Examples with Grouping org.apache.solr.search.Grouping used on opensource projects

Example 1 with Grouping

use of org.apache.solr.search.Grouping in project lucene-solr by apache.

the class QueryComponent method prepareGrouping.

protected void prepareGrouping(ResponseBuilder rb) throws IOException {
    SolrQueryRequest req = rb.req;
    SolrParams params = req.getParams();
    if (null != rb.getCursorMark()) {
        // grouping with a cursor - so for now we just don't allow the combination at all
        throw new SolrException(SolrException.ErrorCode.BAD_REQUEST, "Can not use Grouping with " + CursorMarkParams.CURSOR_MARK_PARAM);
    }
    SolrIndexSearcher searcher = rb.req.getSearcher();
    GroupingSpecification groupingSpec = new GroupingSpecification();
    rb.setGroupingSpec(groupingSpec);
    final SortSpec sortSpec = rb.getSortSpec();
    //TODO: move weighting of sort
    final SortSpec groupSortSpec = searcher.weightSortSpec(sortSpec, Sort.RELEVANCE);
    String withinGroupSortStr = params.get(GroupParams.GROUP_SORT);
    //TODO: move weighting of sort
    final SortSpec withinGroupSortSpec;
    if (withinGroupSortStr != null) {
        SortSpec parsedWithinGroupSortSpec = SortSpecParsing.parseSortSpec(withinGroupSortStr, req);
        withinGroupSortSpec = searcher.weightSortSpec(parsedWithinGroupSortSpec, Sort.RELEVANCE);
    } else {
        withinGroupSortSpec = new SortSpec(groupSortSpec.getSort(), groupSortSpec.getSchemaFields(), groupSortSpec.getCount(), groupSortSpec.getOffset());
    }
    withinGroupSortSpec.setOffset(params.getInt(GroupParams.GROUP_OFFSET, 0));
    withinGroupSortSpec.setCount(params.getInt(GroupParams.GROUP_LIMIT, 1));
    groupingSpec.setWithinGroupSortSpec(withinGroupSortSpec);
    groupingSpec.setGroupSortSpec(groupSortSpec);
    String formatStr = params.get(GroupParams.GROUP_FORMAT, Grouping.Format.grouped.name());
    Grouping.Format responseFormat;
    try {
        responseFormat = Grouping.Format.valueOf(formatStr);
    } catch (IllegalArgumentException e) {
        throw new SolrException(SolrException.ErrorCode.BAD_REQUEST, String.format(Locale.ROOT, "Illegal %s parameter", GroupParams.GROUP_FORMAT));
    }
    groupingSpec.setResponseFormat(responseFormat);
    groupingSpec.setFields(params.getParams(GroupParams.GROUP_FIELD));
    groupingSpec.setQueries(params.getParams(GroupParams.GROUP_QUERY));
    groupingSpec.setFunctions(params.getParams(GroupParams.GROUP_FUNC));
    groupingSpec.setIncludeGroupCount(params.getBool(GroupParams.GROUP_TOTAL_COUNT, false));
    groupingSpec.setMain(params.getBool(GroupParams.GROUP_MAIN, false));
    groupingSpec.setNeedScore((rb.getFieldFlags() & SolrIndexSearcher.GET_SCORES) != 0);
    groupingSpec.setTruncateGroups(params.getBool(GroupParams.GROUP_TRUNCATE, false));
}

Example 2 with Grouping

use of org.apache.solr.search.Grouping in project lucene-solr by apache.

the class SimpleFacets method computeDocSet.

