/// <summary>
/// Searches the index.
/// </summary>
/// <param name="indexPath">The index path.</param>
/// <param name="searchText">The search text.</param>
/// <returns></returns>
public List <IndexInformation> SearchIndex(string indexPath, string searchText)
{
List <IndexInformation> searchResults = new List <IndexInformation>();
try
{
using (Lucene.Net.Store.Directory directory = FSDirectory.Open(new DirectoryInfo(indexPath)))
{
Analyzer analyzer = new StandardAnalyzer(Lucene.Net.Util.Version.LUCENE_29);
var parser = new QueryParser(Lucene.Net.Util.Version.LUCENE_29, RegularExpression, analyzer);
Query query = parser.Parse(searchText);
using (var searcher = new IndexSearcher(directory, true))
{
Hits topDocs = searcher.Search(query);
if (topDocs != null)
{
for (int i = 0; i < topDocs.Length(); i++)
{
searchResults.Add(new IndexInformation {
FilePath = topDocs.Doc(i).Get(FilePath), Date = topDocs.Doc(i).Get(Date)
});
}
}
}
}
}
catch (Exception ex)
{
throw;
}
return(searchResults);
}
public Document this [int i]
{
get
{
return(hits.Doc(i));
}
}
//public List<int> SearchIuses(string searchTerm)
//{
// List<int> results = new List<int>();
// IndexSearcher searcher = new IndexSearcher(FSDirectory.GetDirectory(indexPath));
// QueryParser parser = new QueryParser("Rubro", analyzer);
// parser.SetEnablePositionIncrements(false);
// PhraseQuery q = new PhraseQuery();
// String[] words = searchTerm.Split(' ');
// foreach (string word in words)
// {
// q.Add(new Term("Rubro", word));
// }
// Console.WriteLine(q.ToString());
// //Query query = parser.Parse(searchTerm);
// Hits hitsFound = searcher.Search(q);
// TesisIndx sampleDataFileRow = null;
// for (int i = 0; i < hitsFound.Length(); i++)
// {
// sampleDataFileRow = new TesisIndx();
// Document doc = hitsFound.Doc(i);
// sampleDataFileRow.Ius = int.Parse(doc.Get("Ius"));
// sampleDataFileRow.Rubro = doc.Get("Rubro");
// sampleDataFileRow.Texto = doc.Get("Texto");
// float score = hitsFound.Score(i);
// sampleDataFileRow.Score = score;
// results.Add(sampleDataFileRow.Ius);
// }
// parser = new QueryParser("Texto", analyzer);
// parser.SetEnablePositionIncrements(false);
// q = new PhraseQuery();
// words = searchTerm.Split(' ');
// foreach (string word in words)
// {
// q.Add(new Term("Texto", word));
// }
// // query = parser.Parse(searchTerm);
// hitsFound = searcher.Search(q);
// for (int i = 0; i < hitsFound.Length(); i++)
// {
// sampleDataFileRow = new TesisIndx();
// Document doc = hitsFound.Doc(i);
// sampleDataFileRow.Ius = int.Parse(doc.Get("Ius"));
// sampleDataFileRow.Rubro = doc.Get("Rubro");
// sampleDataFileRow.Texto = doc.Get("Texto");
// float score = hitsFound.Score(i);
// sampleDataFileRow.Score = score;
// results.Add(sampleDataFileRow.Ius);
// }
// results.Distinct();
// return results;
//}
/// <summary>
/// Busca en el índice previamente construido las tesis que tengan coincidencia ya sea en el Rubro o Texto
/// del término buscado
/// </summary>
/// <param name="searchTerm"></param>
/// <returns></returns>
public List <int> SearchIuses(string searchTerm)
{
