newRequest) {
rb.addRequest(this, refine);
}
}
}
return ResponseBuilder.STAGE_DONE;
}
@Override
public void modifyRequest(ResponseBuilder rb, SearchComponent who, ShardRequest sreq) {
if (!rb.doFacets) return;
if ((sreq.purpose & ShardRequest.PURPOSE_GET_TOP_IDS) != 0) {
sreq.purpose |= ShardRequest.PURPOSE_GET_FACETS;
FacetInfo fi = rb._facetInfo;
if (fi == null) {
rb._facetInfo = fi = new FacetInfo();
fi.parse(rb.req.getParams(), rb);
// should already be true...
// sreq.params.set(FacetParams.FACET, "true");
}
sreq.params.remove(FacetParams.FACET_MINCOUNT);
sreq.params.remove(FacetParams.FACET_OFFSET);
sreq.params.remove(FacetParams.FACET_LIMIT);
for (DistribFieldFacet dff : fi.facets.values()) {
String paramStart = "f." + dff.field + '.';
sreq.params.remove(paramStart + FacetParams.FACET_MINCOUNT);
sreq.params.remove(paramStart + FacetParams.FACET_OFFSET);
dff.initialLimit = dff.limit <= 0 ? dff.limit : dff.offset + dff.limit;
if (dff.sort.equals(FacetParams.FACET_SORT_COUNT)) {
if (dff.limit > 0) {
// set the initial limit higher to increase accuracy
dff.initialLimit = (int)(dff.initialLimit * 1.5) + 10;
dff.initialMincount = 0; // TODO: we could change this to 1, but would then need more refinement for small facet result sets?
} else {
// if limit==-1, then no need to artificially lower mincount to 0 if it's 1
dff.initialMincount = Math.min(dff.minCount, 1);
}
} else {
// we're sorting by index order.
// if minCount==0, we should always be able to get accurate results w/o over-requesting or refining
// if minCount==1, we should be able to get accurate results w/o over-requesting, but we'll need to refine
// if minCount==n (>1), we can set the initialMincount to minCount/nShards, rounded up.
// For example, we know that if minCount=10 and we have 3 shards, then at least one shard must have a count of 4 for the term
// For the minCount>1 case, we can generate too short of a list (miss terms at the end of the list) unless limit==-1
// For example: each shard could produce a list of top 10, but some of those could fail to make it into the combined list (i.e.
// we needed to go beyond the top 10 to generate the top 10 combined). Overrequesting can help a little here, but not as
// much as when sorting by count.
if (dff.minCount <= 1) {
dff.initialMincount = dff.minCount;
} else {
dff.initialMincount = (int)Math.ceil((double)dff.minCount / rb.shards.length);
// dff.initialMincount = 1;
}
}
if (dff.initialMincount != 0) {
sreq.params.set(paramStart + FacetParams.FACET_MINCOUNT, dff.initialMincount);
}
// Currently this is for testing only and allows overriding of the
// facet.limit set to the shards
dff.initialLimit = rb.req.getParams().getInt("facet.shard.limit", dff.initialLimit);
sreq.params.set(paramStart + FacetParams.FACET_LIMIT, dff.initialLimit);
}
} else {
// turn off faceting on other requests
sreq.params.set(FacetParams.FACET, "false");
// we could optionally remove faceting params
}
}
@Override
public void handleResponses(ResponseBuilder rb, ShardRequest sreq) {
if (!rb.doFacets) return;
if ((sreq.purpose & ShardRequest.PURPOSE_GET_FACETS)!=0) {
countFacets(rb, sreq);
} else if ((sreq.purpose & ShardRequest.PURPOSE_REFINE_FACETS)!=0) {
refineFacets(rb, sreq);
}
}
private void countFacets(ResponseBuilder rb, ShardRequest sreq) {
FacetInfo fi = rb._facetInfo;
for (ShardResponse srsp: sreq.responses) {
int shardNum = rb.getShardNum(srsp.getShard());
NamedList facet_counts = (NamedList)srsp.getSolrResponse().getResponse().get("facet_counts");
fi.addExceptions((List)facet_counts.get("exception"));
// handle facet queries
NamedList facet_queries = (NamedList)facet_counts.get("facet_queries");
if (facet_queries != null) {
for (int i=0; i<facet_queries.size(); i++) {
String returnedKey = facet_queries.getName(i);
long count = ((Number)facet_queries.getVal(i)).longValue();
QueryFacet qf = fi.queryFacets.get(returnedKey);
qf.count += count;
}
}
// step through each facet.field, adding results from this shard
NamedList facet_fields = (NamedList)facet_counts.get("facet_fields");
if (facet_fields != null) {
for (DistribFieldFacet dff : fi.facets.values()) {
dff.add(shardNum, (NamedList)facet_fields.get(dff.getKey()), dff.initialLimit);
}
}
}
//
// This code currently assumes that there will be only a single
// request ((with responses from all shards) sent out to get facets...
