FilteredQuery包含两个成员变量:
FilterQuery所得到的结果集同两者取AND查询相同,只不过打分的时候,FilterQuery只考虑query的部分,不考虑filter的部分。
Filter包含很多种如下:
其包含一个成员变量Set<Term> terms=new TreeSet<Term>(),所有包含terms集合中任一term的文档全部属于文档号集合。
其getDocIdSet函数如下:
public DocIdSet getDocIdSet(IndexReader reader) throws IOException { //生成一个bitset,大小为索引中文档总数 OpenBitSet result=new OpenBitSet(reader.maxDoc()); TermDocs td = reader.termDocs(); try { //遍历每个term的文档列表,将文档号都在bitset中置一,从而bitset包含了所有的文档号。 for (Iterator<Term> iter = terms.iterator(); iter.hasNext();) { Term term = iter.next(); td.seek(term); while (td.next()) { result.set(td.doc()); } } } finally { td.close(); } return result; } |
其像BooleanQuery相似,包含should的filter,must的filter,not的filter,在getDocIdSet的时候,先将所有满足should的文档号集合之间取OR的关系,然后同not的文档号集合取NOT的关系,最后同must的文档号集合取AND的关系,得到最后的文档集合。
其getDocIdSet函数如下:
public DocIdSet getDocIdSet(IndexReader reader) throws IOException { OpenBitSetDISI res = null; if (shouldFilters != null) { for (int i = 0; i < shouldFilters.size(); i++) { if (res == null) { res = new OpenBitSetDISI(getDISI(shouldFilters, i, reader), reader.maxDoc()); } else { //将should的filter的文档号全部取OR至bitset中 DocIdSet dis = shouldFilters.get(i).getDocIdSet(reader); if(dis instanceof OpenBitSet) { res.or((OpenBitSet) dis); } else { res.inPlaceOr(getDISI(shouldFilters, i, reader)); } } } } if (notFilters!=null) { for (int i = 0; i < notFilters.size(); i++) { if (res == null) { res = new OpenBitSetDISI(getDISI(notFilters, i, reader), reader.maxDoc()); res.flip(0, reader.maxDoc()); } else { //将not的filter的文档号全部取NOT至bitset中 DocIdSet dis = notFilters.get(i).getDocIdSet(reader); if(dis instanceof OpenBitSet) { res.andNot((OpenBitSet) dis); } else { res.inPlaceNot(getDISI(notFilters, i, reader)); } } } } if (mustFilters!=null) { for (int i = 0; i < mustFilters.size(); i++) { if (res == null) { res = new OpenBitSetDISI(getDISI(mustFilters, i, reader), reader.maxDoc()); } else { //将must的filter的文档号全部取AND至bitset中 DocIdSet dis = mustFilters.get(i).getDocIdSet(reader); if(dis instanceof OpenBitSet) { res.and((OpenBitSet) dis); } else { res.inPlaceAnd(getDISI(mustFilters, i, reader)); } } } } if (res !=null) return finalResult(res, reader.maxDoc()); return DocIdSet.EMPTY_DOCIDSET; } |
DuplicateFilter实现了如下的功能:
比如说我们有这样一批文档,每篇文档都分成多页,每篇文档都有一个id,然而每一页是按照单独的Document进行索引的,于是进行搜索的时候,当一篇文档的两页都包含关键词的时候,此文档id在结果集中出现两次,这是我们不想看到的,DuplicateFilter就是指定一个域如id,在此域相同的文档仅取其中一篇。
DuplicateFilter包含以下成员变量:
其getDocIdSet函数如下:
public DocIdSet getDocIdSet(IndexReader reader) throws IOException { if(processingMode==PM_FAST_INVALIDATION) { return fastBits(reader); } else { return correctBits(reader); } } |
private OpenBitSet correctBits(IndexReader reader) throws IOException { OpenBitSet bits=new OpenBitSet(reader.maxDoc()); Term startTerm=new Term(fieldName); TermEnum te = reader.terms(startTerm); if(te!=null) { Term currTerm=te.term(); //如果属于指定的域 while((currTerm!=null)&&(currTerm.field()==startTerm.field())) { int lastDoc=-1; //则取出包含此term的所有的文档 TermDocs td = reader.termDocs(currTerm); if(td.next()) { if(keepMode==KM_USE_FIRST_OCCURRENCE) { //第一篇设为true bits.set(td.doc()); } else { do { lastDoc=td.doc(); }while(td.next()); bits.set(lastDoc); //最后一篇设为true } } if(!te.next()) { break; } currTerm=te.term(); } } return bits; } |