List <int> results = new List <int>();
IndexSearcher searcher = new IndexSearcher(FSDirectory.GetDirectory(indexPath));
QueryParser parser = new QueryParser("RubroIndx", analyzer);
parser.SetEnablePositionIncrements(false);
Query query = parser.Parse(String.Format("\"{0}\"", searchTerm));
Console.WriteLine(query.ToString());
Hits hitsFound = searcher.Search(query);
TesisIndx sampleDataFileRow = null;
for (int i = 0; i < hitsFound.Length(); i++)
{
sampleDataFileRow = new TesisIndx();
Document doc = hitsFound.Doc(i);
sampleDataFileRow.Ius = int.Parse(doc.Get("Ius"));
sampleDataFileRow.RubroIndx = doc.Get("RubroIndx");
sampleDataFileRow.TextoIndx = doc.Get("TextoIndx");
float score = hitsFound.Score(i);
sampleDataFileRow.Score = score;
results.Add(sampleDataFileRow.Ius);
}
parser = new QueryParser("TextoIndx", analyzer);
parser.SetEnablePositionIncrements(false);
query = parser.Parse(String.Format("\"{0}\"", searchTerm));
Console.WriteLine(query.ToString());
hitsFound = searcher.Search(query);
for (int i = 0; i < hitsFound.Length(); i++)
{
sampleDataFileRow = new TesisIndx();
Document doc = hitsFound.Doc(i);
sampleDataFileRow.Ius = int.Parse(doc.Get("Ius"));
sampleDataFileRow.Rubro = doc.Get("RubroIndx");
sampleDataFileRow.Texto = doc.Get("TextoIndx");
float score = hitsFound.Score(i);
sampleDataFileRow.Score = score;
results.Add(sampleDataFileRow.Ius);
}
results.Distinct();
return(results);
}
/// <summary>
/// search
/// </summary>
/// <param name="indexDir"></param>
/// <param name="q">keyword</param>
/// <param name="pageLen">every page's length</param>
/// <param name="pageNo">page number</param>
/// <param name="recCount">result number</param>
/// <returns></returns>
public List <TDocs> Search(String q, int pageLen, int pageNo, out int recCount)
{
string keywords = q;
IndexSearcher search = new IndexSearcher(INDEX_DIR);
q = GetKeyWordsSplitBySpace(q, new PanGuTokenizer());
QueryParser queryParser = new QueryParser("content", new PanGuAnalyzer(true));
Query query = queryParser.Parse(q);
Hits hits = search.Search(query, Sort.RELEVANCE);
List <TDocs> result = new List <TDocs>();
recCount = hits.Length();
int i = (pageNo - 1) * pageLen;
while (i < recCount && result.Count < pageLen)
{
TDocs docs = null;
try
{
docs = new TDocs();
docs.Path = hits.Doc(i).Get("path");
docs.Name = hits.Doc(i).Get("name");
docs.Title = hits.Doc(i).Get("title");
docs.Extension = hits.Doc(i).Get("ext");
//rem this item in case the search result will be too large & consume too much memory,
// takes loading time, abstract is enough
//docs.Content = hits.Doc(i).Get("content");
PanGu.HighLight.SimpleHTMLFormatter simpleHTMLFormatter =
new PanGu.HighLight.SimpleHTMLFormatter("<color=red>", "</color>");
PanGu.HighLight.Highlighter highlighter =
new PanGu.HighLight.Highlighter(simpleHTMLFormatter,
new Segment());
highlighter.FragmentSize = 100;
docs.Abstract = highlighter.GetBestFragment(keywords, hits.Doc(i).Get("content"));
}
catch (Exception e)
{
Debug.WriteLine(e.Message);
}
finally
{
result.Add(docs);
i++;
}
}
search.Close();
return(result);
}
/// <summary>
/// Searches the index.
