public void IgnoreAllWord()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.Text = "this is a tst of a tst errr";
_SpellChecker.SpellCheck();
Assert.True(3 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("tst" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.IgnoreAllWord();
_SpellChecker.SpellCheck();
Assert.True(7 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("errr" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
}
public void IgnoreAllWord()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.Text = "this is a tst of a tst errr";
_SpellChecker.SpellCheck();
Assertion.AssertEquals("Incorrect WordOffset", 3, _SpellChecker.WordIndex);
Assertion.AssertEquals("Incorrect CurrentWord", "tst", _SpellChecker.CurrentWord);
_SpellChecker.IgnoreAllWord();
_SpellChecker.SpellCheck();
Assertion.AssertEquals("Incorrect WordOffset", 7, _SpellChecker.WordIndex);
Assertion.AssertEquals("Incorrect CurrentWord", "errr", _SpellChecker.CurrentWord);
}
public void DeleteWord()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.Text = "this is is a tst.";
_SpellChecker.SpellCheck();
Assert.True(2 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("is" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
// basic delete test
_SpellChecker.DeleteWord();
Assert.True("this is a tst." == _SpellChecker.Text, "Incorrect Text");
_SpellChecker.SpellCheck();
Assert.True(3 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("tst" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
// before punctuation delete test
_SpellChecker.DeleteWord();
Assert.True("this is a." == _SpellChecker.Text, "Incorrect Text");
_SpellChecker.Text = "Becuase people are realy bad spelers";
_SpellChecker.SpellCheck();
Assert.True(0 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("Becuase" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
//delete first word test
_SpellChecker.DeleteWord();
Assert.True("people are realy bad spelers" == _SpellChecker.Text, "Incorrect Text");
//_SpellChecker.SpellCheck();
//Assert.True(2 == _SpellChecker.WordIndex, "Incorrect WordOffset");
//Assert.True("realy" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
////delete first word test
//_SpellChecker.DeleteWord();
//Assert.True("people are bad spelers" == _SpellChecker.Text, "Incorrect Text");
//_SpellChecker.SpellCheck();
//Assert.True(3 == _SpellChecker.WordIndex, "Incorrect WordOffset");
//Assert.True("spelers" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
////delete last word test
//_SpellChecker.DeleteWord();
//Assert.True("people are bad" == _SpellChecker.Text, "Incorrect Text");
}
public void DeleteWord()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.Text = "this is is a tst.";
_SpellChecker.SpellCheck();
Assert.AreEqual(2, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("is", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
// basic delete test
_SpellChecker.DeleteWord();
Assert.AreEqual("this is a tst.", _SpellChecker.Text, "Incorrect Text");
_SpellChecker.SpellCheck();
Assert.AreEqual(3, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("tst", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
// before punctuation delete test
_SpellChecker.DeleteWord();
Assert.AreEqual("this is a.", _SpellChecker.Text, "Incorrect Text");
_SpellChecker.Text = "Becuase people are realy bad spelers";
_SpellChecker.SpellCheck();
Assert.AreEqual(0, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("Becuase", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
//delete first word test
_SpellChecker.DeleteWord();
Assert.AreEqual("people are realy bad spelers", _SpellChecker.Text, "Incorrect Text");
_SpellChecker.SpellCheck();
Assert.AreEqual(2, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("realy", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
//delete first word test
_SpellChecker.DeleteWord();
Assert.AreEqual("people are bad spelers", _SpellChecker.Text, "Incorrect Text");
_SpellChecker.SpellCheck();
Assert.AreEqual(3, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("spelers", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
//delete last word test
_SpellChecker.DeleteWord();
Assert.AreEqual("people are bad", _SpellChecker.Text, "Incorrect Text");
}
public void CheckSpelling()
{
_customUnderlines.IllFormedLines.Clear();
_customUnderlines.Lines.Clear();
//Do not check spelling of watermark text
if (!IsWatermarkShowing)
{
try
{
if (_spelling != null && TextBox.Text.Length < 5000)
{
_spelling.Text = TextBox.Text;
_spelling.ShowDialog = false;
if (File.Exists(_spelling.Dictionary.DictionaryFile))
{
_spelling.SpellCheck();
}
}
}
catch (Exception ex)
{
Trace.WriteLine(ex);
}
MarkLines();
}
TextBox.Refresh();
}
public void NoText()
{
Spelling _SpellChecker = NewSpellChecker();
Assert.AreEqual(string.Empty, _SpellChecker.CurrentWord, "Incorrect Current Word");
