public virtual Query Build(IQueryNode queryNode)
{
MultiPhraseQueryNode phraseNode = (MultiPhraseQueryNode)queryNode;
MultiPhraseQuery phraseQuery = new MultiPhraseQuery();
IList <IQueryNode> children = phraseNode.GetChildren();
if (children != null)
{
IDictionary <int, JCG.List <Term> > positionTermMap = new JCG.SortedDictionary <int, JCG.List <Term> >();
foreach (IQueryNode child in children)
{
FieldQueryNode termNode = (FieldQueryNode)child;
TermQuery termQuery = (TermQuery)termNode
.GetTag(QueryTreeBuilder.QUERY_TREE_BUILDER_TAGID);
if (!positionTermMap.TryGetValue(termNode.PositionIncrement, out JCG.List <Term> termList) || termList is null)
{
termList = new JCG.List <Term>();
positionTermMap[termNode.PositionIncrement] = termList;
}
termList.Add(termQuery.Term);
}
foreach (int positionIncrement in positionTermMap.Keys)
{
JCG.List <Term> termList = positionTermMap[positionIncrement];
phraseQuery.Add(termList.ToArray(/*new Term[termList.size()]*/),
positionIncrement);
}
}
return(phraseQuery);
}
public MultiPhraseWeight(MultiPhraseQuery outerInstance, IndexSearcher searcher)
{
this.outerInstance = outerInstance;
this.similarity = searcher.Similarity;
IndexReaderContext context = searcher.TopReaderContext;
// compute idf
var allTermStats = new JCG.List <TermStatistics>();
foreach (Term[] terms in outerInstance.termArrays)
{
foreach (Term term in terms)
{
if (!termContexts.TryGetValue(term, out TermContext termContext) || termContext is null)
{
termContext = TermContext.Build(context, term);
termContexts[term] = termContext;
}
allTermStats.Add(searcher.TermStatistics(term, termContext));
}
}
stats = similarity.ComputeWeight(outerInstance.Boost, searcher.CollectionStatistics(outerInstance.field), allTermStats.ToArray());
}
private static List <string> searchOneWord_Lucene(string query, IndexSearcher searcher)
{ //Ищем по одному слову
List <string> result_lucene = new List <string>();
var array = query.Split(' ');
var phrase = new MultiPhraseQuery();
foreach (var word in array)
{
phrase = new MultiPhraseQuery();
if (!String.IsNullOrEmpty(word))
{
phrase.Add(new Term("name_word", word));
var res = searcher.Search(phrase, 10).ScoreDocs;
foreach (var hit in res)
{
var foundDoc = searcher.Doc(hit.Doc);
result_lucene.Add(getFoundDocResult(foundDoc));
}
}
}
return(result_lucene);
}
public virtual void TestMultiPhraseQueryParsing()
{
TokenAndPos[] INCR_0_QUERY_TOKENS_AND = new TokenAndPos[]
{
new TokenAndPos("a", 0),
new TokenAndPos("1", 0),
new TokenAndPos("b", 1),
new TokenAndPos("1", 1),
new TokenAndPos("c", 2)
};
QueryParser qp = new QueryParser(TEST_VERSION_CURRENT, "field", new CannedAnalyzer(INCR_0_QUERY_TOKENS_AND));
Query q = qp.Parse("\"this text is acually ignored\"");
assertTrue("wrong query type!", q is MultiPhraseQuery);
MultiPhraseQuery multiPhraseQuery = new MultiPhraseQuery();
multiPhraseQuery.Add(new Term[] { new Term("field", "a"), new Term("field", "1") }, -1);
multiPhraseQuery.Add(new Term[] { new Term("field", "b"), new Term("field", "1") }, 0);
multiPhraseQuery.Add(new Term[] { new Term("field", "c") }, 1);
assertEquals(multiPhraseQuery, q);
}
private void SearchButton_Click(object sender, EventArgs e)
{
//Variables and pretty stuff
int counter = 0;
Cursor.Current = Cursors.WaitCursor;
SearchButton.Enabled = false;
ResultBox.Items.Clear();
var query = TextSearch.Text;
var array = query.Split(' ').ToList();
List <string> res_list = new List <string>();
//Some sort of error handling
try
{
if (!luceneCheck.Checked)
{
using (var conn = new NpgsqlConnection(connString))
{
conn.Open();
var statement = "";
//Поиск по точному названию
statement = "SELECT * " +
"FROM movies " +
"WHERE name = \'" + query + "\'";
var command = new NpgsqlCommand(statement, conn);
var id = 0;
var year = 0;
var name = "";
using (var reader = command.ExecuteReader())
{
while (reader.Read() && counter < 10)
{
id = reader.GetInt32(0);
year = reader.GetInt32(1);
name = reader.GetString(2);
counter += 1;
res_list.Add("ID: " + id.ToString() + " YEAR: " + year.ToString() + " NAME: " + name);
}
}
//Поиск по году и по названию //, если предыдущий ничего не дал
//if (ResultBox.Items.Count == 0)
//Ищем год в запросе
string year_to_find = "";
int number = 0;
foreach (var word in array)
{
bool result = Int32.TryParse(word, out number);
if (result && number > 1800 && number <= 9999)
{
year_to_find = word;
array.RemoveAt(array.IndexOf(word));
break;
}
else
{
number = 0;
}
}
//Если нашли
if (number != 0)
{
foreach (var word in array)
{
if (!String.IsNullOrEmpty(word))
{
statement = "SELECT * " +
"FROM movies " +
"WHERE year = " + year_to_find + " AND name ILIKE \'%" + word + "%\' ";
command = new NpgsqlCommand(statement, conn);
using (var reader = command.ExecuteReader())
{
while (reader.Read() && counter < 10)
{
counter += 1;
id = reader.GetInt32(0);
year = reader.GetInt32(1);
name = reader.GetString(2);
res_list.Add("ID: " + id.ToString() + " YEAR: " + year.ToString() + " NAME: " + name);
}
}
}
}
}
//Поиск по слову в названии //, если предыдущие ничего не дали
//if (ResultBox.Items.Count == 0)
foreach (var word in array)
{
if (!String.IsNullOrEmpty(word))
{
statement = "SELECT * " +
"FROM movies " +
"WHERE name ILIKE \'" + word + " %\' " +
"OR name = \'" + word + "\' " +
"OR name ILIKE \'% " + word + "\'";
command = new NpgsqlCommand(statement, conn);
using (var reader = command.ExecuteReader())
{
while (reader.Read() && counter < 10)
{
counter += 1;
id = reader.GetInt32(0);
year = reader.GetInt32(1);
name = reader.GetString(2);
res_list.Add("ID: " + id.ToString() + " YEAR: " + year.ToString() + " NAME: " + name);
}
}
}
}
//Поиск по части слова в названии. Потому что надо найти хоть что-то
//if (ResultBox.Items.Count == 0)
foreach (var word in array)
{
if (!String.IsNullOrEmpty(word))
{
statement = "SELECT * " +
"FROM movies " +
"WHERE name ILIKE \'%" + word + "%\' ";
command = new NpgsqlCommand(statement, conn);
using (var reader = command.ExecuteReader())
{
while (reader.Read() && counter < 10)
{
counter += 1;
id = reader.GetInt32(0);
year = reader.GetInt32(1);
name = reader.GetString(2);
res_list.Add("ID: " + id.ToString() + " YEAR: " + year.ToString() + " NAME: " + name);
}
}
}
}
//Дубли не хотим
res_list = res_list.Select(x => x).Distinct().ToList();
ResultBox.Items.Clear();
foreach (var item in res_list)
{
ResultBox.Items.Add(item);
}
conn.Close();
}
}
else
{
//Ищем по одному слову
QueryParser parser = new QueryParser(AppLuceneVersion, "name", analyzer);
var phrase = new MultiPhraseQuery();
foreach (var word in array)
{
var q = parser.Parse(query);
if (!String.IsNullOrEmpty(word))
{
var res = searcher.Search(q, 10).ScoreDocs;
foreach (var hit in res)
{
var foundDoc = searcher.Doc(hit.Doc);
var score = hit.Score;
res_list.Add("Score: " + score + " ID: " + foundDoc.GetField("id").GetInt32Value().ToString() +
" YEAR: " + foundDoc.GetField("year").GetInt32Value().ToString() + " NAME: " + foundDoc.GetValues("name")[0]);
}
}
}
//Ищем полное название
phrase.Add(new Term("name", query));
var hits = searcher.Search(phrase, 10).ScoreDocs;
foreach (var hit in hits)
{
var foundDoc = searcher.Doc(hit.Doc);
var score = hit.Score;
res_list.Add("Score: " + score + " ID: " + foundDoc.GetField("id").GetInt32Value().ToString() +
" YEAR: " + foundDoc.GetField("year").GetInt32Value().ToString() + " NAME: " + foundDoc.GetValues("name")[0]);
}
//Ищем части слов
foreach (var word in array)
{
if (!String.IsNullOrEmpty(word))
{
var wild = new WildcardQuery(new Term("name", word));
var res = searcher.Search(wild, 10).ScoreDocs;
foreach (var hit in res)
{
var foundDoc = searcher.Doc(hit.Doc);
var score = hit.Score;
res_list.Add("Score: " + score + " ID: " + foundDoc.GetField("id").GetInt32Value().ToString() +