/// </summary>
/// <param name="queryText"></param>
/// <param name="categoryNames"></param>
/// <param name="pageIndex"></param>
/// <param name="pageSize"></param>
/// <param name="roleIds"></param>
/// <returns></returns>
public SearchResultCollection Find(string queryText, IList <string> categoryNames, int pageIndex, int pageSize, IEnumerable <int> roleIds)
{
long startTicks = DateTime.Now.Ticks;
// the overall-query
BooleanQuery query = new BooleanQuery();
// add our parsed query
if (!String.IsNullOrEmpty(queryText))
{
Query multiQuery = MultiFieldQueryParser.Parse(new[] { queryText, queryText, queryText }, new[] { "title", "summary", "contents" }, new StandardAnalyzer());
query.Add(multiQuery, BooleanClause.Occur.MUST);
}
// add the security constraint - must be satisfied
query.Add(this.BuildSecurityQuery(roleIds), BooleanClause.Occur.MUST);
// Add the category query (if available)
if (categoryNames != null)
{
query.Add(this.BuildCategoryQuery(categoryNames), BooleanClause.Occur.MUST);
}
IndexSearcher searcher = new IndexSearcher(this._indexDirectory);
Hits hits = searcher.Search(query);
int start = pageIndex * pageSize;
int end = (pageIndex + 1) * pageSize;
if (hits.Length() <= end)
{
end = hits.Length();
}
SearchResultCollection results = new SearchResultCollection(end);
results.TotalCount = hits.Length();
results.PageIndex = pageIndex;
for (int i = start; i < end; i++)
{
SearchResult result = new SearchResult();
result.Title = hits.Doc(i).Get("title");
result.Summary = hits.Doc(i).Get("summary");
result.Author = hits.Doc(i).Get("author");
result.ModuleType = hits.Doc(i).Get("moduletype");
result.Path = hits.Doc(i).Get("path");
string[] categories = hits.Doc(i).GetValues("category");
result.Category = categories != null?String.Join(", ", categories) : String.Empty;
result.DateCreated = DateTime.Parse((hits.Doc(i).Get("datecreated")));
result.Score = hits.Score(i);
result.Boost = hits.Doc(i).GetBoost();
result.SectionId = Int32.Parse(hits.Doc(i).Get("sectionid"));
results.Add(result);
}
searcher.Close();
results.ExecutionTime = DateTime.Now.Ticks - startTicks;
return(results);
}
public TokenResult FindExact(TokenParameter parameter)
{
IndexSearcher indexSearcher = Searcher;
string projectIdentifier = string.Format("{0}_{1}", parameter.Username, parameter.Project).ToLower();
var projectQuery = new TermQuery(new Term(ItemProjectIdentifier, projectIdentifier));
var identfierQuery = new TermQuery(new Term(ItemIdentifier, string.Format("{0}", parameter.FullyQualifiedName)));
var booleanQuery = new BooleanQuery();
booleanQuery.Add(projectQuery, BooleanClause.Occur.MUST);
booleanQuery.Add(identfierQuery, BooleanClause.Occur.MUST);
Hits hits = indexSearcher.Search(booleanQuery);
if (hits.Length() > 0)
{
var document = hits.Doc(0); // Take only the first one
string filePath = document.Get(ItemPath);
int location = Int32.Parse(document.Get(ItemLocation));
var projectCodeDirectory = this.applicationConfigurationProvider.GetProjectSourceCodePath(parameter.Username, parameter.Project);
//string relativePath = filePath //projectCodeDirectory.MakeRelativePath(filePath);
return(new Models.TokenResult
{
FileName = Path.GetFileName(filePath),
Position = location,
Path = filePath
});
}
return(null);
}
void DoTitleMatches()
{
string term = searchterm;
try {
string[] terms = Regex.Split(term, " +");
term = "";
foreach (string t in terms)
{
term += t + "~ ";
}
searchterm = "title:(" + term + ")";
DateTime now = DateTime.UtcNow;
Query query = state.Parse(searchterm);
Hits hits = state.Searcher.Search(query);
int numhits = hits.Length();
LogRequest(searchterm, query, numhits, now);
SendHeaders(200, "OK");
for (int i = 0; i < numhits && i < 10; i++)
{
Document doc = hits.Doc(i);
float score = hits.Score(i);
string pageNamespace = doc.Get("namespace");
string title = doc.Get("title");
SendResultLine(score, pageNamespace, title);
}
} catch (Exception e) {
log.Error(e.Message + e.StackTrace);
}
}
public override SqlString ToSqlString(ICriteria criteria, ICriteriaQuery criteriaQuery,
IDictionary <string, IFilter> enabledFilters)
{
System.Type type = GetCriteriaClass(criteria);
ISearchFactoryImplementor searchFactory = ContextHelper.GetSearchFactory(GetSession(criteria));
Iesi.Collections.Generic.ISet <System.Type> types;
IndexSearcher searcher = FullTextSearchHelper.BuildSearcher(searchFactory, out types, type);
if (searcher == null)
{
throw new SearchException("Could not find a searcher for class: " + type.FullName);
}
Lucene.Net.Search.Query query = FullTextSearchHelper.FilterQueryByClasses(types, luceneQuery);
Hits hits = searcher.Search(query);
List <object> ids = new List <object>();
for (int i = 0; i < hits.Length(); i++)
{
object id = DocumentBuilder.GetDocumentId(searchFactory, type, hits.Doc(i));
ids.Add(id);
}
base.Values = ids.ToArray();
return(base.ToSqlString(criteria, criteriaQuery, enabledFilters));
}
public ArrayList getNotesMatchingTitle(string search)
{
ArrayList snotes = new ArrayList();
try {
QueryParser parser = new QueryParser("title", analyzer);
string lucsearch = search + "*^4" + " content:" + search + "*";
Query query = parser.Parse(lucsearch);
IndexSearcher searcher = new IndexSearcher(lucIdx);
Hits hits = searcher.Search(query);
int results = hits.Length();
Console.WriteLine("Found {0} results", results);
for (int i = 0; i < results; i++)
{
Document doc = hits.Doc(i);
//float score = hits.Score (i);
snotes.Add(new Note(doc.Get("title"), doc.Get("lastmod")));
}
} catch (Exception e) {
Console.WriteLine("ERROR Search: " + e.Message);
}
return(snotes);
}
public static void PrintHits(Hits hits)
{
if (hits == null)
{
Console.WriteLine("Hits: 0");
}
else
{
Console.WriteLine("Hits: " + hits.Length());
// iterate over the first few results.