private OpenBitSet fastBits(IndexReader reader) throws IOException { OpenBitSet bits=new OpenBitSet(reader.maxDoc()); bits.set(0,reader.maxDoc()); //全部设为true Term startTerm=new Term(fieldName); TermEnum te = reader.terms(startTerm); if(te!=null) { Term currTerm=te.term(); //如果属于指定的域 while((currTerm!=null)&&(currTerm.field()==startTerm.field())) { if(te.docFreq()>1) { int lastDoc=-1; //取出所有的文档 TermDocs td = reader.termDocs(currTerm); td.next(); if(keepMode==KM_USE_FIRST_OCCURRENCE) { //除了第一篇不清零 td.next(); } do { lastDoc=td.doc(); bits.clear(lastDoc); //其他全部清零 }while(td.next()); if(keepMode==KM_USE_LAST_OCCURRENCE) { bits.set(lastDoc); //最后一篇设为true } } if(!te.next()) { break; } currTerm=te.term(); } } return bits; } |
举例,我们索引如下的文件:
File indexDir = new File("TestDuplicateFilter/index"); doc = new Document(); doc = new Document(); doc = new Document(); doc = new Document(); |
如果搜索TermQuery tq = new TermQuery(new Term("contents","hello")),则结果为:
id : 1 |
如果按如下进行搜索:
File indexDir = new File("TestDuplicateFilter/index"); |
则结果为:
id : 1 |
在介绍与FieldCache相关的Filter之前,先介绍FieldCache。
FieldCache缓存的是不是存储域的内容,而是索引域中term的内容,索引中的term是String的类型,然而可以将其他的类型作为String类型索引进去,例如"1","2.3"等,然后搜索的时候将这些信息取出来。
FieldCache支持如下类型:
其中StringIndex包含两个成员:
FieldCache默认的实现FieldCacheImpl,其中包含成员变量Map<Class<?>,Cache> caches保存从类型到Cache的映射。
private synchronized void init() { caches = new HashMap<Class<?>,Cache>(7); caches.put(Byte.TYPE, new ByteCache(this)); caches.put(Short.TYPE, new ShortCache(this)); caches.put(Integer.TYPE, new IntCache(this)); caches.put(Float.TYPE, new FloatCache(this)); caches.put(Long.TYPE, new LongCache(this)); caches.put(Double.TYPE, new DoubleCache(this)); caches.put(String.class, new StringCache(this)); caches.put(StringIndex.class, new StringIndexCache(this)); } |
其实现接口getInts 如下,即先得到Integer类型所对应的IntCache然后,再从其中根据reader和由field和parser组成的Entry得到整型值。
public int[] getInts(IndexReader reader, String field, IntParser parser) throws IOException { return (int[]) caches.get(Integer.TYPE).get(reader, new Entry(field, parser)); } |
各类缓存的父类Cache包含成员变量Map<Object, Map<Entry, Object>> readerCache,其中key是IndexReader,value是一个Map,此Map的key是Entry,也即是field,value是缓存的int[]的值。(也即在这个reader的这个field中有一个数组的int,每一项代表一篇文档)。
Cache的get函数如下:
public Object get(IndexReader reader, Entry key) throws IOException { Map<Entry,Object> innerCache; Object value; final Object readerKey = reader.getFieldCacheKey(); //此函数返回this,也即IndexReader本身 synchronized (readerCache) { innerCache = readerCache.get(readerKey); //通过IndexReader得到Map if (innerCache == null) { //如果没有则新建一个Map innerCache = new HashMap<Entry,Object>(); readerCache.put(readerKey, innerCache); value = null; } else { value = innerCache.get(key); //此Map的key是Entry,value即是缓存的值 } //如果缓存不命中,则创建此值 if (value == null) { value = new CreationPlaceholder(); innerCache.put(key, value); } } if (value instanceof CreationPlaceholder) { synchronized (value) { CreationPlaceholder progress = (CreationPlaceholder) value; if (progress.value == null) { progress.value = createValue(reader, key); //调用此函数创建缓存值 synchronized (readerCache) { innerCache.put(key, progress.value); } } } return progress.value; } return value; } |
Cache的createValue函数根据类型的不同而不同,我们仅分析IntCache和StringIndexCache的实现.