" YEAR: " + foundDoc.GetField("year").GetInt32Value().ToString() + " NAME: " + foundDoc.GetValues("name")[0]);
}
}
}
//Ищем год и часть слова
string year_to_find = "";
int number = 0;
foreach (var word in array)
{
bool result = Int32.TryParse(word, out number);
if (result && number > 1800 && number <= 9999)
{
year_to_find = word;
array.RemoveAt(array.IndexOf(word));
break;
}
else
{
number = 0;
}
}
//Если нашли
if (number != 0)
{
phrase = new MultiPhraseQuery();
foreach (var word in array)
{
if (!String.IsNullOrEmpty(word))
{
BooleanQuery booleanQuery = new BooleanQuery();
var wild = new WildcardQuery(new Term("name", word));
var num = NumericRangeQuery.NewInt32Range("year", 1, number, number, true, true);
booleanQuery.Add(wild, Occur.MUST);
booleanQuery.Add(num, Occur.MUST);
var res = searcher.Search(booleanQuery, 10).ScoreDocs;
foreach (var hit in res)
{
var foundDoc = searcher.Doc(hit.Doc);
var score = hit.Score;
res_list.Add("Score: " + score + " ID: " + foundDoc.GetField("id").GetInt32Value().ToString() +
" YEAR: " + foundDoc.GetField("year").GetInt32Value().ToString() + " NAME: " + foundDoc.GetValues("name")[0]);
}
}
}
}
}
//Не хотим дубли
res_list = res_list.Select(x => x).Distinct().ToList();
ResultBox.Items.Clear();
foreach (var item in res_list)
{
ResultBox.Items.Add(item);
}
//Ну и если всё плохо
if (ResultBox.Items.Count == 0)
{
ResultBox.Items.Add("Нет результатов. Попробуйте расширить поисковый запрос");
}
}
catch (Exception ex)
{
MessageBox.Show("Error occured while searching: " + ex.Message, "Error", MessageBoxButtons.OK, MessageBoxIcon.Error);
}
Cursor.Current = Cursors.Default;
SearchButton.Enabled = true;
}
public virtual Query VisitMultiPhraseQuery(MultiPhraseQuery multiPhraseq)
{
throw new SnNotSupportedException();
}
public static Hashtable Search(bool api)
{
BillType type;
int number;
int session = -1;
if (HttpContext.Current.Request["session"] != null && HttpContext.Current.Request["session"] != "")
{
session = int.Parse(HttpContext.Current.Request["session"]);
}
string q = HttpContext.Current.Request["q"];
int start = 0, count = (!api ? 30 : 1000);
if (HttpContext.Current.Request["start"] != null)
{
start = int.Parse(HttpContext.Current.Request["start"]);
}
if (HttpContext.Current.Request["count"] != null)
{
count = int.Parse(HttpContext.Current.Request["count"]);
}
BooleanQuery query = new BooleanQuery();
Hashtable no_results = new Hashtable();
no_results["count"] = 0;
if (q != null && q.IndexOf("*") > -1)
{
return(no_results);
}
if (!api && session == -1 && q != null)
{
int slash = q.IndexOf('/');
if (slash >= q.Length - 4 && slash > 2)
{
try {
session = int.Parse(q.Substring(slash + 1)); // and if that worked...
q = q.Substring(0, slash);
HttpContext.Current.Response.Redirect("billsearch.xpd?session=" + session + "&q=" + HttpUtility.UrlEncode(q));
return(null);
} catch { }
}
}
if (session == -1)
{
session = Util.CurrentSession;
}
//Console.Error.WriteLine("Find: " + q);
string search_method = "search";
ArrayList specs = new ArrayList();
Hashtable scores = new Hashtable();
// Match a bill number exactly
if (q != null && Bills.ParseID(q, out type, out number))
{
if (!api)
{
// Redirect the user right to the bill page.
// Don't even check if bill exists.
HttpContext.Current.Response.Redirect(
Bills.BillLink2(session, type, number));
return(null);
}
else
{
search_method = "search by bill number";
scores[session + EnumsConv.BillTypeToString(type) + number] = 1.0F;
specs.Add(new Database.AndSpec(
new Database.SpecEQ("session", session),
new Database.SpecEQ("type", EnumsConv.BillTypeToString(type)),
new Database.SpecEQ("number", number)));
}
}
// Match public law number exactly
if (!api && q != null && (q.StartsWith("P.L.") || q.StartsWith("PL")))
{
try {
string num = null;
if (q.StartsWith("P.L."))
{
num = q.Substring(4);
}
if (q.StartsWith("PL"))
{
num = q.Substring(2);
}
num = num.Replace(" ", "");
int dash = num.IndexOf('-');
int s = int.Parse(num.Substring(0, dash));
TableRow bill = Util.Database.DBSelectFirst("billindex", "session, type, number",
new Database.SpecEQ("idx", "publiclawnumber"),
new Database.SpecEQ("session", s),
new Database.SpecEQ("value", num));
if (bill != null)
{
if (!api)
{
HttpContext.Current.Response.Redirect(Bills.BillLink3((int)bill["session"], (string)bill["type"], (int)bill["number"]));
return(null);
}
else
{
search_method = "search by public law number";
scores[(int)bill["session"] + (string)bill["type"] + (int)bill["number"]] = 1.0F;
specs.Add(new Database.AndSpec(
new Database.SpecEQ("session", (int)bill["session"]),
new Database.SpecEQ("type", (string)bill["type"]),
new Database.SpecEQ("number", (int)bill["number"])));
}
}
} catch {
}
}
if (session == -1)
{
session = Util.CurrentSession;
}
// Match USC reference
Regex uscexp = new Regex(@"(\d[0-9A-Za-z\-]*)\s+U\.?S\.?C\.?\s+(\d[0-9A-Za-z\-]*)((\s*\([^\) ]+\))*)",
RegexOptions.IgnoreCase);
Match uscmc = (q == null ? null : uscexp.Match(q));
if (uscmc != null && uscmc.Success)
{
string title = uscmc.Groups[1].Value;
string section = uscmc.Groups[2].Value;
string paragraph = uscmc.Groups[3].Value;
string[] ps = paragraph.Split('[', '(', ')', ' ');
int psi = 0; while (psi < ps.Length - 1 && ps[psi] == "")
{
psi++;
}
int pse = ps.Length - 1; while (pse > 0 && ps[pse] == "")
{
pse--;
}
if (ps.Length != 0)
{
paragraph = "_" + String.Join("_", ps, psi, pse - psi + 1);
}
Table table = Util.Database.DBSelect("billusc", "session, type, number",
new Database.SpecEQ("session", session),
new Database.OrSpec(
new Database.SpecEQ("ref", title + "_" + section + paragraph),
new Database.SpecStartsWith("ref", title + "_" + section + paragraph + "_")));
foreach (TableRow bill in table)
{
search_method = "search by U.S.C. section";
scores[(int)bill["session"] + (string)bill["type"] + (int)bill["number"]] = 1.0F;
specs.Add(new Database.AndSpec(
new Database.SpecEQ("session", (int)bill["session"]),
new Database.SpecEQ("type", (string)bill["type"]),
new Database.SpecEQ("number", (int)bill["number"])));
}
}
int total_count = -1;
if (specs.Count == 0)
{
if (q != null && q.Trim() != "")
{
BooleanQuery query1 = new BooleanQuery();
query.Add(query1, BooleanClause.Occur.MUST);
try {
/*if (!q.StartsWith("-")) {
* PhraseQuery pq = new PhraseQuery();
* pq.Add( new Term("shorttitles", q) );
* pq.SetBoost((float)4);
* query1.Add(pq, false, false);
* }*/
Query query_titles2 = new QueryParser("shorttitles", new StandardAnalyzer()).Parse(q);
query_titles2.SetBoost((float)3);
query1.Add(query_titles2, BooleanClause.Occur.SHOULD);
Query query_titles1 = new QueryParser("officialtitles", new StandardAnalyzer()).Parse(q);
query_titles1.SetBoost((float)2);
query1.Add(query_titles1, BooleanClause.Occur.SHOULD);
Query query_summary = new QueryParser("summary", new StandardAnalyzer()).Parse(q);
query1.Add(query_summary, BooleanClause.Occur.SHOULD);
Query query_text = new QueryParser("fulltext", new StandardAnalyzer()).Parse(q);
query1.Add(query_text, BooleanClause.Occur.SHOULD);
} catch (Exception e) {
return(no_results);
}
}
string chamber = HttpContext.Current.Request["chamber"];
string[] status = HttpContext.Current.Request["status"] == null ? null : HttpContext.Current.Request["status"].Split(',');
string sponsor = HttpContext.Current.Request["sponsor"];
string cosponsor = HttpContext.Current.Request["cosponsor"];
if (chamber != null && (chamber == "s" || chamber == "h"))
{
query.Add(new WildcardQuery(new Term("type", chamber + "*")), BooleanClause.Occur.MUST);
}
if (status != null && status[0] != "")
{
List <Term> terms = new List <Term>();
foreach (string s in status)
{
terms.Add(new Term("state", s));
}
MultiPhraseQuery mpq = new MultiPhraseQuery();
mpq.Add(terms.ToArray());
query.Add(mpq, BooleanClause.Occur.MUST);