for (int i = 0; i < 50 && i < hits.Length(); i++)
{
var doc = hits.Doc(i);
string id = doc.Get("id");
int n;
for (n = 0; n < Data.GetLength(0); ++n)
{
if (id == Data[n, 1])
{
Console.WriteLine("{0}: {1} ==> {2} {3}", n, GetPath(n), GetContent(n), GetExternals(n));
break;
}
}
}
Console.WriteLine("-");
}
}
public override List <SearchResult> Search(string searchStr)
{
List <SearchResult> results = new List <SearchResult>();
string cleanSearchStr = cleaner.Replace(searchStr, "").ToLower().Trim();
IndexSearcher searcher = new IndexSearcher(directory);
//QueryParser parser = new QueryParser("title", analyzer);
//Query query = parser.Parse(cleanSearchStr + "~0.7");
Query query = parser.Parse(cleanSearchStr + "~0.7");
Hits hits = searcher.Search(query);
int resultCount = hits.Length();
for (int i = 0; i < resultCount; i++)
{
SearchResult result = new SearchResult();
result.Item = DatabaseManager.Get <T>(int.Parse(hits.Doc(i).Get("id")));
result.Score = hits.Score(i);
results.Add(result);
}
return(results);
}
public Hit GetHitById(string id)
{
IndexSearcher s = GetSearcher();
Hits hits = s.Search(new TermQuery(new Term(FieldName.Id, id)));
return(hits.Length() == 1 ? new Hit(hits.Doc(0), null) : null);
}
public override List <ISearchEntity> GetSearchResult(out int MatchCount)
{
Analyzer analyzer = new StandardAnalyzer();
IndexSearcher searcher = new IndexSearcher(searchInfo.ConfigElement.IndexDirectory);
MultiFieldQueryParser parserName = new MultiFieldQueryParser(new string[] { "title", "content", "keywords" }, analyzer);
Query queryName = parserName.Parse(searchInfo.QueryString);
Hits hits = searcher.Search(queryName);
List <ISearchEntity> ResultList = new List <ISearchEntity>();
for (int i = 0; i < hits.Length(); i++)
{
Document doc = hits.Doc(i);
ResultList.Add((ISearchEntity) new NewsModel()
{
EntityIdentity = Convert.ToInt32(doc.Get("newsid")),
Title = Convert.ToString(doc.Get("title")),
Content = Convert.ToString(doc.Get("content")),
Keywords = doc.Get("keywords")
});
}
searcher.Close();
MatchCount = hits.Length();
return(ResultList);
}
public IEnumerable <DataFileRow> Search(string searchTerm)
{
IndexSearcher searcher = new IndexSearcher(luceneIndexDirectory);
QueryParser parser = new QueryParser("LineText", analyzer);
Query query = parser.Parse(searchTerm);
Hits hitsFound = searcher.Search(query);
List <DataFileRow> results = new List <DataFileRow>();
DataFileRow sampleDataFileRow = null;
for (int i = 0; i < hitsFound.Length(); i++)
{
sampleDataFileRow = new DataFileRow();
Document doc = hitsFound.Doc(i);
sampleDataFileRow.LineNumber = int.Parse(doc.Get("LineNumber"));
sampleDataFileRow.LineText = doc.Get("LineText");
float score = hitsFound.Score(i);
sampleDataFileRow.Score = score;
results.Add(sampleDataFileRow);
}
luceneIndexDirectory.Close();
searcher.Close();
return(results.OrderByDescending(x => x.Score).ToList());
}
public static void Main(string[] args)
{
string indexpath = args[0];
string query = args[1];
IndexSearcher searcher = new IndexSearcher(indexpath);
Query parsedquery = QueryParser.Parse(query,
"summary", new StandardAnalyzer());
Hits hits = searcher.Search(parsedquery);
Console.WriteLine("Found " + hits.Length() + " document(s) that matched query '" + query + "':\n");
for (int i = 0; i < hits.Length(); i++)
{
Document doc = hits.Doc(i);
Console.WriteLine(hits.Score(i) + ": " + doc.Get("excerpt") + "\n");
if (i == 50)
{
break;
}
}
searcher.Close();
}
public string DidYouMean(string pattern)
{
try
{