IntCache的createValue函数如下:
protected Object createValue(IndexReader reader, Entry entryKey) throws IOException { Entry entry = entryKey; String field = entry.field; IntParser parser = (IntParser) entry.custom; int[] retArray = null; TermDocs termDocs = reader.termDocs(); TermEnum termEnum = reader.terms (new Term (field)); try { //依次将域中所有的term都取出来,用IntParser进行解析,缓存retArray[]位置即文档号,retArray[i]即第i篇文档所包含的int值. do { Term term = termEnum.term(); if (term==null || term.field() != field) break; int termval = parser.parseInt(term.text()); if (retArray == null) retArray = new int[reader.maxDoc()]; termDocs.seek (termEnum); while (termDocs.next()) { retArray[termDocs.doc()] = termval; } } while (termEnum.next()); } catch (StopFillCacheException stop) { } finally { termDocs.close(); termEnum.close(); } if (retArray == null) retArray = new int[reader.maxDoc()]; return retArray; } }; |
StringIndexCache的createValue函数如下:
protected Object createValue(IndexReader reader, Entry entryKey) throws IOException { String field = StringHelper.intern(entryKey.field); final int[] retArray = new int[reader.maxDoc()]; String[] mterms = new String[reader.maxDoc()+1]; TermDocs termDocs = reader.termDocs(); TermEnum termEnum = reader.terms (new Term (field)); int t = 0; mterms[t++] = null; try { do { Term term = termEnum.term(); if (term==null || term.field() != field) break; mterms[t] = term.text(); //mterms[i]保存的是按照字典顺序第i个term所对应的字符串。 termDocs.seek (termEnum); while (termDocs.next()) { retArray[termDocs.doc()] = t; //retArray[i]保存的是第i篇文档所包含的字符串在mterms中的位置。 } t++; } while (termEnum.next()); } finally { termDocs.close(); termEnum.close(); } if (t == 0) { mterms = new String[1]; } else if (t < mterms.length) { String[] terms = new String[t]; System.arraycopy (mterms, 0, terms, 0, t); mterms = terms; } StringIndex value = new StringIndex (retArray, mterms); return value; } |
FieldCacheRangeFilter的可以是各种类型的Range,其中Int类型用下面的函数生成:
public static FieldCacheRangeFilter<Integer> newIntRange(String field, FieldCache.IntParser parser, Integer lowerVal, Integer upperVal, boolean includeLower, boolean includeUpper) { return new FieldCacheRangeFilter<Integer>(field, parser, lowerVal, upperVal, includeLower, includeUpper) { @Override public DocIdSet getDocIdSet(IndexReader reader) throws IOException { final int inclusiveLowerPoint, inclusiveUpperPoint; //计算左边界 if (lowerVal != null) { int i = lowerVal.intValue(); if (!includeLower && i == Integer.MAX_VALUE) return DocIdSet.EMPTY_DOCIDSET; inclusiveLowerPoint = includeLower ? i : (i + 1); } else { inclusiveLowerPoint = Integer.MIN_VALUE; } //计算右边界 if (upperVal != null) { int i = upperVal.intValue(); if (!includeUpper && i == Integer.MIN_VALUE) return DocIdSet.EMPTY_DOCIDSET; inclusiveUpperPoint = includeUpper ? i : (i - 1); } else { inclusiveUpperPoint = Integer.MAX_VALUE; } if (inclusiveLowerPoint > inclusiveUpperPoint) return DocIdSet.EMPTY_DOCIDSET; //从cache中取出values,values[i]表示第i篇文档在此域中的值 final int[] values = FieldCache.DEFAULT.getInts(reader, field, (FieldCache.IntParser) parser); return new FieldCacheDocIdSet(reader, (inclusiveLowerPoint <= 0 && inclusiveUpperPoint >= 0)) { @Override boolean matchDoc(int doc) { //仅在文档i所对应的值在区间内的时候才返回。 return values[doc] >= inclusiveLowerPoint && values[doc] <= inclusiveUpperPoint; } }; } }; } |