_SpellChecker.WordIndex = 1;
Assert.AreEqual(0, _SpellChecker.WordIndex, "Incorrect Word Index");
Assert.AreEqual(0, _SpellChecker.WordCount, "Incorrect Word Count");
Assert.AreEqual(0, _SpellChecker.TextIndex, "Incorrect Text Index");
_SpellChecker.DeleteWord();
Assert.AreEqual(string.Empty, _SpellChecker.Text, "Incorrect Text");
_SpellChecker.IgnoreWord();
Assert.AreEqual(string.Empty, _SpellChecker.Text, "Incorrect Text");
_SpellChecker.ReplaceWord("Test");
Assert.AreEqual(string.Empty, _SpellChecker.Text, "Incorrect Text");
Assert.IsFalse(_SpellChecker.SpellCheck(), "Spell Check not false");
_SpellChecker.Suggest();
Assert.AreEqual(0, _SpellChecker.Suggestions.Count, "Generated Suggestions with no text");
}
public void SpellCheck()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.Text = "this is an errr tst";
_SpellChecker.SpellCheck();
Assert.True(3 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("errr" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
}
public void SpellCheck()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.Text = "this is an errr tst";
_SpellChecker.SpellCheck();
Assertion.AssertEquals("Incorrect WordOffset", 3, _SpellChecker.WordIndex);
Assertion.AssertEquals("Incorrect CurrentWord", "errr", _SpellChecker.CurrentWord);
}
public void HtmlSpellCheck()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.IgnoreHtml = true;
_SpellChecker.Text = "<a href=\"#\">this <span id=\"txt\">is</span> an errr tst</a>";
_SpellChecker.SpellCheck();
Assertion.AssertEquals("Incorrect WordOffset", 9, _SpellChecker.WordIndex);
Assertion.AssertEquals("Incorrect CurrentWord", "errr", _SpellChecker.CurrentWord);
}
public void HtmlSpellCheck()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.IgnoreHtml = true;
_SpellChecker.Text = "<a href=\"#\">this <span id=\"txt\">is</span> an errr tst</a>";
_SpellChecker.SpellCheck();
Assert.True(9 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("errr" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
}
public void Suggest()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.Text = "this is tst";
_SpellChecker.SpellCheck();
Assert.True(2 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("tst" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.Suggest();
Assert.True(25 == _SpellChecker.Suggestions.Count, "Incorrect Suggestion Count");
Assert.True(true == _SpellChecker.Suggestions.Contains("test"), "Could not find 'test' in suggestions");
}
public void Suggest()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.Text = "this is tst";
_SpellChecker.SpellCheck();
Assertion.AssertEquals("Incorrect WordOffset", 2, _SpellChecker.WordIndex);
Assertion.AssertEquals("Incorrect CurrentWord", "tst", _SpellChecker.CurrentWord);
_SpellChecker.Suggest();
Assertion.AssertEquals("Incorrect Suggestion Count", 25, _SpellChecker.Suggestions.Count);
Assertion.AssertEquals("Could not find 'test' in suggestions", true, _SpellChecker.Suggestions.Contains("test"));
}
protected void Page_Load(object sender, EventArgs e)
{
wordCounterString = GetString("SpellCheck.wordCounterString");
ScriptHelper.RegisterWOpenerScript(Page);
ScriptHelper.RegisterScriptFile(Page, "Controls/MessagesPlaceHolder.js");
ScriptHelper.RegisterSpellChecker(Page);
// Control initialization
IgnoreButton.Text = GetString("SpellCheck.IgnoreButton");
IgnoreAllButton.Text = GetString("SpellCheck.IgnoreAllButton");
ReplaceButton.Text = GetString("SpellCheck.ReplaceButton");
ReplaceAllButton.Text = GetString("SpellCheck.ReplaceAllButton");
StatusText.Text = GetString("SpellCheck.StatusText");
RemoveButton.Text = GetString("general.remove");
headTitle.ResourceString = "SpellCheck.lblNotInDictionary";
lblChangeTo.Text = GetString("SpellCheck.lblChangeTo");
lblSuggestions.Text = GetString("SpellCheck.lblSuggestions");
if (mDictExists)
{
// Add client side events
Suggestions.Attributes.Add("onChange", "javascript: changeWord(this);");
// Load spell checker settings
LoadValues();
switch (SpellMode.Value)
{
case "start":
EnableButtons();
SpellChecker.SpellCheck();
break;
case "suggest":
EnableButtons();
break;
default:
DisableButtons();
break;
}
}
}
public void CheckSpelling()
{
_customUnderlines.IllFormedLines.Clear();
_customUnderlines.Lines.Clear();
try
{
if (_spelling != null)
{
_spelling.Text = TextBox.Text;
_spelling.ShowDialog = false;
if (File.Exists(_spelling.Dictionary.DictionaryFile))
{
_spelling.SpellCheck();
}
}
}
catch (Exception ex)
{
Trace.WriteLine(ex);
}
MarkLines();
TextBox.Refresh();
}
static void Main(string[] args)
{
// start connection with database
SqlConnection sqlConnection = new SqlConnection("Data Source=AHMEDFATHY-PC;Initial Catalog=newDB;Integrated Security=True; MultipleActiveResultSets=true");
sqlConnection.Open();
// select statment to retrieve everything from database
string queryString = "SELECT * FROM crawler_Table";