IndexSearcher searcher = new IndexSearcher(m_HistoryPath);
Term t = new Term(Constants.SearchedText, pattern);
FuzzyQuery query = new FuzzyQuery(t);
Hits hits = searcher.Search(query);
if (hits.Length() != 0)
{
return(hits.Doc(0).Get(Constants.SearchedText));
}
else
{
return("");
}
}
catch (Exception)
{
return("");
}
}
public void Initialize_Indexes_All_Nodes()
{
string elementIdForTestingSearch = _deepNodeFinder.GetNodesForIndexing()[0].Id;
int expectedNumNodes = _deepNodeFinder.GetNodesForIndexing().Length;
Assert.AreEqual("usfr-pte_NetCashFlowsProvidedUsedOperatingActivitiesDirectAbstract", elementIdForTestingSearch,
"TEST SANITY: element id for test search");
Assert.AreEqual(1595, expectedNumNodes, "TEST SANITY: Number of nodes in found in the test taxonomy");
IndexReader indexReader = IndexReader.Open(_indexMgr.LuceneDirectory_ForTesting);
Assert.AreEqual(expectedNumNodes, indexReader.NumDocs(),
"An incorrect number of documents were found in the Lucene directory after initialization");
IndexSearcher searcher = new IndexSearcher(_indexMgr.LuceneDirectory_ForTesting);
try
{
Hits results =
searcher.Search(new TermQuery(new Term(LuceneNodeIndexer.ELEMENTID_FOR_DELETING_FIELD, elementIdForTestingSearch)));
Assert.AreEqual(1, results.Length(), "Search results should only have 1 hit");
Assert.AreEqual(elementIdForTestingSearch, results.Doc(0).Get(LuceneNodeIndexer.ELEMENTID_FIELD),
"Search results yielded the wrong element!");
}
finally
{
searcher.Close();
}
}
protected void BindLucene()
{
StandardAnalyzer analyzer = new Lucene.Net.Analysis.Standard.StandardAnalyzer();
IndexSearcher searcher = new IndexSearcher("C:\\Clients\\NWTD5.0.200\\SearchIndex\\eCommerceFramework\\CatalogEntryIndexer");
BooleanQuery boolQuery = new BooleanQuery();
boolQuery.Add(new TermQuery(new Term("_catalog", "NWTD")), BooleanClause.Occur.MUST);
Hits hits = searcher.Search(boolQuery, new Sort("TypeSort"));
DataTable dt = new DataTable();
dt.Columns.Add("Code");
dt.Columns.Add("Name");
dt.Columns.Add("TypeSort");
for (int hitIndex = 0; hitIndex < hits.Length() && hitIndex < 10; hitIndex++)
{
Document hitDocument = hits.Doc(hitIndex);
object[] param = new[] {
hitDocument.GetField("Code").StringValue(),
hitDocument.GetField("Name").StringValue(),
hitDocument.GetField("TypeSort").StringValue()
};
dt.Rows.Add(param);
}
this.gvLuceneSearchResults.DataSource = dt;
this.gvLuceneSearchResults.DataBind();
}
public static IList <string> Search(string searchTerm)
{
System.Text.RegularExpressions.Regex rgx = new System.Text.RegularExpressions.Regex("[^a-zA-Z0-9]");
searchTerm = rgx.Replace(searchTerm, " ");
IndexSearcher searcher = new IndexSearcher(luceneIndexDirectory);
QueryParser parser = new QueryParser("Data", analyzer);
Query query = parser.Parse(searchTerm.ToLower());
Hits hitsFound = searcher.Search(query);
IList <string> results = new List <string>();
for (int i = 0; i < hitsFound.Length(); i++)
{
Document doc = hitsFound.Doc(i);
float score = hitsFound.Score(i);
string fileName = doc.Get("FileName");
if (score > 0.6)
{
results.Add(doc.Get("FileName"));
}
}
searcher.Close();
return(results);
}
public List <string> Search(string text)
{
List <string> lstFilteredValue = new List <string>();
try
{
IndexSearcher MyIndexSearcher = new IndexSearcher(LuceneDirectory);
Query mainQuery = this.GetParsedQuerywc(text);
//Do the search
Hits hits = MyIndexSearcher.Search(mainQuery);
int results = hits.Length();
for (int i = 0; i < results; i++)
{
Document doc = hits.Doc(i);