}
if (sponsor != null && sponsor != "")
{
query.Add(new TermQuery(new Term("sponsor", sponsor)), BooleanClause.Occur.MUST);
}
if (cosponsor != null && cosponsor != "")
{
query.Add(new TermQuery(new Term("cosponsor", cosponsor)), BooleanClause.Occur.MUST);
}
IndexSearcher searcher = new IndexSearcher(Util.DataPath + Path.DirectorySeparatorChar + session + Path.DirectorySeparatorChar + "index.bills.lucene");
Sort sort = null;
if (HttpContext.Current.Request["sort"] != null && HttpContext.Current.Request["sort"] == "introduced")
{
sort = new Sort(new SortField("introduced", SortField.STRING, true));
}
if (HttpContext.Current.Request["sort"] != null && HttpContext.Current.Request["sort"] == "lastaction")
{
sort = new Sort(new SortField("lastaction", SortField.STRING, true));
}
Hits hits = searcher.Search(query, sort == null ? new Sort() : sort);
int end = hits.Length();
if (start + count < end)
{
end = start + count;
}
total_count = hits.Length();
for (int i = start; i < end; i++)
{
Document doc = hits.Doc(i);
string billsession = doc.Get("session");
string billtype = doc.Get("type");
string billnumber = doc.Get("number");
int istatus = (int)EnumsConv.BillStatusFromString(doc.Get("status"));
float score;
if (sort == null) // readjust the score based on status
{
score = hits.Score(i) + istatus / (float)8 * (float).2;
}
else // keep order from Lucene
{
score = -i;
}
scores[billsession + billtype + billnumber] = score;
specs.Add(new Database.AndSpec(
new Database.SpecEQ("session", billsession),
new Database.SpecEQ("type", billtype),
new Database.SpecEQ("number", billnumber)));
}
if (HttpContext.Current.Request["sort"] != null && HttpContext.Current.Request["sort"] == "hits" && specs.Count > 0)
{
Table hitsinfo = Util.Database.DBSelect("billhits", "*",
Database.OrSpec.New((Database.Spec[])specs.ToArray(typeof(Database.Spec))));
foreach (TableRow billhits in hitsinfo)
{
scores["" + billhits["session"] + billhits["type"] + billhits["number"]] = (float)(int)billhits["hits1"];
}
}
}
if (specs.Count == 0)
{
return(no_results);
}
Table billinfo = Util.Database.DBSelect("billstatus", "*",
Database.OrSpec.New((Database.Spec[])specs.ToArray(typeof(Database.Spec))));
if (total_count == -1)
{
total_count = billinfo.Rows;
}
ArrayList ret = new ArrayList();
foreach (TableRow r in billinfo)
{
ret.Add(r);
}
BillHitComparer bhc = new BillHitComparer();
bhc.scores = scores;
ret.Sort(bhc);
Hashtable ret2 = new Hashtable();
ret2["count"] = total_count;
ret2["method"] = search_method;
ret2["results"] = ret;
return(ret2);
}
/// <summary>
/// Fills a <c>Map</c> with <see cref="WeightedSpanTerm"/>s using the terms from the supplied <c>Query</c>.
/// </summary>
/// <param name="query">Query to extract Terms from</param>
/// <param name="terms">Map to place created WeightedSpanTerms in</param>
private void Extract(Query query, IDictionary <String, WeightedSpanTerm> terms)
{
if (query is BooleanQuery)
{
BooleanClause[] queryClauses = ((BooleanQuery)query).GetClauses();
for (int i = 0; i < queryClauses.Length; i++)
{
if (!queryClauses[i].IsProhibited)
{
Extract(queryClauses[i].Query, terms);
}
}
}
else if (query is PhraseQuery)
{
PhraseQuery phraseQuery = ((PhraseQuery)query);
Term[] phraseQueryTerms = phraseQuery.GetTerms();
SpanQuery[] clauses = new SpanQuery[phraseQueryTerms.Length];
for (int i = 0; i < phraseQueryTerms.Length; i++)
{
clauses[i] = new SpanTermQuery(phraseQueryTerms[i]);
}
int slop = phraseQuery.Slop;
int[] positions = phraseQuery.GetPositions();
// add largest position increment to slop
if (positions.Length > 0)
{
int lastPos = positions[0];
int largestInc = 0;
int sz = positions.Length;
for (int i = 1; i < sz; i++)
{
int pos = positions[i];
int inc = pos - lastPos;
if (inc > largestInc)
{
largestInc = inc;
}
lastPos = pos;
}
if (largestInc > 1)
{
slop += largestInc;
}
}
bool inorder = slop == 0;
SpanNearQuery sp = new SpanNearQuery(clauses, slop, inorder);
sp.Boost = query.Boost;
ExtractWeightedSpanTerms(terms, sp);
}
else if (query is TermQuery)
{
ExtractWeightedTerms(terms, query);
}
else if (query is SpanQuery)
{
ExtractWeightedSpanTerms(terms, (SpanQuery)query);
}
else if (query is FilteredQuery)
{
Extract(((FilteredQuery)query).Query, terms);
}
else if (query is DisjunctionMaxQuery)
{
foreach (var q in ((DisjunctionMaxQuery)query))
{
Extract(q, terms);
}
}
else if (query is MultiTermQuery && expandMultiTermQuery)
{
MultiTermQuery mtq = ((MultiTermQuery)query);
if (mtq.RewriteMethod != MultiTermQuery.SCORING_BOOLEAN_QUERY_REWRITE)
{
mtq = (MultiTermQuery)mtq.Clone();
mtq.RewriteMethod = MultiTermQuery.SCORING_BOOLEAN_QUERY_REWRITE;
query = mtq;
}
FakeReader fReader = new FakeReader();
MultiTermQuery.SCORING_BOOLEAN_QUERY_REWRITE.Rewrite(fReader, mtq);
if (fReader.Field != null)
{
IndexReader ir = GetReaderForField(fReader.Field);
Extract(query.Rewrite(ir), terms);
}
}
else if (query is MultiPhraseQuery)
{
MultiPhraseQuery mpq = (MultiPhraseQuery)query;
IList <Term[]> termArrays = mpq.GetTermArrays();
int[] positions = mpq.GetPositions();
if (positions.Length > 0)
{
int maxPosition = positions[positions.Length - 1];
for (int i = 0; i < positions.Length - 1; ++i)
{
if (positions[i] > maxPosition)
{
maxPosition = positions[i];
}
}
var disjunctLists = new List <SpanQuery> [maxPosition + 1];
int distinctPositions = 0;
for (int i = 0; i < termArrays.Count; ++i)
{
Term[] termArray = termArrays[i];
List <SpanQuery> disjuncts = disjunctLists[positions[i]];
if (disjuncts == null)
{
disjuncts = (disjunctLists[positions[i]] = new List <SpanQuery>(termArray.Length));
++distinctPositions;
}
for (int j = 0; j < termArray.Length; ++j)
{
disjuncts.Add(new SpanTermQuery(termArray[j]));
}
}
int positionGaps = 0;
int position = 0;
SpanQuery[] clauses = new SpanQuery[distinctPositions];
for (int i = 0; i < disjunctLists.Length; ++i)
{
List <SpanQuery> disjuncts = disjunctLists[i];
if (disjuncts != null)
{
clauses[position++] = new SpanOrQuery(disjuncts.ToArray());
}
else
{
++positionGaps;
}
}
int slop = mpq.Slop;
bool inorder = (slop == 0);
SpanNearQuery sp = new SpanNearQuery(clauses, slop + positionGaps, inorder);
sp.Boost = query.Boost;
ExtractWeightedSpanTerms(terms, sp);
}
}
}
private void QueryToSpanQuery(Query query, ICollection <byte[]> payloads)
{
if (query is BooleanQuery)
{
BooleanClause[] queryClauses = ((BooleanQuery)query).GetClauses();
for (int i = 0; i < queryClauses.Length; i++)
{
if (!queryClauses[i].IsProhibited)
{
QueryToSpanQuery(queryClauses[i].Query, payloads);
}
}
}
else if (query is PhraseQuery)
{
Term[] phraseQueryTerms = ((PhraseQuery)query).GetTerms();
SpanQuery[] clauses = new SpanQuery[phraseQueryTerms.Length];
for (int i = 0; i < phraseQueryTerms.Length; i++)
{
clauses[i] = new SpanTermQuery(phraseQueryTerms[i]);
}
int slop = ((PhraseQuery)query).Slop;
bool inorder = false;
if (slop == 0)
{
inorder = true;
}
SpanNearQuery sp = new SpanNearQuery(clauses, slop, inorder);
sp.Boost = query.Boost;
GetPayloads(payloads, sp);
}
else if (query is TermQuery)
{
SpanTermQuery stq = new SpanTermQuery(((TermQuery)query).Term);
stq.Boost = query.Boost;
GetPayloads(payloads, stq);
}
else if (query is SpanQuery)
{
GetPayloads(payloads, (SpanQuery)query);
}
else if (query is FilteredQuery)
{
QueryToSpanQuery(((FilteredQuery)query).Query, payloads);
}
else if (query is DisjunctionMaxQuery)
{
for (IEnumerator <Query> iterator = ((DisjunctionMaxQuery)query).GetEnumerator(); iterator.MoveNext();)
{
QueryToSpanQuery(iterator.Current, payloads);
}
}
else if (query is MultiPhraseQuery)
{
MultiPhraseQuery mpq = (MultiPhraseQuery)query;
System.Collections.Generic.IList <Term[]> termArrays = mpq.GetTermArrays();
int[] positions = mpq.GetPositions();
if (positions.Length > 0)