// otherwise we would need to wait until all facet responses were received.
//
for (DistribFieldFacet dff : fi.facets.values()) {
// no need to check these facets for refinement
if (dff.initialLimit <= 0 && dff.initialMincount == 0) continue;
// only other case where index-sort doesn't need refinement is if minCount==0
if (dff.minCount == 0 && dff.sort.equals(FacetParams.FACET_SORT_INDEX)) continue;
@SuppressWarnings("unchecked") // generic array's are annoying
List<String>[] tmp = (List<String>[]) new List[rb.shards.length];
dff._toRefine = tmp;
ShardFacetCount[] counts = dff.getCountSorted();
int ntop = Math.min(counts.length, dff.limit >= 0 ? dff.offset + dff.limit : Integer.MAX_VALUE);
long smallestCount = counts.length == 0 ? 0 : counts[ntop-1].count;
for (int i=0; i<counts.length; i++) {
ShardFacetCount sfc = counts[i];
boolean needRefinement = false;
if (i<ntop) {
// automatically flag the top values for refinement
// this should always be true for facet.sort=index
needRefinement = true;
} else {
// this logic should only be invoked for facet.sort=index (for now)
// calculate the maximum value that this term may have
// and if it is >= smallestCount, then flag for refinement
long maxCount = sfc.count;
for (int shardNum=0; shardNum<rb.shards.length; shardNum++) {
OpenBitSet obs = dff.counted[shardNum];
if (!obs.get(sfc.termNum)) {
// if missing from this shard, add the max it could be
maxCount += dff.maxPossible(sfc,shardNum);
}
}
if (maxCount >= smallestCount) {
// TODO: on a tie, we could check the term values
needRefinement = true;
}
}
if (needRefinement) {
// add a query for each shard missing the term that needs refinement
for (int shardNum=0; shardNum<rb.shards.length; shardNum++) {
OpenBitSet obs = dff.counted[shardNum];
if (!obs.get(sfc.termNum) && dff.maxPossible(sfc,shardNum)>0) {
dff.needRefinements = true;
List<String> lst = dff._toRefine[shardNum];
if (lst == null) {
lst = dff._toRefine[shardNum] = new ArrayList<String>();
}
lst.add(sfc.name);
}
}
}
}
}
}
private void refineFacets(ResponseBuilder rb, ShardRequest sreq) {
FacetInfo fi = rb._facetInfo;
for (ShardResponse srsp: sreq.responses) {
// int shardNum = rb.getShardNum(srsp.shard);
NamedList facet_counts = (NamedList)srsp.getSolrResponse().getResponse().get("facet_counts");
NamedList facet_fields = (NamedList)facet_counts.get("facet_fields");
fi.addExceptions((List)facet_counts.get("exception"));
if (facet_fields == null) continue; // this can happen when there's an exception
for (int i=0; i<facet_fields.size(); i++) {
String key = facet_fields.getName(i);
DistribFieldFacet dff = fi.facets.get(key);
if (dff == null) continue;
NamedList shardCounts = (NamedList)facet_fields.getVal(i);
for (int j=0; j<shardCounts.size(); j++) {
String name = shardCounts.getName(j);
long count = ((Number)shardCounts.getVal(j)).longValue();
ShardFacetCount sfc = dff.counts.get(name);
sfc.count += count;
}
}
}
}
@Override
public void finishStage(ResponseBuilder rb) {
if (!rb.doFacets || rb.stage != ResponseBuilder.STAGE_GET_FIELDS) return;
// wait until STAGE_GET_FIELDS
// so that "result" is already stored in the response (for aesthetics)
FacetInfo fi = rb._facetInfo;
NamedList facet_counts = new SimpleOrderedMap();
if (fi.exceptionList != null) {
facet_counts.add("exception",fi.exceptionList);
}
NamedList facet_queries = new SimpleOrderedMap();
facet_counts.add("facet_queries",facet_queries);
for (QueryFacet qf : fi.queryFacets.values()) {
facet_queries.add(qf.getKey(), num(qf.count));
}
NamedList facet_fields = new SimpleOrderedMap();