float score = hits.Score(i);
lstFilteredValue.Add(doc.Get("Name") + "," + doc.Get("Id"));
}
}
catch (Exception GeneralException)
{
}
return(lstFilteredValue);
}
public static IDictionary <string, string> Query(string searchTerm)
{
BuildIndexTask.Wait();
System.Text.RegularExpressions.Regex rgx = new System.Text.RegularExpressions.Regex("[^a-zA-Z0-9]");
searchTerm = rgx.Replace(searchTerm, " ");
IndexSearcher searcher = new IndexSearcher(luceneIndexDirectory);
QueryParser parser = new QueryParser("Guideline", analyzer);
Query query = parser.Parse(searchTerm.ToLower());
Hits hitsFound = searcher.Search(query);
IDictionary <string, string> results = new Dictionary <string, string>();
for (int i = 0; i < hitsFound.Length(); i++)
{
Document doc = hitsFound.Doc(i);
float score = hitsFound.Score(i);
string CodeSnippetName = doc.Get("CodeSnippetName");
string CodeSnippet = doc.Get("CodeSnippet");
if (score > 0.6)
{
results.Add(CodeSnippetName, CodeSnippet);
}
}
searcher.Close();
return(results);
}
public StoryCollection Find(int hostId, int storyId)
{
int?docId = ConvertStoryIdtoDocId(hostId, storyId);
if (docId.HasValue)
{
IndexSearcher indexSearch = SearchQuery.GetSearcher(hostId);
IndexReader indexReader = indexSearch.GetIndexReader();
MoreLikeThis mlt = new MoreLikeThis(indexReader);
mlt.SetAnalyzer(new DnkAnalyzer());
//mlt.SetFieldNames(new string[] { "title", "description" });
//these values control the query used to find related/similar stories
//
//-we are only using the title and tags fields,
//-the term must appear 1 or more times,
//-the query will only have 3 terms
//-a word less than 3 char in len with be ignored
//-the term must appear at in at least 4 doc
mlt.SetFieldNames(new string[] { "title", "tags" });
mlt.SetMinTermFreq(1);
mlt.SetMaxQueryTerms(5);
mlt.SetMinWordLen(3);
mlt.SetMinDocFreq(4);
mlt.SetStopWords(StopWords());
mlt.SetBoost(true);
Query mltQuery = mlt.Like(docId.Value);
Hits hits = indexSearch.Search(mltQuery);
List <int> results = new List <int>();
for (int i = 0; i < hits.Length(); i++)
{
Document d = hits.Doc(i);
int hitStoryId = int.Parse(d.GetField("id").StringValue());
if (hitStoryId != storyId)
{
results.Add(hitStoryId);
if (results.Count == NUMBER_OF_RELATED_STORIES_TO_RETURN)
{
break;
}
}
}
return(SearchQuery.LoadStorySearchResults(results));
}
else
{
return(null);
}
}
public override CardDescription[] Read(string dirname, ProgressBar process, ToolStripStatusLabel lbl)
{
if (dirname == null || dirname.Length <= 0)
{
return(null);
}
if (!Directory.Exists(dirname))
{
return(null);
}
if (dirname[dirname.Length - 1] != '\\')
{
dirname += "\\";
}
if (File.Exists(dirname + "list.txt"))
{
string[] files = File.ReadAllLines(dirname + "list.txt", Encoding.UTF8);
foreach (string s in Directory.GetFiles(dirname))
{
string ss = s.Substring(s.LastIndexOf('\\') + 1);
bool inlist = false;
foreach (string s2 in files)
{
if (string.Equals(ss, s2, StringComparison.OrdinalIgnoreCase))
{
inlist = true;
break;
}
}
if (!(inlist || string.Equals(ss, "list.txt", StringComparison.OrdinalIgnoreCase)))
{
File.Delete(dirname + ss);
}
}
}
ArrayList cards = new ArrayList(MinCapacity);
Query query = new MatchAllDocsQuery();
Searcher searcher = new IndexSearcher(dirname);
Hits hits = searcher.Search(query);
int length = hits.Length();
for (int i = 0; i < length; i++)
{
Document doc = hits.Doc(i);
cards.Add(ParseCard(doc));
}
return((CardDescription[])cards.ToArray(typeof(CardDescription)));
}
/// <summary> Perform synonym expansion on a query.
