///////////////////////////////////////////buttons part 2
private void choose_btn_Click(object sender, EventArgs e)
{
Ranker.cleanList();
openFileDialog1.FileName = "";
openFileDialog1.ShowDialog();
List <string> qry = new List <string>();
if (openFileDialog1.FileName != "")
{
try
{
qry = Files.qureys(openFileDialog1.FileName);
}
catch (Exception)
{
MessageBox.Show("wrong file");
}
}
List <string> qid = new List <string>();
List <string> q = new List <string>();
for (int i = 0; i < qry.Count; i = i + 2)
{
qid.Add(qry[i]);
q.Add(qry[i + 1]);
}
Searcher.parseQuerys(qid, q, 50);
}
public static void parseQuery(string qryid, string qry, int num, bool show)
{
Parse p = new Parse();
List <string> words = new List <string>();
p.parse(words, new List <bool>(), qry);
//groupby
HashSet <string> hash = new HashSet <string>();
for (int i = 0; i < words.Count; i++)//optional groupby
{
hash.Add(words[i]);
}
List <string> docs = Ranker.getdocs(num, hash.ToArray());
for (int i = 0; i < docs.Count; i++)
{
Console.WriteLine(docs[i]);
}
Ranker.addTolist(qryid, docs);
if (show)
{
Searcher.show(docs);
}
}
/// <summary>
/// Searches for the specified query, returns a ranked List of Documents matching any word in the query.
/// </summary>
/// <param name="query">The query.</param>
/// <returns></returns>
public static List <Document> Search(String query)
{
//Separate words, remove punctiations,make lowercase
List <String> words = Semanter.Splitwords(query, ":").ToList();
//Obtains possible types searched by this query
HashSet <String> typesPossible = TypeChecker(words);
//Stem words and remove stopwords
//slower method words = words.Except(invt.Stopwords).ToList();
List <String> splitwords = new List <String>();
string stem;
foreach (string word in words)
{
stem = invt.Samantha.StemWord(word);
if (!(invt.Stopwords.Contains(stem)))
{
splitwords.Add(stem);
}
}
if (splitwords.Count == 0)
{
return(new List <Document>());
}
//search for documents
Dictionary <Document, Dictionary <string, List <int> > > Results = DocsFound(splitwords, typesPossible);
if (Results.Keys.Count < 2)
{
return(Results.Keys.ToList());
}
return(Ranker.RankQuery(splitwords, Results, invt.DocumentCount));
}
public static void calculate_w(string postlist)
{
//split term and post doc
string[] post = postlist.Split('=');
if (!Indexer.idf.ContainsKey(post[0]))
{
return;
}
//split by docs
string[] docs = post[1].Split('#');
docs[0] = docs[0].Substring(docs[0].IndexOf(">") + 1);
//split by args
char[] del = { ':', ',' };
double weight = 0;
string term = post[0];
double idf = Ranker.calculate_idf(term);
if (idf == -1)
{
return;
}
foreach (string i in docs)
{
if (i.Equals(""))
{ //end of postlist
return;
}
string[] doci = i.Split(del);
if (!Parse.DocDic.ContainsKey(doci[0]))
{
continue;
}
double tf = (int.Parse(doci[1]) + int.Parse(doci[2]));
tf = Ranker.calculate_itf(post[0], doci[0], tf);
if (tf == -1)
{
return;
}
weight = Math.Pow(idf * tf, 2);
//add the Wi to the relevant doc
string[] doctemp = Parse.DocDic[doci[0]].Split(',');
//if already exist add the weight
if (doctemp.Length == 4)
{
weight = weight + double.Parse(doctemp[3]);
string ans = doctemp[0] + "," + doctemp[1] + "," + doctemp[2] + "," + weight;
Parse.DocDic[doci[0]] = ans;
}
//first W-term of the doc
else
{
string ans = Parse.DocDic[doci[0]] + "," + weight;
Parse.DocDic[doci[0]] = ans;
}
}
}
private void save_qry_Click(object sender, EventArgs e)
{
browser.ShowDialog();
try
{
Ranker.writeList(@"C:\Users\aviv9_000\Desktop\all");
}
catch (Exception) {}
Ranker.cleanList();
}
private void reset_btn_Click(object sender, EventArgs e)
{
//if index doesnt exixt -> alert
Ranker.cleanList();
qry_txt.Text = "";
openFileDialog1.FileName = "";
Doc_Search_chk.Checked = false;
expantion_chk.Checked = false;
}
private void run_query_btn_Click(object sender, EventArgs e)
{
Ranker.cleanList();
//if index doesnt exixt -> alert
if (qry_txt.Text == "")
{
return;
}
if (Doc_Search_chk.Checked)
{
if (Parse.DocDic.ContainsKey(qry_txt.Text))
{
string s = Parse.DocDic[qry_txt.Text].Split(',')[0]; //name of file
string s2 = browseCurpus.Text; //corpus location
string docpath = s2 + "/corpus/" + s + "/" + s;
//read the file , find the doc, parse 5 sentences from the doc;
}
return;
}
Parse p = new Parse();
List <string> ans = new List <string>();
int back = 50;
if (expantion_chk.Checked && !qry_txt.Text.Contains(" "))
{
ans = Files.viki(qry_txt.Text);//need 50 docs instead of 70
}
if (ans.Count != 0)
{
back = 70;
}
Indexer.path = pathPost.Text;
string final = qry_txt.Text;
for (int i = 0; i < ans.Count; i++)
{
final += " " + ans[i];
}
ans.Clear();
Random rnd = new Random();
Searcher.parseQuery("" + rnd.Next(100000, 999999), final, back, true);
}