SqlCommand cmd = new SqlCommand(queryString, sqlConnection);
// declare variable from reader to read from database (all the content from database)
SqlDataReader rdr = cmd.ExecuteReader();
int counterofopages = 0; // counter for number of pages that i read it from database (at least 1500)
// datastructure to save term and doc_id and frequency and list of positions for this term
List <KeyValuePair <string, KeyValuePair <int[], List <int> > > > indexmap = new List <KeyValuePair <string, KeyValuePair <int[], List <int> > > >();
// while loop to read row by row from the reader
while (rdr.Read())
{
// this condition to break from loop when take at least 1500 page
if (counterofopages == 1600)
{
break;
}
// try and catch to throw any exceptions out if it retreive null from innertext or something else
int boolll = 0; // boolean to check if the inner text has exception change boolean = 1 and skip tha link
try
{
//===================================================//
// retreive from each row docid , url (link) , content of the page (html page)
int doc_id = (int)rdr["doc_id"];
string url = (string)rdr["URL"];
string content = (string)rdr["Page_Content"];
//===================================================//
// pasre html page from database and get the inner text (step 1)
IHTMLDocument2 myDoc = new HTMLDocumentClass();
myDoc.write(content);
string elements = myDoc.body.innerText;
//===================================================//
//(it will be)
/// split in (step 2) (to take tokens and save it in array of strings named (tokens)
string[] tokens = elements.Split(',', ' ', '.', ':', '\t', '\n', '\r');
int i = 0; // counter to calculate the position for every term
// check if any string it will be null or empty
tokens = tokens.Where(x => !string.IsNullOrEmpty(x)).ToArray();
//===================================================//
/// saves every term and its list (positions) (s in dictionary named (termsandpos) before removing stop words
Dictionary <string, List <int> > termsandpos = new Dictionary <string, List <int> >();
foreach (var words in tokens)
{
List <int> listofpos = new List <int>();
i++;
// using regex to remove punctuation characters from every word (step 3) -> req 1
string word = Regex.Replace(words, @"[^\w\d\s]", "");
word = Regex.Replace(word, @"\d", "");
// if the word is empty after removing punctuation characters continues and don't save it
if (word == "")
{
continue;
}
// using spelling class from netspell reference and create object from it and using it to check if this word is real word in english or not.
Spelling ss = new Spelling();
// when the object from spelling class is used , the dialog window will opened and has many feature and i will closed by using next line to continue my run it's not used for my code.
ss.ShowDialog = false;
// check if this word is not found in dictionary in the spell library , continue ( go to the next word).
// esle continue the rest of the code (that is mean the word is found in the dictionary).
if (ss.SpellCheck(word))
{
continue;
}
word = word.ToLower(); //case folding in (step 3) -> req 2
//If the word is already existed ,add the new position in the list of this word
if (termsandpos.ContainsKey(word))
{
listofpos = termsandpos[word];
listofpos.Add(i);
termsandpos[word] = listofpos;
}
// else, add the word and the first position
else
{
listofpos.Add(i);
termsandpos.Add(word, listofpos);
}
}
//===================================================//
///// stop words removing in (step 3) -> req 3
/// list of stop words
/// create anthor dictinary to copy all terms without stop words
Dictionary <string, List <int> > temp = new Dictionary <string, List <int> >();
List <string> stopwords = new List <string>()
{
"a", "about", "above", "above", "across", "after", "afterwards", "again", "against", "all", "almost", "alone", "along", "already", "also", "although", "always", "am", "among", "amongst", "amoungst", "amount", "an", "and", "another", "any", "anyhow", "anyone", "anything", "anyway", "anywhere", "are", "around", "as", "at", "back", "be", "became", "because", "become", "becomes", "becoming", "been", "before", "beforehand", "behind", "being", "below", "beside", "besides", "between", "beyond", "bill", "both", "bottom", "but", "by", "call", "can", "cannot", "cant", "co", "con", "could", "couldnt", "cry", "de", "describe", "detail", "do", "done", "down", "due", "during", "each", "eg", "eight", "either", "eleven", "else", "elsewhere", "empty", "enough", "etc", "even", "ever", "every", "everyone", "everything", "everywhere", "except", "few", "fifteen", "fify", "fill", "find", "fire", "first", "five", "for", "former", "formerly", "forty", "found", "four", "from", "front", "full", "further", "get", "give", "go", "had", "has", "hasnt", "have", "he", "hence", "her", "here", "hereafter", "hereby", "herein", "hereupon", "hers", "herself", "him", "himself", "his", "how", "however", "hundred", "ie", "if", "in", "inc", "indeed", "interest", "into", "is", "it", "its", "itself", "keep", "last", "latter", "latterly", "least", "less", "ltd", "made", "many", "may", "me", "meanwhile", "might", "mill", "mine", "more", "moreover", "most", "mostly", "move", "much", "must", "my", "myself", "name", "namely", "neither", "never", "nevertheless", "next", "nine", "no", "nobody", "none", "noone", "nor", "not", "nothing", "now", "nowhere", "of", "off", "often", "on", "once", "one", "only", "onto", "or", "other", "others", "otherwise", "our", "ours", "ourselves", "out", "over", "own", "part", "per", "perhaps", "please", "put", "rather", "re", "same", "see", "seem", "seemed", "seeming", "seems", "serious", "several", "she", "should", "show", "side", "since", "sincere", "six", "sixty", "so", "some", "somehow", "someone", "something", "sometime", "sometimes", "somewhere", "still", "such", "system", "take", "ten", "than", "that", "the", "their", "them", "themselves", "then", "thence", "there", "thereafter", "thereby", "therefore", "therein", "thereupon", "these", "they", "thick", "thin", "third", "this", "those", "though", "three", "through", "throughout", "thru", "thus", "to", "together", "too", "top", "toward", "towards", "twelve", "twenty", "two", "un", "under", "until", "up", "upon", "us", "very", "via", "was", "we", "well", "were", "what", "whatever", "when", "whence", "whenever", "where", "whereafter", "whereas", "whereby", "wherein", "whereupon", "wherever", "whether", "which", "while", "whither", "who", "whoever", "whole", "whom", "whose", "why", "will", "with", "within", "without", "would", "yet", "you", "your", "yours", "yourself", "yourselves", "the"
};
for (int f = 0; f < termsandpos.Count; f++)
{
// if the term is already existed in the stopwords list or the term is a single character like ii or i , continue (and go to the next term).
if (stopwords.Contains(termsandpos.Keys.ElementAt(f)) || termsandpos.Keys.ElementAt(f).Length <= 2)
{
continue;
}
// else ,that's mean the term is not a stop word then add it and its positions in the temp dictionary.
else
{
List <int> copyofpositions = new List <int>();
copyofpositions = termsandpos[termsandpos.Keys.ElementAt(f)];
temp.Add(termsandpos.Keys.ElementAt(f), copyofpositions);
}
}
//===================================================//
//// al stemming algorithm (step 3) --> req 4
var stemmer = new PorterStemmer(); // declare object from claas of porterstemmer algorithm
Dictionary <string, List <int> > finalterm = new Dictionary <string, List <int> >();
foreach (KeyValuePair <string, List <int> > iter1 in temp)
{
//===================================================//
// add every term and its docid in table called (TermsBStemming_Table) in db before stemming (the note in step 3 -->req 4)
string insertString3 = "INSERT INTO TermsBStemming_Table (termBstemming,docID) VALUES (@termBstemming,@docID)";
SqlCommand cmd3 = new SqlCommand(insertString3, sqlConnection);
SqlParameter par1 = new SqlParameter("@termBstemming", iter1.Key);
SqlParameter par2 = new SqlParameter("@docID", doc_id);
cmd3.Parameters.Add(par1);
cmd3.Parameters.Add(par2);
cmd3.ExecuteNonQuery();
//===================================================//
List <int> listofpositions = new List <int>();
// called function (StemWord) and send the term and return term after stemming
string stem = stemmer.StemWord(iter1.Key);
// check if this stem is already existed in finalterm dictionary (the new datastructure to save the term and its list after stemmnig)
if (finalterm.ContainsKey(stem))
{
List <int> tempforsimlir = new List <int>();
tempforsimlir = finalterm[stem]; // take the list of positions for this term (old positions added before for this term)
listofpositions = temp[iter1.Key]; // take the list of new positions for this term
/// added the new positions and old position in one list
for (int j = 0; j < listofpositions.Count; j++)
{
tempforsimlir.Add(listofpositions[j]);
}
// and save it again for the term
finalterm[stem] = tempforsimlir;
}
// addd the term ans its list to finalterm dictionary
else
{
listofpositions = temp[iter1.Key];
finalterm.Add(stem, listofpositions);
}
}
//===================================================//
//// inverted index (step 4)
foreach (KeyValuePair <string, List <int> > iter in finalterm)
{
int freq = iter.Value.Count; // calculate freq through count number of positions
int[] arr = new int[2]; // save in this array doc id and the frequency
arr[0] = doc_id;
arr[1] = freq;
// convert list of the positions for every term to string
var resultofpositions = string.Join(", ", iter.Value);
//===================================================//
// save term and docid ans=d frequency and (list of positions as string ) in table called Inverted_Index in db.
