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

Example 71 with SolrIndexSearcher

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

the class UnInvertedField method visitTerm.

@Override
protected void visitTerm(TermsEnum te, int termNum) throws IOException {
    if (termNum >= maxTermCounts.length) {
        // resize by doubling - for very large number of unique terms, expanding
        // by 4K and resultant GC will dominate uninvert times.  Resize at end if material
        int[] newMaxTermCounts = new int[Math.min(Integer.MAX_VALUE - 16, maxTermCounts.length * 2)];
        System.arraycopy(maxTermCounts, 0, newMaxTermCounts, 0, termNum);
        maxTermCounts = newMaxTermCounts;
    }
    final BytesRef term = te.term();
    if (te.docFreq() > maxTermDocFreq) {
        // this makes a deep copy of the term bytes
        Term t = new Term(field, term);
        TopTerm topTerm = new TopTerm();
        topTerm.term = t.bytes();
        topTerm.termNum = termNum;
        topTerm.termQuery = new TermQuery(t);
        bigTerms.put(topTerm.termNum, topTerm);
        if (deState == null) {
            deState = new SolrIndexSearcher.DocsEnumState();
            deState.fieldName = field;
            deState.liveDocs = searcher.getSlowAtomicReader().getLiveDocs();
            // TODO: check for MultiTermsEnum in SolrIndexSearcher could now fail?
            deState.termsEnum = te;
            deState.postingsEnum = postingsEnum;
            deState.minSetSizeCached = maxTermDocFreq;
        }
        postingsEnum = deState.postingsEnum;
        DocSet set = searcher.getDocSet(deState);
        maxTermCounts[termNum] = set.size();
    }
}

Example 72 with SolrIndexSearcher

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

the class SortSlotAcc method collect.

public int collect(DocSet docs, int slot) throws IOException {
    int count = 0;
    SolrIndexSearcher searcher = fcontext.searcher;
    final List<LeafReaderContext> leaves = searcher.getIndexReader().leaves();
    final Iterator<LeafReaderContext> ctxIt = leaves.iterator();
    LeafReaderContext ctx = null;
    int segBase = 0;
    int segMax;
    int adjustedMax = 0;
    for (DocIterator docsIt = docs.iterator(); docsIt.hasNext(); ) {
        final int doc = docsIt.nextDoc();
        if (doc >= adjustedMax) {
            do {
                ctx = ctxIt.next();
                if (ctx == null) {
                    // should be impossible
                    throw new RuntimeException("INTERNAL FACET ERROR");
                }
                segBase = ctx.docBase;
                segMax = ctx.reader().maxDoc();
                adjustedMax = segBase + segMax;
            } while (doc >= adjustedMax);
            assert doc >= ctx.docBase;
            setNextReader(ctx);
        }
        count++;
        // per-seg collectors
        collect(doc - segBase, slot);
    }
    return count;
}

Example 73 with SolrIndexSearcher

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

the class PivotFacetProcessor method processSingle.

/**
   * Process a single branch of refinement values for a specific pivot
   * @param pivotFields the ordered list of fields in this pivot
   * @param refinements the comma separate list of refinement values corresponding to each field in the pivot, or null if there are no refinements
   * @param statsFields List of {@link StatsField} instances to compute for each pivot value
   * @param facetQueries the list of facet queries hung under this pivot
   * @param facetRanges the list of facet ranges hung under this pivot
   */
private SimpleOrderedMap<List<NamedList<Object>>> processSingle(List<String> pivotFields, String refinements, List<StatsField> statsFields, final ParsedParams parsed, List<FacetComponent.FacetBase> facetQueries, List<RangeFacetRequest> facetRanges) throws IOException {
    SolrIndexSearcher searcher = rb.req.getSearcher();
    SimpleOrderedMap<List<NamedList<Object>>> pivotResponse = new SimpleOrderedMap<>();
    String field = pivotFields.get(0);
    SchemaField sfield = searcher.getSchema().getField(field);
    Deque<String> fnames = new LinkedList<>();
    for (int i = pivotFields.size() - 1; i > 1; i--) {
        fnames.push(pivotFields.get(i));
    }
    NamedList<Integer> facetCounts;
    Deque<String> vnames = new LinkedList<>();
    if (null != refinements) {
        // All values, split by the field they should go to
        List<String> refinementValuesByField = PivotFacetHelper.decodeRefinementValuePath(refinements);
        for (int i = refinementValuesByField.size() - 1; i > 0; i--) {
            //Only for [1] and on
            vnames.push(refinementValuesByField.get(i));
        }
        String firstFieldsValues = refinementValuesByField.get(0);
        facetCounts = new NamedList<>();
        facetCounts.add(firstFieldsValues, getSubsetSize(parsed.docs, sfield, firstFieldsValues));
    } else {
        // no refinements needed
        facetCounts = this.getTermCountsForPivots(field, parsed);
    }
    if (pivotFields.size() > 1) {
        String subField = pivotFields.get(1);
        pivotResponse.add(parsed.key, doPivots(facetCounts, field, subField, fnames, vnames, parsed, statsFields, facetQueries, facetRanges));
    } else {
        pivotResponse.add(parsed.key, doPivots(facetCounts, field, null, fnames, vnames, parsed, statsFields, facetQueries, facetRanges));
    }
    return pivotResponse;
}

Example 74 with SolrIndexSearcher

use of org.apache.solr.search.SolrIndexSearcher 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);
}

Example 75 with SolrIndexSearcher

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

the class PivotFacetProcessor method process.