{
int maxPosition = positions[positions.Length - 1];
for (int i = 0; i < positions.Length - 1; ++i)
{
if (positions[i] > maxPosition)
{
maxPosition = positions[i];
}
}
IList <Query>[] disjunctLists = new IList <Query> [maxPosition + 1];
int distinctPositions = 0;
for (int i = 0; i < termArrays.Count; ++i)
{
Term[] termArray = termArrays[i];
IList <Query> disjuncts = disjunctLists[positions[i]];
if (disjuncts == null)
{
disjuncts = (disjunctLists[positions[i]] = new List <Query>(termArray.Length));
++distinctPositions;
}
foreach (Term term in termArray)
{
disjuncts.Add(new SpanTermQuery(term));
}
}
int positionGaps = 0;
int position = 0;
SpanQuery[] clauses = new SpanQuery[distinctPositions];
for (int i = 0; i < disjunctLists.Length; ++i)
{
IList <Query> disjuncts = disjunctLists[i];
if (disjuncts != null)
{
clauses[position++] = new SpanOrQuery((SpanQuery[])(disjuncts.ToArray()));
}
else
{
++positionGaps;
}
}
int slop = mpq.Slop;
bool inorder = (slop == 0);
SpanNearQuery sp = new SpanNearQuery(clauses, slop + positionGaps, inorder);
sp.Boost = query.Boost;
GetPayloads(payloads, sp);
}
}
}
static void Main(string[] args)
{
var jsonProps = "[{\"name\":\"Ondrej\",\"surname\":\"Kubicek\",\"app_data\":[\"112233\", \"56612\"]},"
+ "{\"name\":\"Lukas\",\"surname\":\"Bily\",\"app_data\":[\"12355\", \"112233\", \"89466\"]}," +
"{\"name\":\"Lenak\",\"surname\":\"Nejaka\",\"app_data\":[\"89700\"]}]";
var version = LuceneVersion.LUCENE_48;
var dir = new RAMDirectory();
var analyzer = new StandardAnalyzer(version);
var indexConfig = new IndexWriterConfig(version, analyzer);
var writer = new IndexWriter(dir, indexConfig);
var d = JsonDocument.Parse(jsonProps);
var root = d.RootElement;
foreach (var line in root.EnumerateArray())
{
var doc = new Document();
doc.Add(new StringField("name", line.GetProperty("name").GetString(), Field.Store.NO));
doc.Add(new StringField("surname", line.GetProperty("surname").GetString(), Field.Store.NO));
foreach (var f in line.GetProperty("app_data").EnumerateArray())
{
doc.Add(new StringField("app_data", f.GetString(), Field.Store.NO));
}
doc.Add(new StringField("payload", line.ToString(), Field.Store.YES));
writer.AddDocument(doc);
// Console.WriteLine(line.GetProperty("name"));
// if (line.GetProperty("app_data").EnumerateArray().Any(x => x.GetString() == "1"))
// {
// }
// foreach(var data in line.GetProperty("app_data").EnumerateArray())
// {
// }
// Console.WriteLine(line.GetProperty("app_data").GetArrayLength());
}
writer.Flush(false, false);
var searcher = new IndexSearcher(writer.GetReader(true));
var query = new MultiPhraseQuery();
query.Add(new Term("app_data", "12355"));
var booleanQuery = new BooleanQuery();
booleanQuery.Add(new TermQuery(new Term("app_data", "12355")), Occur.SHOULD);
booleanQuery.Add(new TermQuery(new Term("app_data", "89700")), Occur.SHOULD);
var res = searcher.Search(booleanQuery, 100);
Console.WriteLine(res.TotalHits);
foreach (var hit in res.ScoreDocs)
{
var item = searcher.Doc(hit.Doc);
Console.WriteLine(item.Get("payload"));
}
}
private void GetDataIndexId(DirectoryInfo directoryInfo, ref string textSearcher, ref List <string> lReturn)
{
using (Directory directory = FSDirectory.Open(directoryInfo))
using (Analyzer analyzer = new StandardAnalyzer(Version.LUCENE_30))
using (IndexReader indexReader = IndexReader.Open(directory, true))
using (Searcher indexSearcher = new IndexSearcher(indexReader))
{
TopScoreDocCollector collectorMultiPhraseQuery = TopScoreDocCollector.Create(100, true);
TopScoreDocCollector collectorQueryParser = TopScoreDocCollector.Create(100, true);
int docId = 0;
string tempObjectId = string.Empty;
List <string> listTemp = new List <string>();
char[] delimiterChars = { ' ', ',', '.', ':', '\t' };
MultiPhraseQuery multiPhraseQuery = new MultiPhraseQuery();
//Here implement the search lines for graphs at level 3
multiPhraseQuery.Slop = 3;
foreach (var word in textSearcher.Split(delimiterChars))
{
multiPhraseQuery.Add(new Term("FullRequest", word));
}
indexSearcher.Search(multiPhraseQuery, collectorMultiPhraseQuery);
ScoreDoc[] listResultPharseQuery = collectorMultiPhraseQuery.TopDocs().ScoreDocs;
foreach (var itemPharseQuery in listResultPharseQuery)
{
docId = itemPharseQuery.Doc;
Document docPharseQuery = indexSearcher.Doc(docId);
tempObjectId = docPharseQuery.Get("ObjetcId");
if (!string.IsNullOrEmpty(tempObjectId))
{
listTemp.Add(tempObjectId);
}
}
//This lineas implement QueryPArse
docId = 0;
tempObjectId = string.Empty;
var queryParser = new QueryParser(Version.LUCENE_30, "FullRequest", analyzer);
var query = queryParser.Parse(textSearcher);
indexSearcher.Search(query, collectorQueryParser);
ScoreDoc[] listResultquery = collectorQueryParser.TopDocs().ScoreDocs;
foreach (var itemQuery in listResultquery)
{
docId = itemQuery.Doc;
Document docQuery = indexSearcher.Doc(docId);
tempObjectId = docQuery.Get("ObjetcId");
if (!string.IsNullOrEmpty(tempObjectId))
{
listTemp.Add(tempObjectId);
}
}
lReturn.AddRange(listTemp.Distinct().ToList());
}
}
/// <summary>
/// Creates a query from the analysis chain.
/// <para/>
/// Expert: this is more useful for subclasses such as queryparsers.
/// If using this class directly, just use <see cref="CreateBooleanQuery(string, string)"/>
/// and <see cref="CreatePhraseQuery(string, string)"/>. </summary>
/// <param name="analyzer"> Analyzer used for this query. </param>
/// <param name="operator"> Default boolean operator used for this query. </param>
/// <param name="field"> Field to create queries against. </param>
/// <param name="queryText"> Text to be passed to the analysis chain. </param>
/// <param name="quoted"> <c>true</c> if phrases should be generated when terms occur at more than one position. </param>
/// <param name="phraseSlop"> Slop factor for phrase/multiphrase queries. </param>
protected Query CreateFieldQuery(Analyzer analyzer, Occur @operator, string field, string queryText, bool quoted, int phraseSlop)
{
Debug.Assert(@operator == Occur.SHOULD || @operator == Occur.MUST);
// Use the analyzer to get all the tokens, and then build a TermQuery,
// PhraseQuery, or nothing based on the term count
CachingTokenFilter buffer = null;
ITermToBytesRefAttribute termAtt = null;
IPositionIncrementAttribute posIncrAtt = null;
int numTokens = 0;
int positionCount = 0;
bool severalTokensAtSamePosition = false;
bool hasMoreTokens = false;
TokenStream source = null;
try
{
source = analyzer.GetTokenStream(field, new StringReader(queryText));
source.Reset();
buffer = new CachingTokenFilter(source);
buffer.Reset();
if (buffer.HasAttribute <ITermToBytesRefAttribute>())
{
termAtt = buffer.GetAttribute <ITermToBytesRefAttribute>();
}
if (buffer.HasAttribute <IPositionIncrementAttribute>())
{
posIncrAtt = buffer.GetAttribute <IPositionIncrementAttribute>();
}
if (termAtt != null)
{
try
{
hasMoreTokens = buffer.IncrementToken();
while (hasMoreTokens)
{
numTokens++;
int positionIncrement = (posIncrAtt != null) ? posIncrAtt.PositionIncrement : 1;
if (positionIncrement != 0)
{
positionCount += positionIncrement;
}
else
{
severalTokensAtSamePosition = true;
}
hasMoreTokens = buffer.IncrementToken();
}
}
catch (System.IO.IOException)
{
// ignore
}
}
}
catch (System.IO.IOException e)
{
throw new Exception("Error analyzing query text", e);
}
finally
{
IOUtils.DisposeWhileHandlingException(source);
}
// rewind the buffer stream
buffer.Reset();
BytesRef bytes = termAtt == null ? null : termAtt.BytesRef;
if (numTokens == 0)
{
return(null);
}
else if (numTokens == 1)
{
try
{
bool hasNext = buffer.IncrementToken();
Debug.Assert(hasNext == true);
termAtt.FillBytesRef();
}
catch (System.IO.IOException)
{
// safe to ignore, because we know the number of tokens
}