///
/// </summary>
/// <param name="">query
/// </param>
/// <param name="">syns
/// </param>
/// <param name="">a
/// </param>
/// <param name="">field
/// </param>
/// <param name="">boost
/// </param>
public static Query Expand(System.String query, Searcher syns, Analyzer a, System.String field, float boost)
{
System.Collections.Hashtable already = new System.Collections.Hashtable(); // avoid dups
System.Collections.IList top = new System.Collections.ArrayList(); // needs to be separately listed..
// [1] Parse query into separate words so that when we expand we can avoid dups
TokenStream ts = a.TokenStream(field, new System.IO.StringReader(query));
Lucene.Net.Analysis.Token t;
while ((t = ts.Next()) != null)
{
System.String word = t.TermText();
if (already.Contains(word) == false)
{
already.Add(word, word);
top.Add(word);
}
}
BooleanQuery tmp = new BooleanQuery();
// [2] form query
System.Collections.IEnumerator it = top.GetEnumerator();
while (it.MoveNext())
{
// [2a] add to level words in
System.String word = (System.String)it.Current;
TermQuery tq = new TermQuery(new Term(field, word));
tmp.Add(tq, BooleanClause.Occur.SHOULD);
// [2b] add in unique synonums
Hits hits = syns.Search(new TermQuery(new Term(Syns2Index.F_WORD, word)));
for (int i = 0; i < hits.Length(); i++)
{
Document doc = hits.Doc(i);
System.String[] values = doc.GetValues(Syns2Index.F_SYN);
for (int j = 0; j < values.Length; j++)
{
System.String syn = values[j];
if (already.Contains(syn) == false)
{
already.Add(syn, syn);
tq = new TermQuery(new Term(field, syn));
if (boost > 0)
{
// else keep normal 1.0
tq.SetBoost(boost);
}
tmp.Add(tq, BooleanClause.Occur.SHOULD);
}
}
}
}
return(tmp);
}
protected void Page_Load(object sender, EventArgs e)
{
//if (Session["KeyWords"] == null ? false : true)
//{
// Response.Redirect("Search.aspx");
//}
String text = Session["KeyWords"].ToString();
ChineseAnalyzer analyzer = new ChineseAnalyzer();
TokenStream ts = analyzer.TokenStream("ItemName", new System.IO.StringReader(text));
Lucene.Net.Analysis.Token token;
try
{
int n = 0;
while ((token = ts.Next()) != null)
{
this.lbMsg.Text += (n++) + "->" + token.TermText() + " " + token.StartOffset() + " " + token.EndOffset() + " " + token.Type() + "<br>";
// Response.Write((n++) + "->" + token.TermText() + " " + token.StartOffset() + " "
//+ token.EndOffset() + " " + token.Type() + "<br>");
}
}
catch
{
this.lbMsg.Text = "wrong";
}
// Analyzer analyzer = new StandardAnalyzer();
Directory directory = FSDirectory.GetDirectory(Server.MapPath("/indexFile/"), false);
IndexSearcher isearcher = new IndexSearcher(directory);
Query query;
query = QueryParser.Parse(Session["KeyWords"].ToString(), "ItemName", analyzer);
//query = QueryParser.Parse("2", "nid", analyzer);
Hits hits = isearcher.Search(query);
this.lbMsg.Text += "<font color=red>共找到" + hits.Length() + "条记录</font><br>";
//Response.Write("<font color=red>共找到" + hits.Length() + "条记录</font><br>");
for (int i = 0; i < hits.Length(); i++)
{
Document hitDoc = hits.Doc(i);
this.lbMsg.Text += "编号：" + hitDoc.Get("ItemID").ToString() + "<br>"
+ "分类：" + hitDoc.Get("CategoryName").ToString() + "<br>"
+ "专题：" + hitDoc.Get("ProductName").ToString() + "<br>"