facet_counts.add("facet_fields", facet_fields);
for (DistribFieldFacet dff : fi.facets.values()) {
NamedList fieldCounts = new NamedList(); // order is more important for facets
facet_fields.add(dff.getKey(), fieldCounts);
ShardFacetCount[] counts;
boolean countSorted = dff.sort.equals(FacetParams.FACET_SORT_COUNT);
if (countSorted) {
counts = dff.countSorted;
if (counts == null || dff.needRefinements) {
counts = dff.getCountSorted();
}
} else if (dff.sort.equals(FacetParams.FACET_SORT_INDEX)) {
counts = dff.getLexSorted();
} else { // TODO: log error or throw exception?
counts = dff.getLexSorted();
}
if (countSorted) {
int end = dff.limit < 0 ? counts.length : Math.min(dff.offset + dff.limit, counts.length);
for (int i=dff.offset; i<end; i++) {
if (counts[i].count < dff.minCount) {
break;
}
fieldCounts.add(counts[i].name, num(counts[i].count));
}
} else {
int off = dff.offset;
int lim = dff.limit >= 0 ? dff.limit : Integer.MAX_VALUE;
// index order...
for (int i=0; i<counts.length; i++) {
long count = counts[i].count;
if (count < dff.minCount) continue;
if (off > 0) {
off--;
continue;
}
if (lim <= 0) {
break;
}
lim--;
fieldCounts.add(counts[i].name, num(count));
}
}
if (dff.missing) {
fieldCounts.add(null, num(dff.missingCount));
}
}
// TODO: facet dates & numbers
facet_counts.add("facet_dates", new SimpleOrderedMap());
facet_counts.add("facet_ranges", new SimpleOrderedMap());
rb.rsp.add("facet_counts", facet_counts);
rb._facetInfo = null; // could be big, so release asap
}
// use <int> tags for smaller facet counts (better back compatibility)
private Number num(long val) {
if (val < Integer.MAX_VALUE) return (int)val;
else return val;
}
private Number num(Long val) {
if (val.longValue() < Integer.MAX_VALUE) return val.intValue();
else return val;
}
/////////////////////////////////////////////
/// SolrInfoMBean
////////////////////////////////////////////
@Override
public String getDescription() {
return "Handle Faceting";
}
@Override
public String getVersion() {
return "$Revision: 1094218 $";
}
@Override
public String getSourceId() {
return "$Id: FacetComponent.java 1094218 2011-04-18 01:21:23Z yonik $";
}
@Override
public String getSource() {
return "$URL: https://svn.apache.org/repos/asf/lucene/dev/branches/lucene_solr_3_3/solr/src/java/org/apache/solr/handler/component/FacetComponent.java $";
}
@Override
public URL[] getDocs() {
return null;
}
/**
* <b>This API is experimental and subject to change</b>
*/
public static class FacetInfo {
public LinkedHashMap<String,QueryFacet> queryFacets;
public LinkedHashMap<String,DistribFieldFacet> facets;
public List exceptionList;
void parse(SolrParams params, ResponseBuilder rb) {
queryFacets = new LinkedHashMap<String,QueryFacet>();
facets = new LinkedHashMap<String,DistribFieldFacet>();
String[] facetQs = params.getParams(FacetParams.FACET_QUERY);
if (facetQs != null) {
for (String query : facetQs) {
QueryFacet queryFacet = new QueryFacet(rb, query);
queryFacets.put(queryFacet.getKey(), queryFacet);
}
}
String[] facetFs = params.getParams(FacetParams.FACET_FIELD);
if (facetFs != null) {
for (String field : facetFs) {
DistribFieldFacet ff = new DistribFieldFacet(rb, field);
facets.put(ff.getKey(), ff);
}
}
}
public void addExceptions(List exceptions) {
if (exceptions == null) return;
if (exceptionList == null) exceptionList = new ArrayList();
exceptionList.addAll(exceptions);
}
}
/**
* <b>This API is experimental and subject to change</b>
*/
public static class FacetBase {
String facetType; // facet.field, facet.query, etc (make enum?)