return(NewTermQuery(new Term(field, BytesRef.DeepCopyOf(bytes))));
}
else
{
if (severalTokensAtSamePosition || (!quoted))
{
if (positionCount == 1 || (!quoted))
{
// no phrase query:
if (positionCount == 1)
{
// simple case: only one position, with synonyms
BooleanQuery q = NewBooleanQuery(true);
for (int i = 0; i < numTokens; i++)
{
try
{
bool hasNext = buffer.IncrementToken();
Debug.Assert(hasNext == true);
termAtt.FillBytesRef();
}
catch (System.IO.IOException)
{
// safe to ignore, because we know the number of tokens
}
Query currentQuery = NewTermQuery(new Term(field, BytesRef.DeepCopyOf(bytes)));
q.Add(currentQuery, Occur.SHOULD);
}
return(q);
}
else
{
// multiple positions
BooleanQuery q = NewBooleanQuery(false);
Query currentQuery = null;
for (int i = 0; i < numTokens; i++)
{
try
{
bool hasNext = buffer.IncrementToken();
Debug.Assert(hasNext == true);
termAtt.FillBytesRef();
}
catch (System.IO.IOException)
{
// safe to ignore, because we know the number of tokens
}
if (posIncrAtt != null && posIncrAtt.PositionIncrement == 0)
{
if (!(currentQuery is BooleanQuery))
{
Query t = currentQuery;
currentQuery = NewBooleanQuery(true);
((BooleanQuery)currentQuery).Add(t, Occur.SHOULD);
}
((BooleanQuery)currentQuery).Add(NewTermQuery(new Term(field, BytesRef.DeepCopyOf(bytes))), Occur.SHOULD);
}
else
{
if (currentQuery != null)
{
q.Add(currentQuery, @operator);
}
currentQuery = NewTermQuery(new Term(field, BytesRef.DeepCopyOf(bytes)));
}
}
q.Add(currentQuery, @operator);
return(q);
}
}
else
{
// phrase query:
MultiPhraseQuery mpq = NewMultiPhraseQuery();
mpq.Slop = phraseSlop;
IList <Term> multiTerms = new List <Term>();
int position = -1;
for (int i = 0; i < numTokens; i++)
{
int positionIncrement = 1;
try
{
bool hasNext = buffer.IncrementToken();
Debug.Assert(hasNext == true);
termAtt.FillBytesRef();
if (posIncrAtt != null)
{
positionIncrement = posIncrAtt.PositionIncrement;
}
}
catch (System.IO.IOException)
{
// safe to ignore, because we know the number of tokens
}
if (positionIncrement > 0 && multiTerms.Count > 0)
{
if (enablePositionIncrements)
{
mpq.Add(multiTerms.ToArray(), position);
}
else
{
mpq.Add(multiTerms.ToArray());
}
multiTerms.Clear();
}
position += positionIncrement;
multiTerms.Add(new Term(field, BytesRef.DeepCopyOf(bytes)));
}
if (enablePositionIncrements)
{
mpq.Add(multiTerms.ToArray(), position);
}
else
{
mpq.Add(multiTerms.ToArray());
}
return(mpq);
}
}
else
{
PhraseQuery pq = NewPhraseQuery();
pq.Slop = phraseSlop;
int position = -1;
for (int i = 0; i < numTokens; i++)
{
int positionIncrement = 1;
try
{
bool hasNext = buffer.IncrementToken();
Debug.Assert(hasNext == true);
termAtt.FillBytesRef();
if (posIncrAtt != null)
{
positionIncrement = posIncrAtt.PositionIncrement;
}
}
catch (System.IO.IOException)
{
// safe to ignore, because we know the number of tokens
}
if (enablePositionIncrements)
{
position += positionIncrement;
pq.Add(new Term(field, BytesRef.DeepCopyOf(bytes)), position);
}
else
{
pq.Add(new Term(field, BytesRef.DeepCopyOf(bytes)));
}
}
return(pq);
}
}
}
public virtual Query VisitMultiPhraseQuery(MultiPhraseQuery multiPhraseq)
{
throw new NotImplementedException();
}
public override Query VisitMultiPhraseQuery(MultiPhraseQuery multiPhraseq)
{
throw new SnNotSupportedException();
}
public virtual void TestCJKSynonymsPhrase()
{
MultiPhraseQuery expected = new MultiPhraseQuery();
expected.Add(new Term("field", "中"));
expected.Add(new Term[] { new Term("field", "国"), new Term("field", "國") });
QueryBuilder builder = new QueryBuilder(new MockCJKSynonymAnalyzer());
Assert.AreEqual(expected, builder.CreatePhraseQuery("field", "中国"));
expected.Slop = 3;
Assert.AreEqual(expected, builder.CreatePhraseQuery("field", "中国", 3));
}
public virtual void TestSynonymsPhrase()
{
MultiPhraseQuery expected = new MultiPhraseQuery();
expected.Add(new Term("field", "old"));
expected.Add(new Term[] { new Term("field", "dogs"), new Term("field", "dog") });
QueryBuilder builder = new QueryBuilder(new MockSynonymAnalyzer());
Assert.AreEqual(expected, builder.CreatePhraseQuery("field", "old dogs"));
}
private Query BuildMatchAnywhereQuery(IndexReader indexReader, string expandedSearchString, bool matchCase)
{
List <string> searchTerms = null;
string adjustedSearchString = expandedSearchString;
string fieldToSearch = Constants.IndexFields.Content;
if (!matchCase)
{
fieldToSearch = Constants.IndexFields.ContentCaseInsensitive;
adjustedSearchString = adjustedSearchString.ToLower();
}
searchTerms = adjustedSearchString.Split(new char[] { ' ' }, StringSplitOptions.RemoveEmptyEntries).ToList();
bool onlyOneTerm = searchTerms.Count == 1;
var firstTerm = searchTerms.FirstOrDefault();
var lastTerm = searchTerms.LastOrDefault();
Query contentQuery = null;
if (onlyOneTerm)
{
bool isFirstTermPunctuation = (firstTerm.Length == 1 && LuceneHelper.IsPunctuation(firstTerm.First()));
if (isFirstTermPunctuation)
{
contentQuery = new TermQuery(new Term(fieldToSearch, firstTerm));
}
else
{
contentQuery = new WildcardQuery(new Term(fieldToSearch, "*" + firstTerm + "*"));
}
}
else
{
MultiPhraseQuery phraseQuery = new MultiPhraseQuery();
List <Term> firstTermMatches = new List <Term>();
List <Term> lastTermMatches = new List <Term>();
CollectFirstAndLastTermMatches(indexReader, fieldToSearch, firstTermMatches, lastTermMatches, firstTerm, lastTerm);
if (firstTermMatches.Count > 0)
{
phraseQuery.Add(firstTermMatches.ToArray());
}
bool includeFirstTerm = firstTermMatches.Count == 0;
bool includeLastTerm = lastTermMatches.Count == 0;
int startIndex = includeFirstTerm ? 0 : 1;
int endIndex = searchTerms.Count - (includeLastTerm ? 0 : 1);
for (int i = startIndex; i < endIndex; i++)
{
phraseQuery.Add(new Term(fieldToSearch, searchTerms[i]));
}
if (lastTermMatches.Count > 0)
{
phraseQuery.Add(lastTermMatches.ToArray());
}
contentQuery = phraseQuery;
}
return(contentQuery);
}
public override Query VisitMultiPhraseQuery(MultiPhraseQuery multiPhraseq)
{
throw new NotImplementedException();
}
private void lucene_serach_Click(object sender, EventArgs e)
{
results.Rows.Clear();
var query = search_field.Text.ToLower();
var array = query.Split(' ').ToList();
var searcher = new IndexSearcher(writer1.GetReader(applyAllDeletes: true));
var totalResults = new List <Document>();
//одно слово
QueryParser parser = new QueryParser(AppLuceneVersion, "name", analyzer);
var phrase = new MultiPhraseQuery();
foreach (var word in array)
{
var q = parser.Parse(query);
if (!String.IsNullOrEmpty(word))
{
var res = searcher.Search(q, 10).ScoreDocs;
foreach (var hit in res)
{
var foundDoc = searcher.Doc(hit.Doc);
if (!totalResults.Any(f =>
f.GetField("id").GetInt32Value() == foundDoc.GetField("id").GetInt32Value()))
{
totalResults.Add(foundDoc);
}
}
}
}
// полное название
phrase.Add(new Term("name", query));
var hits = searcher.Search(phrase, 10).ScoreDocs;
foreach (var hit in hits)
{
var foundDoc = searcher.Doc(hit.Doc);
if (!totalResults.Any(f => f.GetField("id").GetInt32Value() == foundDoc.GetField("id").GetInt32Value()))
{
totalResults.Add(foundDoc);
}
}
//части слов
foreach (var word in array)
{
if (!string.IsNullOrEmpty(word))
{
var wild = new WildcardQuery(new Term("name", "*" + word + "*"));
var res = searcher.Search(wild, 10).ScoreDocs;
foreach (var hit in res)
{
var foundDoc = searcher.Doc(hit.Doc);
if (!totalResults.Any(f =>
f.GetField("id").GetInt32Value() == foundDoc.GetField("id").GetInt32Value()))
{
totalResults.Add(foundDoc);
}
}
}
}
//год и часть слова
var year_to_find = "";
var number = 0;
foreach (var word in array)
{
var result = TryParse(word, out number);
if (result && number > 1800 && number <= 9999)
{
year_to_find = word;
array.RemoveAt(array.IndexOf(word));
break;
}
}
Console.WriteLine(number != 0);