+ "标题：<a href=" + hitDoc.Get("visiturl").ToString() + ">" + hitDoc.Get("ItemName").ToString() + "</a><br>";
//Response.Write("编号：" + hitDoc.Get("ItemID").ToString() + "<br>");
//Response.Write("分类：" + hitDoc.Get("CategoryName").ToString() + "<br>");
//Response.Write("标题：<a href=" + hitDoc.Get("visiturl").ToString() + ">" + hitDoc.Get("ItemName").ToString() + "</a><br>");
//Response.Write("专题：" + hitDoc.Get("ProductName").ToString() + "<br>");
}
isearcher.Close();
directory.Close();
}
private void SearchFor(int n, Searcher searcher)
{
System.Console.Out.WriteLine("Searching for " + n);
Hits hits = searcher.Search(QueryParsers.QueryParser.Parse(English.IntToEnglish(n), "contents", Lucene.Net.ThreadSafetyTest.ANALYZER));
System.Console.Out.WriteLine("Search for " + n + ": total=" + hits.Length());
for (int j = 0; j < System.Math.Min(3, hits.Length()); j++)
{
System.Console.Out.WriteLine("Hit for " + n + ": " + hits.Doc(j).Get("id"));
}
}
public string[] GetFragments(Query query, Hits hits)
{
var highlighter = CreateHighlighter(query);
var result = new string[hits.Length()];
for (var i = 0; i < result.Length; i++)
{
var size = PackedSizeConverter.FromSortableString(hits.Doc(i).Get(FieldName.Size));
var loc = hits.Doc(i).Get(FieldName.Id).Split('@');
var info = Svn.GetPathInfo(loc[0], Convert.ToInt32(loc[1]));
if (info.IsDirectory)
{
continue;
}
var text = Svn.GetPathContent(loc[0], Convert.ToInt32(loc[1]), size);
var tokens = new SimpleTokenStream(text);
result[i] = GetFragments(highlighter, tokens, text);
}
return(result);
}
private void PrintHits(System.IO.StringWriter out_Renamed, Hits hits)
{
out_Renamed.WriteLine(hits.Length() + " total results\n");
for (int i = 0; i < hits.Length(); i++)
{
if (i < 10 || (i > 94 && i < 105))
{
Document d = hits.Doc(i);
out_Renamed.WriteLine(i + " " + d.Get(ID_FIELD));
}
}
}
private void CheckHits(Hits hits, int expectedCount)
{
Assert.AreEqual(expectedCount, hits.Length(), "total results");
for (int i = 0; i < hits.Length(); i++)
{
if (i < 10 || (i > 94 && i < 105))
{
Document d = hits.Doc(i);
Assert.AreEqual(System.Convert.ToString(i), d.Get(ID_FIELD), "check " + i);
}
}
}
/////////////////////////////////////////////////////////
static void ExecuteRemove(string arg)
{
LuceneDriver driver = new LuceneDriver(index_dir);
if (arg.IndexOf("://") != -1)
{
Uri uri = new Uri(arg);
ICollection hits = driver.DoQueryByUri(uri);
if (hits == null || hits.Count == 0)
{
Console.WriteLine("Uri not found in the index: {0}", uri);
Environment.Exit(1);
}
driver.Remove(uri);
driver.Flush();
Console.WriteLine("Successfully removed Uri: {0}", uri);
}
else
{
IndexSearcher searcher = new IndexSearcher(driver.Store);
BooleanQuery query = new BooleanQuery();
Term term = new Term("prop:k:Tag", arg); // Argh
TermQuery term_query = new TermQuery(term);
query.Add(term_query, false, false);
Hits hits = searcher.Search(query);
int n_hits = hits.Length();
string uri;
for (int i = 0; i < n_hits; ++i)
{
Document doc = hits.Doc(i);
uri = doc.Get("Uri");
if (uri == null)
{
continue;
}
driver.Remove(UriFu.UriStringToUri(uri));
}
driver.Flush();
Console.WriteLine("Successfully removed {0} items with tag: {1}", n_hits, arg);
}
}