string insertString2 = "INSERT INTO Inverted_Index (Term,DocID,Frequency,position) VALUES (@Term,@DocID,@Frequency,@position)";
SqlCommand cmd2 = new SqlCommand(insertString2, sqlConnection);
SqlParameter paramter1 = new SqlParameter("@Term", iter.Key);
SqlParameter paramter2 = new SqlParameter("@DocID", doc_id);
SqlParameter paramter3 = new SqlParameter("@Frequency", freq);
SqlParameter paramter4 = new SqlParameter("@position", resultofpositions);
cmd2.Parameters.Add(paramter1);
cmd2.Parameters.Add(paramter2);
cmd2.Parameters.Add(paramter3);
cmd2.Parameters.Add(paramter4);
cmd2.ExecuteNonQuery();
//===================================================//
/// store in index list term and arrof ints (arr[0]=docid,arr[1] = freqs of every term) and list of all positions of this term (if i needed in ranks or something else).
indexmap.Add(new KeyValuePair <string, KeyValuePair <int[], List <int> > >(iter.Key, new KeyValuePair <int[], List <int> >(arr, iter.Value)));
}
//===================================================//
}
//===================================================//
//catch any type of exception and change the boolean that i decalred equal zero
catch (NullReferenceException ex)
{
boolll = 1;
Console.WriteLine(ex.Message);
}
catch (ArgumentOutOfRangeException exx)
{
boolll = 1;
Console.WriteLine(exx.Message);
}
// if the boolean became equal 1 , then leave this link and go to anthor link
if (boolll == 1)
{
continue;
}
//===================================================//
/// to count number of pages (at least 1500 page)
counterofopages++;
//===================================================//
}
//===================================================//
// close the reader from database
rdr.Close();
/// close the connection
sqlConnection.Close();
//===================================================//
}
public void TestEvents()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.Text = "ths is is a tst.";
ResetEvents();
_SpellChecker.SpellCheck();
//spelling check
Assert.AreEqual(0, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("ths", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
// event check
Assert.AreEqual(EventNames.MisspelledWord, _lastEvent, "Incorrect Event");
Assert.IsNotNull(_lastSpellingEvent, "Event not fired");
Assert.AreEqual(0, _lastSpellingEvent.WordIndex, "Incorrect Event Word Index");
Assert.AreEqual(0, _lastSpellingEvent.TextIndex, "Incorrect Event Text Index");
Assert.AreEqual("ths", _lastSpellingEvent.Word, "Incorrect Event Word");
ResetEvents();
_SpellChecker.ReplaceWord("this");
//spelling check
Assert.AreEqual("this is is a tst.", _SpellChecker.Text, "Incorrect Text");
// event check
Assert.AreEqual(EventNames.ReplacedWord, _lastEvent, "Incorrect Event");
Assert.IsNotNull(_lastReplaceEvent, "Null Event object fired");
Assert.AreEqual(0, _lastReplaceEvent.WordIndex, "Incorrect Event Word Index");
Assert.AreEqual(0, _lastReplaceEvent.TextIndex, "Incorrect Event Text Index");
Assert.AreEqual("ths", _lastReplaceEvent.Word, "Incorrect Event Word");
Assert.AreEqual("this", _lastReplaceEvent.ReplacementWord, "Incorrect Event Replacement Word");
ResetEvents();
_SpellChecker.SpellCheck();
//spelling check
Assert.AreEqual(2, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("is", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
// event check
Assert.AreEqual(EventNames.DoubledWord, _lastEvent, "Incorrect Event");
Assert.IsNotNull(_lastSpellingEvent, "Null Event object fired");
Assert.AreEqual(2, _lastSpellingEvent.WordIndex, "Incorrect Event Word Index");
Assert.AreEqual(8, _lastSpellingEvent.TextIndex, "Incorrect Event Text Index");
Assert.AreEqual("is", _lastSpellingEvent.Word, "Incorrect Event Word");
ResetEvents();
_SpellChecker.DeleteWord();
//spelling check
Assert.AreEqual("this is a tst.", _SpellChecker.Text, "Incorrect Text");
// event check
Assert.AreEqual(EventNames.DeletedWord, _lastEvent, "Incorrect Event");
Assert.IsNotNull(_lastSpellingEvent, "Null Event object fired");
Assert.AreEqual(2, _lastSpellingEvent.WordIndex, "Incorrect Event Word Index");
Assert.AreEqual(8, _lastSpellingEvent.TextIndex, "Incorrect Event Text Index");
Assert.AreEqual("is ", _lastSpellingEvent.Word, "Incorrect Event Word");
ResetEvents();
_SpellChecker.SpellCheck();
//spelling check
