public EntityInfo Extract(Hits hits, int index)
{
Document doc = hits.Doc(index);
//TODO if we are lonly looking for score (unlikely), avoid accessing doc (lazy load)
EntityInfo entityInfo = Extract(doc);
object[] eip = entityInfo.Projection;
if (eip != null && eip.Length > 0)
{
for (int x = 0; x < projection.Length; x++)
{
switch (projection[x])
{
case ProjectionConstants.SCORE:
eip[x] = hits.Score(index);
break;
case ProjectionConstants.ID:
eip[x] = entityInfo.Id;
break;
case ProjectionConstants.DOCUMENT:
eip[x] = doc;
break;
case ProjectionConstants.DOCUMENT_ID:
eip[x] = hits.Id(index);
break;
case ProjectionConstants.BOOST:
eip[x] = doc.GetBoost();
break;
case ProjectionConstants.THIS:
//THIS could be projected more than once
//THIS loading delayed to the Loader phase
if (entityInfo.IndexesOfThis == null)
{
entityInfo.IndexesOfThis = new List <int>(1);
}
entityInfo.IndexesOfThis.Add(x);
break;
}
}
}
return(entityInfo);
}
/// <summary>
/// Converts storyId as used in the database to the docId as used in lucene
/// </summary>
/// <param name="hostId"></param>
/// <param name="storyId"></param>
/// <returns>returns the docId for the story from lucene. If the story is not
/// found a null value will be returned</returns>
protected int?ConvertStoryIdtoDocId(int hostId, int storyId)
{
QueryParser queryParser = new QueryParser("id", new DnkAnalyzer());
Query q = queryParser.Parse(storyId.ToString());
IndexSearcher searcher = SearchQuery.GetSearcher(hostId);
Hits hits = searcher.Search(q);
if (hits.Length() > 0)
{
return(hits.Id(0));
}
return(null);
}
private int[] GetDocumentIdsForElementId(string elementId)
{
List <int> result = new List <int>();
IndexSearcher searcher = new IndexSearcher(CurrentDirectory);
Hits hits = searcher.Search(new TermQuery(new Term(LuceneNodeIndexer.ELEMENTID_FOR_DELETING_FIELD, elementId)));
for (int x = 0; x < hits.Length(); x++)
{
result.Add(hits.Id(x));
}
return(result.ToArray());
}
/**
* Lucene doesn't let us do deletes from an IndexWriter, only an IndexReader.
* To avoid locking or constant open-close switches, be sure that writing
* is only happening on an in-memory writer.
* @param article
* @throws IOException
*/
public void DeleteArticle(Article article)
{
OpenReader();
string key = article.Key;
Hits hits = searcher.Search(new TermQuery(
new Term("key", key)));
if (hits.Length() == 0)
{
log.Debug("Nothing to delete for " + key);
}
for (int i = 0; i < hits.Length(); i++)
{
int id = hits.Id(i);
log.Debug("Trying to delete article number " + id + " for " + key);
try {
reader.Delete(id);
} catch (IOException e) {
log.Warn("Couldn't delete article number " +
id + " for " + key + "... " + e);
}
}
}
public override void Search(string SearchString)
{
base.Search(SearchString);
SearchString = SearchString.ToLower();
var dir = FSDirectory.GetDirectory(IndexPath, false);
var searcher = new IndexSearcher(dir);
var parser = new QueryParser(ContentField, new StandardAnalyzer());
foreach (var s in SearchString.Split(new [] { ' ' }))
{
var query = parser.GetFuzzyQuery(ContentField, s, MinSimilarity);
Hits hits = searcher.Search(query);
for (int i = 0; i < hits.Length(); i++)
{
Document doc = hits.Doc(i);
var result = new SearchResult {
Score = hits.Score(i), Plugin = doc.Get(PluginField)
};
//Text des aktuellen Dokuments auslesen
string text = doc.Get(ContentField);
//Alle indizierten Wörter dieses Dokumentes auslesen
var tpv = (TermPositionVector)IndexReader.Open(dir).GetTermFreqVector(hits.Id(i), ContentField);
String[] DocTerms = tpv.GetTerms();
//Die Anzahl der Erscheinungen aller Wörter auslesen
int[] freq = tpv.GetTermFrequencies();
var words = new List <string>(DocTerms);
//Hier wollen wir nun die Positionen der Erscheinungen des Suchwortes herausfinden
for (int t = 0; t < freq.Length; t++)
{
//Falls das Suchwort mit dem aktuellen Wort übereinstimmt...