protected DocSet computeDocSet(DocSet baseDocSet, List<String> excludeTagList) throws SyntaxError, IOException {
    Map<?, ?> tagMap = (Map<?, ?>) req.getContext().get("tags");
    // rb can be null if facets are being calculated from a RequestHandler e.g. MoreLikeThisHandler
    if (tagMap == null || rb == null) {
        return baseDocSet;
    }
    IdentityHashMap<Query, Boolean> excludeSet = new IdentityHashMap<>();
    for (String excludeTag : excludeTagList) {
        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;
        for (Object o : (Collection<?>) olst) {
            if (!(o instanceof QParser))
                continue;
            QParser qp = (QParser) o;
            excludeSet.put(qp.getQuery(), Boolean.TRUE);
        }
    }
    if (excludeSet.size() == 0)
        return baseDocSet;
    List<Query> qlist = new ArrayList<>();
    // add the base query
    if (!excludeSet.containsKey(rb.getQuery())) {
        qlist.add(rb.getQuery());
    }
    // add the filters
    if (rb.getFilters() != null) {
        for (Query q : rb.getFilters()) {
            if (!excludeSet.containsKey(q)) {
                qlist.add(q);
            }
        }
    }
    // get the new base docset for this facet
    DocSet base = searcher.getDocSet(qlist);
    if (rb.grouping() && rb.getGroupingSpec().isTruncateGroups()) {
        Grouping grouping = new Grouping(searcher, null, rb.getQueryCommand(), false, 0, false);
        grouping.setWithinGroupSort(rb.getGroupingSpec().getSortWithinGroup());
        if (rb.getGroupingSpec().getFields().length > 0) {
            grouping.addFieldCommand(rb.getGroupingSpec().getFields()[0], req);
        } else if (rb.getGroupingSpec().getFunctions().length > 0) {
            grouping.addFunctionCommand(rb.getGroupingSpec().getFunctions()[0], req);
        } else {
            return base;
        }
        AllGroupHeadsCollector allGroupHeadsCollector = grouping.getCommands().get(0).createAllGroupCollector();
        searcher.search(base.getTopFilter(), allGroupHeadsCollector);
        return new BitDocSet(allGroupHeadsCollector.retrieveGroupHeads(searcher.maxDoc()));
    } else {
        return base;
    }
}

Example 3 with Grouping

use of org.apache.solr.search.Grouping in project lucene-solr by apache.

the class QueryComponent method process.