Assert.AreEqual(3, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("tst", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
// event check
Assert.AreEqual(EventNames.MisspelledWord, _lastEvent, "Incorrect Event");
Assert.IsNotNull(_lastSpellingEvent, "Null Event object fired");
Assert.AreEqual(3, _lastSpellingEvent.WordIndex, "Incorrect Event Word Index");
Assert.AreEqual(10, _lastSpellingEvent.TextIndex, "Incorrect Event Text Index");
Assert.AreEqual("tst", _lastSpellingEvent.Word, "Incorrect Event Word");
ResetEvents();
_SpellChecker.IgnoreWord();
//spelling check
Assert.AreEqual("this is a tst.", _SpellChecker.Text, "Incorrect Text");
// event check
Assert.AreEqual(EventNames.IgnoredWord, _lastEvent, "Incorrect Event");
Assert.IsNotNull(_lastSpellingEvent, "Null Event object fired");
Assert.AreEqual(3, _lastSpellingEvent.WordIndex, "Incorrect Event Word Index");
Assert.AreEqual(10, _lastSpellingEvent.TextIndex, "Incorrect Event Text Index");
Assert.AreEqual("tst", _lastSpellingEvent.Word, "Incorrect Event Word");
ResetEvents();
_SpellChecker.SpellCheck();
// event check
Assert.AreEqual(EventNames.EndOfText, _lastEvent, "Incorrect Event");
}
public void TestEvents()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.Text = "ths is is a tst.";
ResetEvents();
_SpellChecker.SpellCheck();
//spelling check
Assert.True(0 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("ths" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
// event check
Assert.True(EventNames.MisspelledWord == _lastEvent, "Incorrect Event");
Assert.NotNull(_lastSpellingEvent); //, "Event not fired");
Assert.True(0 == _lastSpellingEvent.WordIndex, "Incorrect Event Word Index");
Assert.True(0 == _lastSpellingEvent.TextIndex, "Incorrect Event Text Index");
Assert.True("ths" == _lastSpellingEvent.Word, "Incorrect Event Word");
ResetEvents();
_SpellChecker.ReplaceWord("this");
//spelling check
Assert.True("this is is a tst." == _SpellChecker.Text, "Incorrect Text");
// event check
Assert.True(EventNames.ReplacedWord == _lastEvent, "Incorrect Event");
Assert.NotNull(_lastReplaceEvent); //, "Null Event object fired");
Assert.True(0 == _lastReplaceEvent.WordIndex, "Incorrect Event Word Index");
Assert.True(0 == _lastReplaceEvent.TextIndex, "Incorrect Event Text Index");
Assert.True("ths" == _lastReplaceEvent.Word, "Incorrect Event Word");
Assert.True("this" == _lastReplaceEvent.ReplacementWord, "Incorrect Event Replacement Word");
ResetEvents();
_SpellChecker.SpellCheck();
//spelling check
Assert.True(2 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("is" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
// event check
Assert.True(EventNames.DoubledWord == _lastEvent, "Incorrect Event");
Assert.NotNull(_lastSpellingEvent); //, "Null Event object fired");
Assert.True(2 == _lastSpellingEvent.WordIndex, "Incorrect Event Word Index");
Assert.True(8 == _lastSpellingEvent.TextIndex, "Incorrect Event Text Index");
Assert.True("is" == _lastSpellingEvent.Word, "Incorrect Event Word");
ResetEvents();
_SpellChecker.DeleteWord();
//spelling check
Assert.True("this is a tst." == _SpellChecker.Text, "Incorrect Text");
// event check
Assert.True(EventNames.DeletedWord == _lastEvent, "Incorrect Event");
Assert.NotNull(_lastSpellingEvent); //, "Null Event object fired");
Assert.True(2 == _lastSpellingEvent.WordIndex, "Incorrect Event Word Index");
Assert.True(8 == _lastSpellingEvent.TextIndex, "Incorrect Event Text Index");
Assert.True("is " == _lastSpellingEvent.Word, "Incorrect Event Word");
ResetEvents();
_SpellChecker.SpellCheck();
//spelling check
Assert.True(3 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("tst" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
// event check
Assert.True(EventNames.MisspelledWord == _lastEvent, "Incorrect Event");
Assert.NotNull(_lastSpellingEvent); //, "Null Event object fired");
Assert.True(3 == _lastSpellingEvent.WordIndex, "Incorrect Event Word Index");
Assert.True(10 == _lastSpellingEvent.TextIndex, "Incorrect Event Text Index");
Assert.True("tst" == _lastSpellingEvent.Word, "Incorrect Event Word");
ResetEvents();
_SpellChecker.IgnoreWord();
//spelling check
Assert.True("this is a tst." == _SpellChecker.Text, "Incorrect Text");
// event check
Assert.True(EventNames.IgnoredWord == _lastEvent, "Incorrect Event");