if (ContainsSearchString(SearchString, DocTerms[t], words))
{
//...können wir die Positionen auslesen
TermVectorOffsetInfo[] offsets = tpv.GetOffsets(t);
//Das Array beinhaltet nun für das Suchwort alle Auftreten mit jeweils Anfang und Ende
for (int j = 0; j < offsets.Length; j++)
{
//Jetz muss nur noch ein kleiner Kontextausschnitt ausgelesen werden, damit der User etwas damit anfangen kann
int start = offsets[j].GetStartOffset();
int end = offsets[j].GetEndOffset();
int contextStart = start - ContextLeftOffset;
contextStart = contextStart < 0 ? 0 : contextStart;
int contextEnd = end + ContextRightOffset;
contextEnd = contextEnd > text.Length ? text.Length : contextEnd;
//Nun wollen wir noch bis zum Ende des nächsten Wortes lesen, um das Ergebnis besser lesbar zu machen
int nextEndSpace = text.IndexOf(" ", contextEnd);
contextEnd = nextEndSpace > 0 ? nextEndSpace : contextEnd;
//Maximal so viele Zeichen darf der Text nach einem Leerzeichen links von dem Suchergebnis durchsucht werden
int leftSpaceOffset = contextStart;
//Finden des nächstenLeerzeichens links vom Suchergebnis
int nextStartSpace = text.LastIndexOf(" ", contextStart, leftSpaceOffset);
//Falls es kein Space in der Nöhe gibt brauchen wir natürlich auch nichts verändern
contextStart = nextStartSpace > 0 ? nextStartSpace : contextStart;
int contextLength = contextEnd - contextStart;
contextLength = contextLength > text.Length ? text.Length : contextLength;
//Kontext auslesen
string context = text.Substring(contextStart, contextLength);
//und den Searchresults zusammen mit dem zugehörigen PlugInNamen und dem HitScore hinzufügen
result.Contexts.Add(context);
}
}
}
SearchResults.Add(result);
}
}
}
/// <summary>
/// 查询索引
/// </summary>
/// <param name="fieldName">FieldName</param>
/// <param name="keywords">关键字</param>
/// <param name="pageIndex">当前页</param>
/// <param name="pageSize">分页大小</param>
/// <param name="totalRecord">总的记录</param>
/// <returns>索引列表</returns>
/// <remarks>2013-08-15 朱成果 创建</remarks>
public List <CBPdProductIndex> QueryDoc(string fieldName, string keywords, int pageIndex, int pageSize, out int totalRecord)
{
var search = new IndexSearcher(IndexStorePath);
Query searchQuery;
if (!string.IsNullOrEmpty(fieldName) && !string.IsNullOrEmpty(keywords))
{
#region [关键字查询]
var query = new BooleanQuery();
BooleanQuery childQuery;
BooleanQuery esenQuery;
if (fieldName == "ProductName")
{
#region 2016-4-6 杨浩 新增模糊搜索
childQuery = new BooleanQuery();
esenQuery = new BooleanQuery();
//模糊搜索
//esenQuery.Add(new FuzzyQuery(new Term("ProductName", Regex.Replace(keywords, @"\s", ""))), BooleanClause.Occur.SHOULD);
//esenQuery.SetBoost(4.0F);
//分词 盘古分词
var keyWordsSplitBySpace = GetKeyWordsSplitBySpace(keywords);
//string keyWordsSplitBySpace = string.Format("{0}^{1}.0", keywords, (int)Math.Pow(3, 5));
//不启用分词，直接用模糊搜索
QueryParser productNameQueryParser = new QueryParser(global::Lucene.Net.Util.Version.LUCENE_29, "ProductName", new PanGuAnalyzer(true));
Query productNameQuery = productNameQueryParser.Parse(keyWordsSplitBySpace);
childQuery.Add(productNameQuery, BooleanClause.Occur.SHOULD);