/**
   * Actually run the query
   */
@Override
public void process(ResponseBuilder rb) throws IOException {
    LOG.debug("process: {}", rb.req.getParams());
    SolrQueryRequest req = rb.req;
    SolrParams params = req.getParams();
    if (!params.getBool(COMPONENT_NAME, true)) {
        return;
    }
    SolrIndexSearcher searcher = req.getSearcher();
    StatsCache statsCache = req.getCore().getStatsCache();
    int purpose = params.getInt(ShardParams.SHARDS_PURPOSE, ShardRequest.PURPOSE_GET_TOP_IDS);
    if ((purpose & ShardRequest.PURPOSE_GET_TERM_STATS) != 0) {
        statsCache.returnLocalStats(rb, searcher);
        return;
    }
    // check if we need to update the local copy of global dfs
    if ((purpose & ShardRequest.PURPOSE_SET_TERM_STATS) != 0) {
        // retrieve from request and update local cache
        statsCache.receiveGlobalStats(req);
    }
    SolrQueryResponse rsp = rb.rsp;
    IndexSchema schema = searcher.getSchema();
    // Optional: This could also be implemented by the top-level searcher sending
    // a filter that lists the ids... that would be transparent to
    // the request handler, but would be more expensive (and would preserve score
    // too if desired).
    String ids = params.get(ShardParams.IDS);
    if (ids != null) {
        SchemaField idField = schema.getUniqueKeyField();
        List<String> idArr = StrUtils.splitSmart(ids, ",", true);
        int[] luceneIds = new int[idArr.size()];
        int docs = 0;
        if (idField.getType().isPointField()) {
            for (int i = 0; i < idArr.size(); i++) {
                int id = searcher.search(idField.getType().getFieldQuery(null, idField, idArr.get(i)), 1).scoreDocs[0].doc;
                if (id >= 0) {
                    luceneIds[docs++] = id;
                }
            }
        } else {
            for (int i = 0; i < idArr.size(); i++) {
                int id = searcher.getFirstMatch(new Term(idField.getName(), idField.getType().toInternal(idArr.get(i))));
                if (id >= 0)
                    luceneIds[docs++] = id;
            }
        }
        DocListAndSet res = new DocListAndSet();
        res.docList = new DocSlice(0, docs, luceneIds, null, docs, 0);
        if (rb.isNeedDocSet()) {
            // TODO: create a cache for this!
            List<Query> queries = new ArrayList<>();
            queries.add(rb.getQuery());
            List<Query> filters = rb.getFilters();
            if (filters != null)
                queries.addAll(filters);
            res.docSet = searcher.getDocSet(queries);
        }
        rb.setResults(res);
        ResultContext ctx = new BasicResultContext(rb);
        rsp.addResponse(ctx);
        return;
    }
    // -1 as flag if not set.
    long timeAllowed = params.getLong(CommonParams.TIME_ALLOWED, -1L);
    if (null != rb.getCursorMark() && 0 < timeAllowed) {
        // fundamentally incompatible
        throw new SolrException(SolrException.ErrorCode.BAD_REQUEST, "Can not search using both " + CursorMarkParams.CURSOR_MARK_PARAM + " and " + CommonParams.TIME_ALLOWED);
    }
    QueryCommand cmd = rb.getQueryCommand();
    cmd.setTimeAllowed(timeAllowed);
    req.getContext().put(SolrIndexSearcher.STATS_SOURCE, statsCache.get(req));
    QueryResult result = new QueryResult();
    cmd.setSegmentTerminateEarly(params.getBool(CommonParams.SEGMENT_TERMINATE_EARLY, CommonParams.SEGMENT_TERMINATE_EARLY_DEFAULT));
    if (cmd.getSegmentTerminateEarly()) {
        result.setSegmentTerminatedEarly(Boolean.FALSE);
    }
    //
    // grouping / field collapsing
    //
    GroupingSpecification groupingSpec = rb.getGroupingSpec();
    if (groupingSpec != null) {
        // not supported, silently ignore any segmentTerminateEarly flag
        cmd.setSegmentTerminateEarly(false);
        try {
            boolean needScores = (cmd.getFlags() & SolrIndexSearcher.GET_SCORES) != 0;
            if (params.getBool(GroupParams.GROUP_DISTRIBUTED_FIRST, false)) {
                CommandHandler.Builder topsGroupsActionBuilder = new CommandHandler.Builder().setQueryCommand(cmd).setNeedDocSet(// Order matters here
                false).setIncludeHitCount(true).setSearcher(searcher);
                for (String field : groupingSpec.getFields()) {
                    topsGroupsActionBuilder.addCommandField(new SearchGroupsFieldCommand.Builder().setField(schema.getField(field)).setGroupSort(groupingSpec.getGroupSort()).setTopNGroups(cmd.getOffset() + cmd.getLen()).setIncludeGroupCount(groupingSpec.isIncludeGroupCount()).build());
                }
                CommandHandler commandHandler = topsGroupsActionBuilder.build();
                commandHandler.execute();
                SearchGroupsResultTransformer serializer = new SearchGroupsResultTransformer(searcher);
                rsp.add("firstPhase", commandHandler.processResult(result, serializer));
                rsp.add("totalHitCount", commandHandler.getTotalHitCount());
                rb.setResult(result);
                return;
            } else if (params.getBool(GroupParams.GROUP_DISTRIBUTED_SECOND, false)) {
                CommandHandler.Builder secondPhaseBuilder = new CommandHandler.Builder().setQueryCommand(cmd).setTruncateGroups(groupingSpec.isTruncateGroups() && groupingSpec.getFields().length > 0).setSearcher(searcher);
                int docsToCollect = Grouping.getMax(groupingSpec.getWithinGroupOffset(), groupingSpec.getWithinGroupLimit(), searcher.maxDoc());
                docsToCollect = Math.max(docsToCollect, 1);