Assert.NotNull(_lastSpellingEvent); //, "Null Event object fired");
Assert.True(3 == _lastSpellingEvent.WordIndex, "Incorrect Event Word Index");
Assert.True(10 == _lastSpellingEvent.TextIndex, "Incorrect Event Text Index");
Assert.True("tst" == _lastSpellingEvent.Word, "Incorrect Event Word");
ResetEvents();
_SpellChecker.SpellCheck();
// event check
Assert.True(EventNames.EndOfText == _lastEvent, "Incorrect Event");
}
public void ReplaceWord()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.Text = "ths is an errr tst";
_SpellChecker.SpellCheck();
Assert.True(0 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("ths" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.ReplacementWord = "this";
_SpellChecker.ReplaceWord();
Assert.True("this is an errr tst" == _SpellChecker.Text, "Incorrect Text");
//replace with empty string
_SpellChecker.SpellCheck();
Assert.True(3 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("errr" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.ReplaceWord("");
Assert.True("this is an tst" == _SpellChecker.Text, "Incorrect Text");
_SpellChecker.Text = "Becuase people are realy bad spelers, \r\nths produc was desinged to prevent spelling errors in a text area like ths.";
_SpellChecker.SpellCheck();
Assert.True(0 == _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.True("Becuase" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.ReplaceWord("because");
Assert.True("Because people are realy bad spelers, \r\nths produc was desinged to prevent spelling errors in a text area like ths." == _SpellChecker.Text, "Incorrect Text");
//_SpellChecker.SpellCheck();
//Assert.True(3 == _SpellChecker.WordIndex, "Incorrect WordOffset");
//Assert.True("realy" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
//_SpellChecker.ReplaceWord("really");
//Assert.True("Because people are really bad spelers, \r\nths produc was desinged to prevent spelling errors in a text area like ths." == _SpellChecker.Text, "Incorrect Text");
//_SpellChecker.SpellCheck();
//Assert.True(5 == _SpellChecker.WordIndex, "Incorrect WordOffset");
//Assert.True("spelers" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
//_SpellChecker.ReplaceWord("spellers");
//Assert.True("Because people are really bad spellers, \r\nths produc was desinged to prevent spelling errors in a text area like ths." == _SpellChecker.Text, "Incorrect Text");
//_SpellChecker.SpellCheck();
//Assert.True(6 == _SpellChecker.WordIndex, "Incorrect WordOffset");
//Assert.True("ths" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
//_SpellChecker.ReplaceWord("this");
//Assert.True("Because people are really bad spellers, \r\nthis produc was desinged to prevent spelling errors in a text area like ths." == _SpellChecker.Text, "Incorrect Text");
//_SpellChecker.SpellCheck();
//Assert.True(7 == _SpellChecker.WordIndex, "Incorrect WordOffset");
//Assert.True("produc" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
//_SpellChecker.ReplaceWord("product");
//Assert.True("Because people are really bad spellers, \r\nthis product was desinged to prevent spelling errors in a text area like ths." == _SpellChecker.Text, "Incorrect Text");
//_SpellChecker.SpellCheck();
//Assert.True(9 == _SpellChecker.WordIndex, "Incorrect WordOffset");
//Assert.True("desinged" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
//_SpellChecker.ReplaceWord("designed");
//Assert.True("Because people are really bad spellers, \r\nthis product was designed to prevent spelling errors in a text area like ths." == _SpellChecker.Text, "Incorrect Text");
//_SpellChecker.SpellCheck();
//Assert.True(19 == _SpellChecker.WordIndex, "Incorrect WordOffset");
//Assert.True("ths" == _SpellChecker.CurrentWord, "Incorrect CurrentWord");
//_SpellChecker.ReplaceWord("this");
//Assert.True("Because people are really bad spellers, \r\nthis product was designed to prevent spelling errors in a text area like this." == _SpellChecker.Text, "Incorrect Text");
}
public void ReplaceWord()
{
Spelling _SpellChecker = NewSpellChecker();
_SpellChecker.Text = "ths is an errr tst";
_SpellChecker.SpellCheck();
Assert.AreEqual(0, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("ths", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.ReplacementWord = "this";
_SpellChecker.ReplaceWord();