String facetStr; // original parameter value of facetStr
String facetOn; // the field or query, absent localParams if appropriate
private String key; // label in the response for the result... "foo" for {!key=foo}myfield
SolrParams localParams; // any local params for the facet
public FacetBase(ResponseBuilder rb, String facetType, String facetStr) {
this.facetType = facetType;
this.facetStr = facetStr;
try {
this.localParams = QueryParsing.getLocalParams(facetStr, rb.req.getParams());
} catch (ParseException e) {
throw new SolrException(SolrException.ErrorCode.BAD_REQUEST, e);
}
this.facetOn = facetStr;
this.key = facetStr;
if (localParams != null) {
// remove local params unless it's a query
if (!facetType.equals(FacetParams.FACET_QUERY)) {
facetOn = localParams.get(CommonParams.VALUE);
key = facetOn;
}
key = localParams.get(CommonParams.OUTPUT_KEY, key);
}
}
/** returns the key in the response that this facet will be under */
public String getKey() { return key; }
public String getType() { return facetType; }
}
/**
* <b>This API is experimental and subject to change</b>
*/
public static class QueryFacet extends FacetBase {
public long count;
public QueryFacet(ResponseBuilder rb, String facetStr) {
super(rb, FacetParams.FACET_QUERY, facetStr);
}
}
/**
* <b>This API is experimental and subject to change</b>
*/
public static class FieldFacet extends FacetBase {
public String field; // the field to facet on... "myfield" for {!key=foo}myfield
public FieldType ftype;
public int offset;
public int limit;
public int minCount;
public String sort;
public boolean missing;
public String prefix;
public long missingCount;
public FieldFacet(ResponseBuilder rb, String facetStr) {
super(rb, FacetParams.FACET_FIELD, facetStr);
fillParams(rb, rb.req.getParams(), facetOn);
}
private void fillParams(ResponseBuilder rb, SolrParams params, String field) {
this.field = field;
this.ftype = rb.req.getSchema().getFieldTypeNoEx(this.field);
this.offset = params.getFieldInt(field, FacetParams.FACET_OFFSET, 0);
this.limit = params.getFieldInt(field, FacetParams.FACET_LIMIT, 100);
Integer mincount = params.getFieldInt(field, FacetParams.FACET_MINCOUNT);
if (mincount==null) {
Boolean zeros = params.getFieldBool(field, FacetParams.FACET_ZEROS);
// mincount = (zeros!=null && zeros) ? 0 : 1;
mincount = (zeros!=null && !zeros) ? 1 : 0;
// current default is to include zeros.
}
this.minCount = mincount;
this.missing = params.getFieldBool(field, FacetParams.FACET_MISSING, false);
// default to sorting by count if there is a limit.
this.sort = params.getFieldParam(field, FacetParams.FACET_SORT, limit>0 ? FacetParams.FACET_SORT_COUNT : FacetParams.FACET_SORT_INDEX);
if (this.sort.equals(FacetParams.FACET_SORT_COUNT_LEGACY)) {
this.sort = FacetParams.FACET_SORT_COUNT;
} else if (this.sort.equals(FacetParams.FACET_SORT_INDEX_LEGACY)) {
this.sort = FacetParams.FACET_SORT_INDEX;
}
this.prefix = params.getFieldParam(field,FacetParams.FACET_PREFIX);
}
}
/**
* <b>This API is experimental and subject to change</b>
*/
public static class DistribFieldFacet extends FieldFacet {
public List<String>[] _toRefine; // a List<String> of refinements needed, one for each shard.