if (number != 0)
{
phrase = new MultiPhraseQuery();
foreach (var word in array)
{
if (!string.IsNullOrEmpty(word))
{
var booleanQuery = new BooleanQuery();
var wild = new WildcardQuery(new Term("name", "*" + word + "*"));
var num = NumericRangeQuery.NewInt32Range("year", 1, number, number, true, true);
booleanQuery.Add(wild, Occur.SHOULD);
booleanQuery.Add(num, Occur.SHOULD);
var res = searcher.Search(booleanQuery, 10).ScoreDocs;
foreach (var hit in res)
{
var foundDoc = searcher.Doc(hit.Doc);
if (!totalResults.Any(f =>
f.GetField("id").GetInt32Value() == foundDoc.GetField("id").GetInt32Value()))
{
totalResults.Add(foundDoc);
}
}
}
}
}
foreach (var doc in totalResults)
{
results.Rows.Add(doc.GetField("id").GetInt32Value().ToString(),
doc.GetValues("name")[0],
doc.GetField("year").GetInt32Value().ToString());
}
}
private void QueryToSpanQuery(Query query, ICollection <byte[]> payloads)
{
if (query is BooleanQuery)
{
BooleanClause[] queryClauses = ((BooleanQuery)query).GetClauses();
for (int i = 0; i < queryClauses.Length; i++)
{
if (!queryClauses[i].IsProhibited())
{
QueryToSpanQuery(queryClauses[i].GetQuery(), payloads);
}
}
}
else if (query is PhraseQuery)
{
Term[] phraseQueryTerms = ((PhraseQuery)query).GetTerms();
SpanQuery[] clauses = new SpanQuery[phraseQueryTerms.Length];
for (int i = 0; i < phraseQueryTerms.Length; i++)
{
clauses[i] = new SpanTermQuery(phraseQueryTerms[i]);
}
int slop = ((PhraseQuery)query).GetSlop();
bool inorder = false;
if (slop == 0)
{
inorder = true;
}
SpanNearQuery sp = new SpanNearQuery(clauses, slop, inorder);
sp.SetBoost(query.GetBoost());
GetPayloads(payloads, sp);
}
else if (query is TermQuery)
{
SpanTermQuery stq = new SpanTermQuery(((TermQuery)query).GetTerm());
stq.SetBoost(query.GetBoost());
GetPayloads(payloads, stq);
}
else if (query is SpanQuery)
{
GetPayloads(payloads, (SpanQuery)query);
}
else if (query is FilteredQuery)
{
QueryToSpanQuery(((FilteredQuery)query).GetQuery(), payloads);
}
else if (query is DisjunctionMaxQuery)
{
for (System.Collections.IEnumerator iterator = ((DisjunctionMaxQuery)query).Iterator(); iterator.MoveNext();)
{
QueryToSpanQuery((Query)iterator.Current, payloads);
}
}
else if (query is MultiPhraseQuery)
{
MultiPhraseQuery mpq = (MultiPhraseQuery)query;
System.Collections.IList termArrays = mpq.GetTermArrays();
int[] positions = mpq.GetPositions();
if (positions.Length > 0)
{
int maxPosition = positions[positions.Length - 1];
for (int i = 0; i < positions.Length - 1; ++i)
{
if (positions[i] > maxPosition)
{
maxPosition = positions[i];
}
}
System.Collections.ArrayList[] disjunctLists = new System.Collections.ArrayList[maxPosition + 1];
int distinctPositions = 0;
for (int i = 0; i < termArrays.Count; ++i)
{
Term[] termArray = (Term[])termArrays[i];
System.Collections.IList disjuncts = disjunctLists[positions[i]];
if (disjuncts == null)
{
disjuncts = (disjunctLists[positions[i]] = new System.Collections.ArrayList(termArray.Length));
++distinctPositions;
}
for (int j = 0; j < termArray.Length; ++j)
{
disjuncts.Add(new SpanTermQuery(termArray[j]));
}
}
int positionGaps = 0;
int position = 0;
SpanQuery[] clauses = new SpanQuery[distinctPositions];
for (int i = 0; i < disjunctLists.Length; ++i)
{
System.Collections.ArrayList disjuncts = disjunctLists[i];
if (disjuncts != null)
{
clauses[position++] = new SpanOrQuery((SpanQuery[])(disjuncts.ToArray(typeof(SpanQuery[]))));
}
else
{
++positionGaps;
}
}
int slop = mpq.GetSlop();
bool inorder = (slop == 0);
SpanNearQuery sp = new SpanNearQuery(clauses, slop + positionGaps, inorder);
sp.SetBoost(query.GetBoost());
GetPayloads(payloads, sp);
}
}
}
public async Task <SearchResult <T> > SearchAsync(SearchQuery queryDefinition, CancellationToken cancellationToken = default)
{
using (await writerLock.ReaderLockAsync(cancellationToken))
{
var result = new SearchResult <T>();
List <T> hits = new List <T>();
using (var writer = getWriter())
{
Query query = new MatchAllDocsQuery();
// Term queries
if (queryDefinition.TermQueries.Any())
{
var phraseQuery = new MultiPhraseQuery();
foreach (var termQuery in queryDefinition.TermQueries)
{
phraseQuery.Add(
termQuery.value
.Split(" ".ToCharArray(), StringSplitOptions.RemoveEmptyEntries)
.Select(phrase => new Term(termQuery.field, phrase.ToLower()))
.ToArray()
);
}
query = phraseQuery;
}
var reader = writer.DocsWriter.GetReader(applyAllDeletes: true);
var searcher = new IndexSearcher(reader);
var luceneResult = searcher.Search(query, queryDefinition.Limit);
foreach (var doc in luceneResult.ScoreDocs)
{
var foundDoc = searcher.Doc(doc.Doc);
hits.Add(await inflateDocument(foundDoc));
}
result.TotalHits = luceneResult.TotalHits;
result.Hits = hits;
// Facets
if (queryDefinition.Facets.Any())
{
FacetsConfig facetsConfig = new FacetsConfig();
FacetsCollector fc = new FacetsCollector();
FacetsCollector.Search(searcher, query, queryDefinition.FacetMax, fc);
using (var taxonomyReader = new DirectoryTaxonomyReader(FSDirectory.Open(Path.Combine(options.IndexPath, indexType, "taxonomy"))))
{
var facets = new FastTaxonomyFacetCounts(taxonomyReader, facetsConfig, fc);
foreach (var facet in queryDefinition.Facets)
{
var facetGroup = new FacetGroup {
Field = facet
};
facetGroup.Facets =
facets.GetTopChildren(queryDefinition.FacetMax, facet).LabelValues
.Select(x => new Facet {
Key = x.Label, Count = (long)x.Value
})
.ToArray();
result.FacetGroups.Add(facetGroup);
}
}
}
}
return(result);
}
}
private void QueryToSpanQuery(Query query, ICollection <byte[]> payloads)
{
if (query is BooleanQuery)
{
BooleanClause[] queryClauses = ((BooleanQuery)query).GetClauses();
for (int i = 0; i < queryClauses.Length; i++)
{
if (!queryClauses[i].IsProhibited)
{
QueryToSpanQuery(queryClauses[i].Query, payloads);
}
}
}
else if (query is PhraseQuery)
{
Term[] phraseQueryTerms = ((PhraseQuery)query).GetTerms();
SpanQuery[] clauses = new SpanQuery[phraseQueryTerms.Length];
for (int i = 0; i < phraseQueryTerms.Length; i++)
{
clauses[i] = new SpanTermQuery(phraseQueryTerms[i]);
}
int slop = ((PhraseQuery)query).Slop;
bool inorder = false;
if (slop == 0)
{
inorder = true;
}
SpanNearQuery sp = new SpanNearQuery(clauses, slop, inorder);
sp.Boost = query.Boost;
GetPayloads(payloads, sp);
}
else if (query is TermQuery)
{
SpanTermQuery stq = new SpanTermQuery(((TermQuery)query).Term);
stq.Boost = query.Boost;
GetPayloads(payloads, stq);
}
else if (query is SpanQuery)
{
GetPayloads(payloads, (SpanQuery)query);
}
else if (query is FilteredQuery)
{
QueryToSpanQuery(((FilteredQuery)query).Query, payloads);
}
else if (query is DisjunctionMaxQuery)
{
foreach (var q in ((DisjunctionMaxQuery)query))
{
QueryToSpanQuery(q, payloads);
}
}
else if (query is MultiPhraseQuery)
{
MultiPhraseQuery mpq = (MultiPhraseQuery)query;
IList <Term[]> termArrays = mpq.GetTermArrays();
int[] positions = mpq.GetPositions();
if (positions.Length > 0)
{
int maxPosition = positions[positions.Length - 1];
for (int i = 0; i < positions.Length - 1; ++i)
{
if (positions[i] > maxPosition)
{
maxPosition = positions[i];
}
}
// LUCENENET: Changed from Query to SpanQuery to eliminate the O(n) cast
// required to instantiate SpanOrQuery below
IList <SpanQuery>[] disjunctLists = new List <SpanQuery> [maxPosition + 1];
int distinctPositions = 0;
for (int i = 0; i < termArrays.Count; ++i)
{
Term[] termArray = termArrays[i];
IList <SpanQuery> disjuncts = disjunctLists[positions[i]]; // LUCENENET: Changed from Query to SpanQuery
if (disjuncts == null)
{
disjuncts = (disjunctLists[positions[i]] = new List <SpanQuery>(termArray.Length)); // LUCENENET: Changed from Query to SpanQuery
++distinctPositions;
}
foreach (Term term in termArray)
{
disjuncts.Add(new SpanTermQuery(term));
}
}
int positionGaps = 0;
int position = 0;