                for (String field : groupingSpec.getFields()) {
                    SchemaField schemaField = schema.getField(field);
                    String[] topGroupsParam = params.getParams(GroupParams.GROUP_DISTRIBUTED_TOPGROUPS_PREFIX + field);
                    if (topGroupsParam == null) {
                        topGroupsParam = new String[0];
                    }
                    List<SearchGroup<BytesRef>> topGroups = new ArrayList<>(topGroupsParam.length);
                    for (String topGroup : topGroupsParam) {
                        SearchGroup<BytesRef> searchGroup = new SearchGroup<>();
                        if (!topGroup.equals(TopGroupsShardRequestFactory.GROUP_NULL_VALUE)) {
                            BytesRefBuilder builder = new BytesRefBuilder();
                            schemaField.getType().readableToIndexed(topGroup, builder);
                            searchGroup.groupValue = builder.get();
                        }
                        topGroups.add(searchGroup);
                    }
                    secondPhaseBuilder.addCommandField(new TopGroupsFieldCommand.Builder().setField(schemaField).setGroupSort(groupingSpec.getGroupSort()).setSortWithinGroup(groupingSpec.getSortWithinGroup()).setFirstPhaseGroups(topGroups).setMaxDocPerGroup(docsToCollect).setNeedScores(needScores).setNeedMaxScore(needScores).build());
                }
                for (String query : groupingSpec.getQueries()) {
                    secondPhaseBuilder.addCommandField(new Builder().setDocsToCollect(docsToCollect).setSort(groupingSpec.getGroupSort()).setQuery(query, rb.req).setDocSet(searcher).build());
                }
                CommandHandler commandHandler = secondPhaseBuilder.build();
                commandHandler.execute();
                TopGroupsResultTransformer serializer = new TopGroupsResultTransformer(rb);
                rsp.add("secondPhase", commandHandler.processResult(result, serializer));
                rb.setResult(result);
                return;
            }
            int maxDocsPercentageToCache = params.getInt(GroupParams.GROUP_CACHE_PERCENTAGE, 0);
            boolean cacheSecondPassSearch = maxDocsPercentageToCache >= 1 && maxDocsPercentageToCache <= 100;
            Grouping.TotalCount defaultTotalCount = groupingSpec.isIncludeGroupCount() ? Grouping.TotalCount.grouped : Grouping.TotalCount.ungrouped;
            // this is normally from "rows"
            int limitDefault = cmd.getLen();
            Grouping grouping = new Grouping(searcher, result, cmd, cacheSecondPassSearch, maxDocsPercentageToCache, groupingSpec.isMain());
            grouping.setGroupSort(groupingSpec.getGroupSort()).setWithinGroupSort(groupingSpec.getSortWithinGroup()).setDefaultFormat(groupingSpec.getResponseFormat()).setLimitDefault(limitDefault).setDefaultTotalCount(defaultTotalCount).setDocsPerGroupDefault(groupingSpec.getWithinGroupLimit()).setGroupOffsetDefault(groupingSpec.getWithinGroupOffset()).setGetGroupedDocSet(groupingSpec.isTruncateGroups());
            if (groupingSpec.getFields() != null) {
                for (String field : groupingSpec.getFields()) {
                    grouping.addFieldCommand(field, rb.req);
                }
            }
            if (groupingSpec.getFunctions() != null) {
                for (String groupByStr : groupingSpec.getFunctions()) {
                    grouping.addFunctionCommand(groupByStr, rb.req);
                }
            }
            if (groupingSpec.getQueries() != null) {
                for (String groupByStr : groupingSpec.getQueries()) {
                    grouping.addQueryCommand(groupByStr, rb.req);
                }
            }
            if (rb.isNeedDocList() || rb.isDebug()) {
                // we need a single list of the returned docs
                cmd.setFlags(SolrIndexSearcher.GET_DOCLIST);
            }
            grouping.execute();
            if (grouping.isSignalCacheWarning()) {
                rsp.add("cacheWarning", String.format(Locale.ROOT, "Cache limit of %d percent relative to maxdoc has exceeded. Please increase cache size or disable caching.", maxDocsPercentageToCache));
            }
            rb.setResult(result);
            if (grouping.mainResult != null) {
                ResultContext ctx = new BasicResultContext(rb, grouping.mainResult);
                rsp.addResponse(ctx);
                rsp.getToLog().add("hits", grouping.mainResult.matches());
            } else if (!grouping.getCommands().isEmpty()) {
                // Can never be empty since grouping.execute() checks for this.
                rsp.add("grouped", result.groupedResults);
                rsp.getToLog().add("hits", grouping.getCommands().get(0).getMatches());
            }
            return;
        } catch (SyntaxError e) {
            throw new SolrException(SolrException.ErrorCode.BAD_REQUEST, e);
        }
    }
    // normal search result
    searcher.search(result, cmd);
    rb.setResult(result);
    ResultContext ctx = new BasicResultContext(rb);
    rsp.addResponse(ctx);
    rsp.getToLog().add("hits", rb.getResults().docList.matches());
    if (!rb.req.getParams().getBool(ShardParams.IS_SHARD, false)) {
        if (null != rb.getNextCursorMark()) {
            rb.rsp.add(CursorMarkParams.CURSOR_MARK_NEXT, rb.getNextCursorMark().getSerializedTotem());
        }
    }
    if (rb.mergeFieldHandler != null) {
        rb.mergeFieldHandler.handleMergeFields(rb, searcher);
    } else {
        doFieldSortValues(rb, searcher);
    }
    doPrefetch(rb);
}