/**
   * Processes all of the specified {@link FacetParams#FACET_PIVOT} strings, generating 
   * a complete response tree for each pivot.  The values in this response will either
   * be the complete tree of fields and values for the specified pivot in the local index, 
   * or the requested refinements if the pivot params include the {@link PivotFacet#REFINE_PARAM}
   */
public SimpleOrderedMap<List<NamedList<Object>>> process(String[] pivots) throws IOException {
    if (!rb.doFacets || pivots == null)
        return null;
    // rb._statsInfo may be null if stats=false, ie: refine requests
    // if that's the case, but we need to refine w/stats, then we'll lazy init our 
    // own instance of StatsInfo
    StatsInfo statsInfo = rb._statsInfo;
    SimpleOrderedMap<List<NamedList<Object>>> pivotResponse = new SimpleOrderedMap<>();
    for (String pivotList : pivots) {
        final ParsedParams parsed;
        try {
            parsed = this.parseParams(FacetParams.FACET_PIVOT, pivotList);
        } catch (SyntaxError e) {
            throw new SolrException(ErrorCode.BAD_REQUEST, e);
        }
        List<String> pivotFields = StrUtils.splitSmart(parsed.facetValue, ",", true);
        if (pivotFields.size() < 1) {
            throw new SolrException(ErrorCode.BAD_REQUEST, "Pivot Facet needs at least one field name: " + pivotList);
        } else {
            SolrIndexSearcher searcher = rb.req.getSearcher();
            for (String fieldName : pivotFields) {
                SchemaField sfield = searcher.getSchema().getField(fieldName);
                if (sfield == null) {
                    throw new SolrException(ErrorCode.BAD_REQUEST, "\"" + fieldName + "\" is not a valid field name in pivot: " + pivotList);
                }
            }
        }
        // start by assuming no local params...
        // no local => no refinement
        String refineKey = null;
        // no local => no stats
        List<StatsField> statsFields = Collections.emptyList();
        List<FacetComponent.FacetBase> facetQueries = Collections.emptyList();
        List<RangeFacetRequest> facetRanges = Collections.emptyList();
        if (null != parsed.localParams) {
            // we might be refining..
            refineKey = parsed.localParams.get(PivotFacet.REFINE_PARAM);
            String statsLocalParam = parsed.localParams.get(StatsParams.STATS);
            if (null != refineKey && null != statsLocalParam && null == statsInfo) {
                // we are refining and need to compute stats, 
                // but stats component hasn't inited StatsInfo (because we
                // don't need/want top level stats when refining) so we lazy init
                // our own copy of StatsInfo
                statsInfo = new StatsInfo(rb);
            }
            statsFields = getTaggedStatsFields(statsInfo, statsLocalParam);
            try {
                FacetComponent.FacetContext facetContext = FacetComponent.FacetContext.getFacetContext(req);
                String taggedQueries = parsed.localParams.get(QUERY);
                if (StringUtils.isEmpty(taggedQueries)) {
                    facetQueries = Collections.emptyList();
                } else {
                    List<String> localParamValue = StrUtils.splitSmart(taggedQueries, ',');
                    if (localParamValue.size() > 1) {
                        String msg = QUERY + " local param of " + FacetParams.FACET_PIVOT + "may not include tags separated by a comma - please use a common tag on all " + FacetParams.FACET_QUERY + " params you wish to compute under this pivot";
                        throw new SolrException(ErrorCode.BAD_REQUEST, msg);
                    }
                    taggedQueries = localParamValue.get(0);
                    facetQueries = facetContext.getQueryFacetsForTag(taggedQueries);
                }
                String taggedRanges = parsed.localParams.get(RANGE);
                if (StringUtils.isEmpty(taggedRanges)) {
                    facetRanges = Collections.emptyList();
                } else {
                    List<String> localParamValue = StrUtils.splitSmart(taggedRanges, ',');
                    if (localParamValue.size() > 1) {
                        String msg = RANGE + " local param of " + FacetParams.FACET_PIVOT + "may not include tags separated by a comma - please use a common tag on all " + FacetParams.FACET_RANGE + " params you wish to compute under this pivot";
                        throw new SolrException(ErrorCode.BAD_REQUEST, msg);
                    }
                    taggedRanges = localParamValue.get(0);
                    facetRanges = facetContext.getRangeFacetRequestsForTag(taggedRanges);
                }
            } catch (IllegalStateException e) {
                throw new SolrException(ErrorCode.SERVER_ERROR, "Faceting context not set, cannot calculate pivot values");
            }
        }
        if (null != refineKey) {
            String[] refinementValuesByField = params.getParams(PivotFacet.REFINE_PARAM + refineKey);
            for (String refinements : refinementValuesByField) {
                pivotResponse.addAll(processSingle(pivotFields, refinements, statsFields, parsed, facetQueries, facetRanges));
            }
        } else {
            pivotResponse.addAll(processSingle(pivotFields, null, statsFields, parsed, facetQueries, facetRanges));
        }
    }
    return pivotResponse;
}