SpanQuery[] clauses = new SpanQuery[distinctPositions];
for (int i = 0; i < disjunctLists.Length; ++i)
{
IList <SpanQuery> disjuncts = disjunctLists[i]; // LUCENENET: Changed from Query to SpanQuery
if (disjuncts != null)
{
clauses[position++] = new SpanOrQuery(disjuncts);
}
else
{
++positionGaps;
}
}
int slop = mpq.Slop;
bool inorder = (slop == 0);
SpanNearQuery sp = new SpanNearQuery(clauses, slop + positionGaps, inorder);
sp.Boost = query.Boost;
GetPayloads(payloads, sp);
}
}
}
private void button4_Click(object sender, EventArgs e)
{
int counter = 0;
var query = find_text.ToLower();
var array = query.Split(' ').ToList();
List <string> res_list = new List <string>();
var searcher = new IndexSearcher(writer.GetReader());
var totalResults = new List <Document>();
//поиск по одному слову из названия
var phrase = new MultiPhraseQuery();
foreach (var word in array)
{
phrase = new MultiPhraseQuery();
if (!String.IsNullOrEmpty(word))
{
phrase.Add(new Term("name", word));
var res = searcher.Search(phrase, 10).ScoreDocs;
foreach (var hit in res)
{
var foundDoc = searcher.Doc(hit.Doc);
if (!totalResults.Any(f =>
f.GetField("id").ToString() == foundDoc.GetField("id").ToString()))
{
totalResults.Add(foundDoc);
}
}
}
}
//поиск по всем словам названия
phrase = new MultiPhraseQuery();
phrase.Add(new Term("name", query));
var hits = searcher.Search(phrase, 10).ScoreDocs;
foreach (var hit in hits)
{
var foundDoc = searcher.Doc(hit.Doc);
if (!totalResults.Any(f => f.GetField("id").ToString() == foundDoc.GetField("id").ToString()))
{
totalResults.Add(foundDoc);
}
}
//поиск по частичным словам названия
foreach (var word in array)
{
if (!String.IsNullOrEmpty(word))
{
var wild = new WildcardQuery(new Term("name", "*" + word + "*"));
var res = searcher.Search(wild, 10).ScoreDocs;
foreach (var hit in res)
{
var foundDoc = searcher.Doc(hit.Doc);
if (!totalResults.Any(f =>
f.GetField("id").ToString() == foundDoc.GetField("id").ToString()))
{
totalResults.Add(foundDoc);
}
}
}
}
//поиск по году и названию (части названия)
string year_to_find = "";
int number = 0;
foreach (var word in array)
{
bool result = Int32.TryParse(word, out number);
if (result && number > 1800 && number <= 9999)
{
year_to_find = word;
array.RemoveAt(array.IndexOf(word));
break;
}
}
Console.WriteLine(number != 0);
if (number != 0)
{
phrase = new MultiPhraseQuery();
foreach (var word in array)
{
if (!String.IsNullOrEmpty(word))
{
BooleanQuery booleanQuery = new BooleanQuery();
var wild = new WildcardQuery(new Term("name", "*" + word + "*"));
var num = NumericRangeQuery.NewIntRange("year", 1, number, number, true, true);
booleanQuery.Add(wild, Occur.SHOULD);
booleanQuery.Add(num, Occur.SHOULD);
var res = searcher.Search(booleanQuery, 10).ScoreDocs;
foreach (var hit in res)
{
var foundDoc = searcher.Doc(hit.Doc);
if (!totalResults.Any(f =>
f.GetField("id").ToString() == foundDoc.GetField("id").ToString()))
{
totalResults.Add(foundDoc);
}
}
}
}
}
foreach (var doc in totalResults)
{
textBox1.AppendText(doc.ToString());
}
}
/// <summary>
/// Fills a <see cref="T:IDictionary{string, WeightedSpanTerm}"/> with <see cref="WeightedSpanTerm"/>s using the terms from the supplied <paramref name="query"/>.
/// </summary>
/// <param name="query"><see cref="Query"/> to extract Terms from</param>
/// <param name="terms">Map to place created <see cref="WeightedSpanTerm"/>s in</param>
/// <exception cref="System.IO.IOException">If there is a low-level I/O error</exception>
protected virtual void Extract(Query query, IDictionary <string, WeightedSpanTerm> terms)
{
if (query is BooleanQuery)
{
IList <BooleanClause> queryClauses = ((BooleanQuery)query).Clauses;
for (int i = 0; i < queryClauses.Count; i++)
{
if (!queryClauses[i].IsProhibited)
{
Extract(queryClauses[i].Query, terms);
}
}
}
else if (query is PhraseQuery)
{
PhraseQuery phraseQuery = (PhraseQuery)query;
Term[] phraseQueryTerms = phraseQuery.GetTerms();
SpanQuery[] clauses = new SpanQuery[phraseQueryTerms.Length];
for (int i = 0; i < phraseQueryTerms.Length; i++)
{
clauses[i] = new SpanTermQuery(phraseQueryTerms[i]);
}
int slop = phraseQuery.Slop;
int[] positions = phraseQuery.GetPositions();
// add largest position increment to slop
if (positions.Length > 0)
{
int lastPos = positions[0];
int largestInc = 0;
int sz = positions.Length;
for (int i = 1; i < sz; i++)
{
int pos = positions[i];
int inc = pos - lastPos;
if (inc > largestInc)
{
largestInc = inc;
}
lastPos = pos;
}
if (largestInc > 1)
{
slop += largestInc;
}
}
bool inorder = slop == 0;
SpanNearQuery sp = new SpanNearQuery(clauses, slop, inorder);
sp.Boost = query.Boost;
ExtractWeightedSpanTerms(terms, sp);
}
else if (query is TermQuery)
{
ExtractWeightedTerms(terms, query);
}
else if (query is SpanQuery)
{
ExtractWeightedSpanTerms(terms, (SpanQuery)query);
}
else if (query is FilteredQuery)
{
Extract(((FilteredQuery)query).Query, terms);
}
else if (query is ConstantScoreQuery)
{
Query q = ((ConstantScoreQuery)query).Query;
if (q != null)
{
Extract(q, terms);
}
}
else if (query is CommonTermsQuery)
{
// specialized since rewriting would change the result query
// this query is TermContext sensitive.
ExtractWeightedTerms(terms, query);
}
else if (query is DisjunctionMaxQuery)
{
foreach (var q in ((DisjunctionMaxQuery)query))
{
Extract(q, terms);
}
}
else if (query is MultiPhraseQuery)
{
MultiPhraseQuery mpq = (MultiPhraseQuery)query;
IList <Term[]> termArrays = mpq.GetTermArrays();
int[] positions = mpq.GetPositions();
if (positions.Length > 0)
{
int maxPosition = positions[positions.Length - 1];
for (int i = 0; i < positions.Length - 1; ++i)
{
if (positions[i] > maxPosition)
{
maxPosition = positions[i];
}
}
var disjunctLists = new List <SpanQuery> [maxPosition + 1];
int distinctPositions = 0;
for (int i = 0; i < termArrays.Count; ++i)
{
Term[] termArray = termArrays[i];
List <SpanQuery> disjuncts = disjunctLists[positions[i]];
if (disjuncts == null)
{
disjuncts = (disjunctLists[positions[i]] = new List <SpanQuery>(termArray.Length));
++distinctPositions;
}
foreach (var term in termArray)
{
disjuncts.Add(new SpanTermQuery(term));
}
}
int positionGaps = 0;
int position = 0;
SpanQuery[] clauses = new SpanQuery[distinctPositions];
foreach (var disjuncts in disjunctLists)
{
if (disjuncts != null)
{
clauses[position++] = new SpanOrQuery(disjuncts.ToArray());
}
else
{
++positionGaps;
}
}
int slop = mpq.Slop;
bool inorder = (slop == 0);
SpanNearQuery sp = new SpanNearQuery(clauses, slop + positionGaps, inorder);
sp.Boost = query.Boost;
ExtractWeightedSpanTerms(terms, sp);
}
}
else
{
Query origQuery = query;
if (query is MultiTermQuery)
{
if (!expandMultiTermQuery)
{
return;
}
MultiTermQuery copy = (MultiTermQuery)query.Clone();
copy.MultiTermRewriteMethod = MultiTermQuery.SCORING_BOOLEAN_QUERY_REWRITE;
origQuery = copy;
}
IndexReader reader = GetLeafContext().Reader;
Query rewritten = origQuery.Rewrite(reader);
if (rewritten != origQuery)
{
// only rewrite once and then flatten again - the rewritten query could have a speacial treatment
// if this method is overwritten in a subclass or above in the next recursion
Extract(rewritten, terms);
}
}
ExtractUnknownQuery(query, terms);
}
public void TestQueryScorerMultiPhraseQueryHighlighting()
{
MultiPhraseQuery mpq = new MultiPhraseQuery();
mpq.Add(new Term[] {new Term(FIELD_NAME, "wordx"), new Term(FIELD_NAME, "wordb")});
mpq.Add(new Term(FIELD_NAME, "wordy"));
DoSearching(mpq);
int maxNumFragmentsRequired = 2;
AssertExpectedHighlightCount(maxNumFragmentsRequired, 6);
}
/// <summary>
/// On base of https://lucenenet.apache.org
/// </summary>
/// <param name="query"></param>
/// <returns></returns>
public IEnumerable <(string name, int year)> SearchWithLucy(string query)
{