// SchemaField sf; // currently unneeded
// the max possible count for a term appearing on no list
public long missingMaxPossible;
// the max possible count for a missing term for each shard (indexed by shardNum)
public long[] missingMax;
public OpenBitSet[] counted; // a bitset for each shard, keeping track of which terms seen
public HashMap<String,ShardFacetCount> counts = new HashMap<String,ShardFacetCount>(128);
public int termNum;
public int initialLimit; // how many terms requested in first phase
public int initialMincount; // mincount param sent to each shard
public boolean needRefinements;
public ShardFacetCount[] countSorted;
DistribFieldFacet(ResponseBuilder rb, String facetStr) {
super(rb, facetStr);
// sf = rb.req.getSchema().getField(field);
missingMax = new long[rb.shards.length];
counted = new OpenBitSet[rb.shards.length];
}
void add(int shardNum, NamedList shardCounts, int numRequested) {
// shardCounts could be null if there was an exception
int sz = shardCounts == null ? 0 : shardCounts.size();
int numReceived = sz;
OpenBitSet terms = new OpenBitSet(termNum+sz);
long last = 0;
for (int i=0; i<sz; i++) {
String name = shardCounts.getName(i);
long count = ((Number)shardCounts.getVal(i)).longValue();
if (name == null) {
missingCount += count;
numReceived--;
} else {
ShardFacetCount sfc = counts.get(name);
if (sfc == null) {
sfc = new ShardFacetCount();
sfc.name = name;
sfc.indexed = ftype == null ? sfc.name : ftype.toInternal(sfc.name);
sfc.termNum = termNum++;
counts.put(name, sfc);
}
sfc.count += count;
terms.fastSet(sfc.termNum);
last = count;
}
}
// the largest possible missing term is initialMincount if we received less
// than the number requested.
if (numRequested<0 || numRequested != 0 && numReceived < numRequested) {
last = initialMincount;
}
missingMaxPossible += last;
missingMax[shardNum] = last;
counted[shardNum] = terms;
}
public ShardFacetCount[] getLexSorted() {
ShardFacetCount[] arr = counts.values().toArray(new ShardFacetCount[counts.size()]);
Arrays.sort(arr, new Comparator<ShardFacetCount>() {
public int compare(ShardFacetCount o1, ShardFacetCount o2) {
return o1.indexed.compareTo(o2.indexed);
}
});
countSorted = arr;
return arr;
}
public ShardFacetCount[] getCountSorted() {
ShardFacetCount[] arr = counts.values().toArray(new ShardFacetCount[counts.size()]);
Arrays.sort(arr, new Comparator<ShardFacetCount>() {
public int compare(ShardFacetCount o1, ShardFacetCount o2) {
if (o2.count < o1.count) return -1;
else if (o1.count < o2.count) return 1;
return o1.indexed.compareTo(o2.indexed);
}
});
countSorted = arr;
return arr;
}
// returns the max possible value this ShardFacetCount could have for this shard
// (assumes the shard did not report a count for this value)
long maxPossible(ShardFacetCount sfc, int shardNum) {
return missingMax[shardNum];
// TODO: could store the last term in the shard to tell if this term
// comes before or after it. If it comes before, we could subtract 1
}
}
/**
* <b>This API is experimental and subject to change</b>
*/
public static class ShardFacetCount {
public String name;
public String indexed; // the indexed form of the name... used for comparisons.
public long count;
public int termNum; // term number starting at 0 (used in bit arrays)
@Override
public String toString() {
return "{term="+name+",termNum="+termNum+",count="+count+"}";
}
}
}