//以什么开头，输入“ja”就可以搜到包含java和javascript两项结果了
Query prefixQuery_productName = new PrefixQuery(new Term("ProductName", keywords.Trim()));
//直接模糊匹配,假设你想搜索跟‘wuzza’相似的词语,你可能得到‘fuzzy’和‘wuzzy’。
Query fuzzyQuery_productName = new FuzzyQuery(new Term("ProductName", keywords.Trim()));
//通配符搜索
Query wildcardQuery_productName = new WildcardQuery(new Term("ProductName", string.Format("{0}", keywords.Trim())));
childQuery.Add(prefixQuery_productName, BooleanClause.Occur.SHOULD);
childQuery.Add(fuzzyQuery_productName, BooleanClause.Occur.SHOULD);
childQuery.Add(wildcardQuery_productName, BooleanClause.Occur.SHOULD);
childQuery.SetBoost(4.0F);
//childQuery.Add(esenQuery, BooleanClause.Occur.SHOULD);
query.Add(childQuery, BooleanClause.Occur.MUST);
#endregion
//childQuery = new BooleanQuery();
//esenQuery = new BooleanQuery();
////全词去空格
//esenQuery.Add(new TermQuery(new Term("ProductName", Regex.Replace(keywords, @"\s", ""))),
// BooleanClause.Occur.SHOULD);
//esenQuery.SetBoost(3.0F);
//childQuery.Add(esenQuery, BooleanClause.Occur.SHOULD);
//esenQuery = new BooleanQuery();
////分词 盘古分词
//esenQuery.Add(new QueryParser("ProductName", new PanGuAnalyzer(true)).Parse(keywords),
// BooleanClause.Occur.SHOULD);
////分词 按空格
//var keyColl = keywords.Split(new char[] { ' ' }, StringSplitOptions.RemoveEmptyEntries);
//foreach (var item in keyColl)
//{
// esenQuery.Add(new TermQuery(new Term("ProductName", item)),
// BooleanClause.Occur.SHOULD);
//}
//esenQuery.SetBoost(2.9F);
//childQuery.Add(esenQuery, BooleanClause.Occur.SHOULD);
//query.Add(childQuery, BooleanClause.Occur.MUST);
}
else if (fieldName == "Category")
{
childQuery = new BooleanQuery();
esenQuery = new BooleanQuery();
esenQuery.Add(new TermQuery(new Term("Category", keywords)),
BooleanClause.Occur.SHOULD);
esenQuery.SetBoost(3.0F);
childQuery.Add(esenQuery, BooleanClause.Occur.SHOULD);
esenQuery = new BooleanQuery();
esenQuery.Add(new WildcardQuery(new Term("AssociationCategory", string.Format("*,{0},*", keywords))),
BooleanClause.Occur.SHOULD);
esenQuery.SetBoost(2.8F);
childQuery.Add(esenQuery, BooleanClause.Occur.SHOULD);
query.Add(childQuery, BooleanClause.Occur.MUST);
}
else if (fieldName == "BrandSysNo")
{
childQuery = new BooleanQuery();
childQuery.Add(new TermQuery(new Term("BrandSysNo", keywords)),
BooleanClause.Occur.SHOULD);
childQuery.SetBoost(3.0F);
query.Add(childQuery, BooleanClause.Occur.MUST);
}
else if (fieldName == "DealerSysNos")
{
childQuery = new BooleanQuery();
childQuery.Add(new WildcardQuery(new Term("DealerSysNos", string.Format("*,{0},*", keywords))),
BooleanClause.Occur.SHOULD);
childQuery.SetBoost(2.8F);
query.Add(childQuery, BooleanClause.Occur.MUST);
}
else if (fieldName == "ProductGroupCode")
{
childQuery = new BooleanQuery();
childQuery.Add(new WildcardQuery(new Term("ProductGroupCode", string.Format("*,{0},*", keywords))),
BooleanClause.Occur.SHOULD);
childQuery.SetBoost(2.8F);
query.Add(childQuery, BooleanClause.Occur.MUST);
}
else
{
query.Add(new TermQuery(new Term(fieldName, keywords)),
BooleanClause.Occur.SHOULD);
}
#endregion