var words = query.Split(' ').ToList();
var searcher = new IndexSearcher(lucyAdapter.lucyWriter.GetReader(applyAllDeletes: true));
var totalResults = new List <Document>();
//word
MultiPhraseQuery multiPhraseQuery;
foreach (var word in words)
{
multiPhraseQuery = new MultiPhraseQuery();
if (string.IsNullOrEmpty(word))
{
continue;
}
multiPhraseQuery.Add(new Term("name_word", word));
var docs = searcher.Search(multiPhraseQuery, 10).ScoreDocs;
foreach (var doc in docs)
{
var document = searcher.Doc(doc.Doc);
if (totalResults.All(f => f.GetField("id").GetInt32Value() != document.GetField("id").GetInt32Value()))
{
totalResults.Add(document);
}
}
}
// full name
multiPhraseQuery = new MultiPhraseQuery();
multiPhraseQuery.Add(new Term("full_name", query));
var scoreDocs = searcher.Search(multiPhraseQuery, 10).ScoreDocs;
foreach (var scoreDoc in scoreDocs)
{
var doc = searcher.Doc(scoreDoc.Doc);
if (totalResults.All(f => f.GetField("id").GetInt32Value() != doc.GetField("id").GetInt32Value()))
{
totalResults.Add(doc);
}
}
//word parts
foreach (var word in words)
{
if (string.IsNullOrEmpty(word))
{
continue;
}
var wildcardQuery = new WildcardQuery(new Term("name_word", "*" + word + "*"));
var docs = searcher.Search(wildcardQuery, 10).ScoreDocs;
foreach (var doc in docs)
{
var document = searcher.Doc(doc.Doc);
if (totalResults.All(f => f.GetField("id").GetInt32Value() != document.GetField("id").GetInt32Value()))
{
totalResults.Add(document);
}
}
}
//year and word part
var number = 0;
foreach (var word in words)
{
var result = int.TryParse(word, out number);
if (!result)
{
continue;
}
words.RemoveAt(words.IndexOf(word));
break;
}
if (number != 0)
{
foreach (var word in words)
{
if (string.IsNullOrEmpty(word))
{
continue;
}
var booleanQuery = new BooleanQuery();
var wildcardQuery = new WildcardQuery(new Term("name_word", "*" + word + "*"));
var rangeQuery = NumericRangeQuery.NewInt32Range("year", 1, number, number, true, true);
booleanQuery.Add(wildcardQuery, Occur.SHOULD);
booleanQuery.Add(rangeQuery, Occur.SHOULD);
var docs = searcher.Search(booleanQuery, 10).ScoreDocs;
foreach (var doc in docs)
{
var foundDoc = searcher.Doc(doc.Doc);
if (totalResults.All(f => f.GetField("id").GetInt32Value() != foundDoc.GetField("id").GetInt32Value()))
{
totalResults.Add(foundDoc);
}
}
}
}
foreach (var doc in totalResults.Take(10))
{
yield return(doc.GetValues("full_name")[0], (int)doc.GetField("year").GetInt32Value());
}
}
public void TestQueryScorerMultiPhraseQueryHighlightingWithGap()
{
MultiPhraseQuery mpq = new MultiPhraseQuery();
/*
* The toString of MultiPhraseQuery doesn't work so well with these
* out-of-order additions, but the Query itself seems to match accurately.
*/
mpq.Add(new Term[] {new Term(FIELD_NAME, "wordz")}, 2);
mpq.Add(new Term[] {new Term(FIELD_NAME, "wordx")}, 0);
DoSearching(mpq);
int maxNumFragmentsRequired = 1;
int expectedHighlights = 2;
AssertExpectedHighlightCount(maxNumFragmentsRequired, expectedHighlights);
}
public void LuceneMinimalExampleTest()
{
//var dir = new RAMDirectory();
//var analyzer = new StandardAnalyzer(LuceneVersion.LUCENE_48);
//X) Create an index and define a text analyzer ------------------------
// Ensures index backward compatibility
const LuceneVersion AppLuceneVersion = LuceneVersion.LUCENE_48;
// Construct a machine-independent path for the index
var basePath = Environment.GetFolderPath(Environment.SpecialFolder.CommonApplicationData);
var indexPath = Path.Combine(basePath, "index");
System.IO.Directory.Delete(indexPath);
using var dir = FSDirectory.Open(indexPath);
// Create an analyzer to process the text
var analyzer = new StandardAnalyzer(AppLuceneVersion);
// Create an index writer
var indexConfig = new IndexWriterConfig(AppLuceneVersion, analyzer);
using var writer = new IndexWriter(dir, indexConfig);
//X) Add to the index ------------------------
var source = new
{
Name = "Kermit the Frog",
FavoritePhrase = "The quick brown fox jumps over the lazy dog"
};
var doc = new Document
{
// StringField indexes but doesn't tokenize
new StringField("name", source.Name, Field.Store.YES),
new TextField("favoritePhrase", source.FavoritePhrase, Field.Store.YES)
};
writer.AddDocument(doc);
writer.Flush(triggerMerge: false, applyAllDeletes: false);
//X) Construct a query ------------------------
// Search with a phrase
var phrase = new MultiPhraseQuery
{
new Term("favoritePhrase", "brown"),
new Term("favoritePhrase", "fox")
};
//X) Fetch the results ------------------------
// Re-use the writer to get real-time updates
using var reader = writer.GetReader(applyAllDeletes: true);
var searcher = new IndexSearcher(reader);
ScoreDoc[] hits = searcher.Search(phrase, 20 /* top 20 */).ScoreDocs;
// Display the output in a table
string strDisp = $"{"Score",10}" + $" {"Name",-15}" + $" {"Favorite Phrase",-40}";
Console.WriteLine(strDisp);
foreach (var hit in hits)
{
var foundDoc = searcher.Doc(hit.Doc);
string str = $"{hit.Score:f8}" + $" {foundDoc.Get("name"),-15}" + $" {foundDoc.Get("favoritePhrase"),-40}";
Console.WriteLine(str);
}
}
private void lucene_serach_Click(object sender, EventArgs e)
{
int counter = 0;
this.results.Rows.Clear();
var query = this.search_field.Text.ToLower();
var array = query.Split(' ').ToList();
List <string> res_list = new List <string>();
var searcher = new IndexSearcher(writer1.GetReader(applyAllDeletes: true));
//одно слово
var phrase = new MultiPhraseQuery();
foreach (var word in array)
{
phrase = new MultiPhraseQuery();
if (!String.IsNullOrEmpty(word))
{
phrase.Add(new Term("name_word", word));
var res = searcher.Search(phrase, 10).ScoreDocs;
foreach (var hit in res)
{
var foundDoc = searcher.Doc(hit.Doc);
var score = hit.Score;
this.results.Rows.Add(foundDoc.GetField("id").GetInt32Value().ToString(), foundDoc.GetValues("full_name")[0], foundDoc.GetField("year").GetInt32Value().ToString());
}
}
}
// полное название
phrase = new MultiPhraseQuery();
phrase.Add(new Term("full_name", query));
var hits = searcher.Search(phrase, 10).ScoreDocs;
foreach (var hit in hits)
{
var foundDoc = searcher.Doc(hit.Doc);
var score = hit.Score;
this.results.Rows.Add(foundDoc.GetField("id").GetInt32Value().ToString(), foundDoc.GetValues("full_name")[0], foundDoc.GetField("year").GetInt32Value().ToString());
}
//части слов
foreach (var word in array)
{
if (!String.IsNullOrEmpty(word))
{
var wild = new WildcardQuery(new Term("name_word", "*" + word + "*"));
var res = searcher.Search(wild, 10).ScoreDocs;
foreach (var hit in res)
{
var foundDoc = searcher.Doc(hit.Doc);
var score = hit.Score;
this.results.Rows.Add(foundDoc.GetField("id").GetInt32Value().ToString(), foundDoc.GetValues("full_name")[0], foundDoc.GetField("year").GetInt32Value().ToString());
}
}
}
//год и часть слова
string year_to_find = "";
int number = 0;
foreach (var word in array)
{
bool result = Int32.TryParse(word, out number);
if (result && number > 1800 && number <= 9999)
{
year_to_find = word;
array.RemoveAt(array.IndexOf(word));
break;
}
}
Console.WriteLine(number != 0);
if (number != 0)
{
phrase = new MultiPhraseQuery();
foreach (var word in array)
{
if (!String.IsNullOrEmpty(word))
{
BooleanQuery booleanQuery = new BooleanQuery();
var wild = new WildcardQuery(new Term("name_word", "*" + word + "*"));
var num = NumericRangeQuery.NewInt32Range("year", 1, number, number, true, true);
booleanQuery.Add(wild, Occur.SHOULD);
booleanQuery.Add(num, Occur.SHOULD);
var res = searcher.Search(booleanQuery, 10).ScoreDocs;
foreach (var hit in res)
{
var foundDoc = searcher.Doc(hit.Doc);
var score = hit.Score;
this.results.Rows.Add(foundDoc.GetField("id").GetInt32Value().ToString(), foundDoc.GetValues("full_name")[0], foundDoc.GetField("year").GetInt32Value().ToString());
}
}
}
}
}