FieldCacheRangeFilter同NumericRangeFilter或者TermRangeFilter功能类似,只不过后两者取得docid的bitset都是从索引中取出,而前者是缓存了的,加快了速度。
同样FieldCacheTermsFilter同TermFilter功能类似,也是前者进行了缓存,加快了速度。
MultiTermQueryWrapperFilter包含成员变量Q query,其getDocIdSet得到满足此query的文档号bitset。
public DocIdSet getDocIdSet(IndexReader reader) throws IOException { final TermEnum enumerator = query.getEnum(reader); try { if (enumerator.term() == null) return DocIdSet.EMPTY_DOCIDSET; final OpenBitSet bitSet = new OpenBitSet(reader.maxDoc()); final int[] docs = new int[32]; final int[] freqs = new int[32]; TermDocs termDocs = reader.termDocs(); try { int termCount = 0; //遍历满足query的所有term do { Term term = enumerator.term(); if (term == null) break; termCount++; termDocs.seek(term); while (true) { //得到每个term的文档号列表,放入bitset final int count = termDocs.read(docs, freqs); if (count != 0) { for(int i=0;i<count;i++) { bitSet.set(docs[i]); } } else { break; } } } while (enumerator.next()); query.incTotalNumberOfTerms(termCount); } finally { termDocs.close(); } return bitSet; } finally { enumerator.close(); } } |
MultiTermQueryWrapperFilter有三个重要的子类:
其包含一个查询对象,getDocIdSet会获得所有满足此查询的文档号:
public DocIdSet getDocIdSet(final IndexReader reader) throws IOException { final Weight weight = query.weight(new IndexSearcher(reader)); return new DocIdSet() { public DocIdSetIterator iterator() throws IOException { return weight.scorer(reader, true, false); //Scorer的next即返回一个个文档号。 } }; } |
其包含一个SpanQuery query,作为过滤器,其除了通过getDocIdSet得到文档号之外,bitSpans函数得到的SpanFilterResult还包含位置信息,可以用于在FilterQuery中起过滤作用。
public DocIdSet getDocIdSet(IndexReader reader) throws IOException { SpanFilterResult result = bitSpans(reader); return result.getDocIdSet(); } |
public SpanFilterResult bitSpans(IndexReader reader) throws IOException { final OpenBitSet bits = new OpenBitSet(reader.maxDoc()); Spans spans = query.getSpans(reader); List<SpanFilterResult.PositionInfo> tmp = new ArrayList<SpanFilterResult.PositionInfo>(20); int currentDoc = -1; SpanFilterResult.PositionInfo currentInfo = null; while (spans.next()) { //将docid放入bitset int doc = spans.doc(); bits.set(doc); if (currentDoc != doc) { currentInfo = new SpanFilterResult.PositionInfo(doc); tmp.add(currentInfo); currentDoc = doc; } //将start和end信息放入PositionInfo currentInfo.addPosition(spans.start(), spans.end()); } return new SpanFilterResult(bits, tmp); } |
由Filter的接口DocIdSet getDocIdSet(IndexReader reader)得知,一个docid的bitset是同一个reader相对应的。
有前面对docid的描述可知,其仅对一个打开的reader有意义。
CachingSpanFilter有一个成员变量Map<IndexReader,SpanFilterResult> cache保存从reader到SpanFilterResult的映射,另一个成员变量SpanFilter filter用于缓存不命中的时候得到SpanFilterResult。
其getDocIdSet如下:
public DocIdSet getDocIdSet(IndexReader reader) throws IOException { SpanFilterResult result = getCachedResult(reader); return result != null ? result.getDocIdSet() : null; } |
private SpanFilterResult getCachedResult(IndexReader reader) throws IOException { lock.lock(); try { if (cache == null) { cache = new WeakHashMap<IndexReader,SpanFilterResult>(); } //如果缓存命中,则返回缓存中的结果。 final SpanFilterResult cached = cache.get(reader); if (cached != null) return cached; } finally { lock.unlock(); } //如果缓存不命中,则用SpanFilter直接从reader中得到结果。 final SpanFilterResult result = filter.bitSpans(reader); lock.lock(); try { //将新得到的结果放入缓存 cache.put(reader, result); } finally { lock.unlock(); } return result; } |