searchQuery = query;
}
else
{
searchQuery = new WildcardQuery(new Term("ProductName", "*雪花秀*"));
}
//排序方式
var sort = new Sort();
//搜索
Hits hits = search.Search(searchQuery, sort);
totalRecord = hits.Length();//总的记录
int startIndex = (pageIndex - 1) * pageSize;
if (startIndex < 0)
{
startIndex = 0;
}
int endIndex = startIndex + pageSize;
if (endIndex > totalRecord - 1)
{
endIndex = totalRecord - 1;
}
List <CBPdProductIndex> lst = new List <CBPdProductIndex>();
for (int i = startIndex; i <= endIndex; i++)
{
var doc = hits.Doc(i);
lst.Add(
new CBPdProductIndex
{
DocID = hits.Id(i),
Score = hits.Score(i),
AssociationCategory = doc.Get("AssociationCategory"),
Attributes = doc.Get("Attributes"),
Barcode = doc.Get("Barcode"),
BrandSysNo = Convert.ToInt32(doc.Get("BrandSysNo")),
Category = Convert.ToInt32(doc.Get("Category")),
DisplayOrder = Convert.ToInt32(doc.Get("DisplayOrder")),
NameAcronymy = doc.Get("NameAcronymy"),
Prices = doc.Get("Prices"),
ProductImage = doc.Get("ProductImage"),
ProductName = doc.Get("ProductName"),
QRCode = doc.Get("QRCode"),
Status = Convert.ToInt32(doc.Get("Status")),
SysNo = Convert.ToInt32(doc.Get("SysNo")),
BasicPrice = Convert.ToDecimal(doc.Get("BasicPrice")),
Price = Convert.ToDecimal(doc.Get("Price")),
DispalySymbol = 0,
RankPrice = 0.00M,
ProductGroupCode = Convert.ToString(doc.Get("ProductGroupCode")),
DealerSysNos = doc.Get("DealerSysNos"),
WarehouseSysNos = doc.Get("WarehouseSysNos")
});
}
search.Close();
return(lst);
}
public List <Post> Similar(int postid, int itemsToReturn)
{
List <Post> TList = new List <Post>();
int docId = -1;
IndexSearcher searcher = null;
IndexReader reader = null;
if (rd == null)
{
BuildIndex();
}
lck.AcquireReaderLock(ReaderTimeOut);
try
{
Analyzer analyzer = GetAnalyzer();
QueryParser parser = GetQueryParser(analyzer);
parser.SetDefaultOperator(QueryParser.AND_OPERATOR);
Query q = parser.Parse("postid:" + postid);
searcher = new IndexSearcher(rd, true);
//TODO
#pragma warning disable CS0618 // Type or member is obsolete
Hits hits = searcher.Search(q);
#pragma warning restore CS0618 // Type or member is obsolete
if (hits != null && hits.Length() > 0)
{
docId = hits.Id(0);
}
if (docId > -1)
{
reader = IndexReader.Open(rd, true);
TermFreqVector tfv = reader.GetTermFreqVector(docId, "exact");
BooleanQuery booleanQuery = new BooleanQuery();
for (int j = 0; j < tfv.Size(); j++)
{
TermQuery tq = new TermQuery(new Term("exact", tfv.GetTerms()[j]));
booleanQuery.Add(tq, BooleanClause.Occur.SHOULD);
}
//TODO
#pragma warning disable CS0618 // Type or member is obsolete
Hits similarhits = searcher.Search(booleanQuery, Sort.RELEVANCE);
#pragma warning restore CS0618 // Type or member is obsolete
for (int i = 0; i < similarhits.Length(); i++)
{
Document doc = similarhits.Doc(i);
if (similarhits.Id(i) != docId)
{
TList.Add(CreateFromDocument(doc, analyzer, null));
}
if (TList.Count >= itemsToReturn)
{
break;
}
}
}
}
catch (Exception)
{
}
finally
{
if (searcher != null)
{
searcher.Close();
}
if (reader != null)
{
reader.Close();
}
lck.ReleaseReaderLock();
}
return(TList);
}