Assert.AreEqual("this is an errr tst", _SpellChecker.Text, "Incorrect Text");
//replace with empty string
_SpellChecker.SpellCheck();
Assert.AreEqual(3, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("errr", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.ReplaceWord("");
Assert.AreEqual("this is an tst", _SpellChecker.Text, "Incorrect Text");
_SpellChecker.Text = "Becuase people are realy bad spelers, \r\nths produc was desinged to prevent spelling errors in a text area like ths.";
_SpellChecker.SpellCheck();
Assert.AreEqual(0, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("Becuase", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.ReplaceWord("because");
Assert.AreEqual("Because people are realy bad spelers, \r\nths produc was desinged to prevent spelling errors in a text area like ths.", _SpellChecker.Text, "Incorrect Text");
_SpellChecker.SpellCheck();
Assert.AreEqual(3, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("realy", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.ReplaceWord("really");
Assert.AreEqual("Because people are really bad spelers, \r\nths produc was desinged to prevent spelling errors in a text area like ths.", _SpellChecker.Text, "Incorrect Text");
_SpellChecker.SpellCheck();
Assert.AreEqual(5, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("spelers", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.ReplaceWord("spellers");
Assert.AreEqual("Because people are really bad spellers, \r\nths produc was desinged to prevent spelling errors in a text area like ths.", _SpellChecker.Text, "Incorrect Text");
_SpellChecker.SpellCheck();
Assert.AreEqual(6, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("ths", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.ReplaceWord("this");
Assert.AreEqual("Because people are really bad spellers, \r\nthis produc was desinged to prevent spelling errors in a text area like ths.", _SpellChecker.Text, "Incorrect Text");
_SpellChecker.SpellCheck();
Assert.AreEqual(7, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("produc", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.ReplaceWord("product");
Assert.AreEqual("Because people are really bad spellers, \r\nthis product was desinged to prevent spelling errors in a text area like ths.", _SpellChecker.Text, "Incorrect Text");
_SpellChecker.SpellCheck();
Assert.AreEqual(9, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("desinged", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.ReplaceWord("designed");
Assert.AreEqual("Because people are really bad spellers, \r\nthis product was designed to prevent spelling errors in a text area like ths.", _SpellChecker.Text, "Incorrect Text");
_SpellChecker.SpellCheck();
Assert.AreEqual(19, _SpellChecker.WordIndex, "Incorrect WordOffset");
Assert.AreEqual("ths", _SpellChecker.CurrentWord, "Incorrect CurrentWord");
_SpellChecker.ReplaceWord("this");
Assert.AreEqual("Because people are really bad spellers, \r\nthis product was designed to prevent spelling errors in a text area like this.", _SpellChecker.Text, "Incorrect Text");
}
public void SuggestionRank()
{
string invalidFile = @"..\src\NetSpell.Tests\Data\SuggestionTest.txt";
// open file
FileStream fs = new FileStream(invalidFile, FileMode.Open, FileAccess.Read, FileShare.Read);
StreamReader sr = new StreamReader(fs, Encoding.UTF7);
int totalFound = 0;
int totalChecked = 0;
int totalFirst = 0;
int totalTopFive = 0;
int totalTopTen = 0;
int totalTopTwentyFive = 0;
Console.WriteLine("Misspelled\tCorrect\tPosition\tCount");
// read line by line
while (sr.Peek() >= 0)
{
string tempLine = sr.ReadLine().Trim();
if (tempLine.Length > 0)
{
string[] parts = tempLine.Split();
string misSpelled = parts[0];
string correctSpelled = parts[1];
if (parts.Length > 2)
{
correctSpelled += " " + parts[2];
}
bool found = false;
if (_SpellChecker.SpellCheck(misSpelled))
{
totalChecked++;
_SpellChecker.Suggest();
int position = 0;
foreach (string suggestion in _SpellChecker.Suggestions)
{
position++;
if (suggestion.ToLower() == correctSpelled.ToLower())
{
Console.WriteLine("{0}\t{1}\t{2}\t{3}",
misSpelled, correctSpelled, position.ToString(), _SpellChecker.Suggestions.Count.ToString());
found = true;
totalFound++;
if (position == 1)
{
totalFirst++;
}
else if (position <= 5)
{
totalTopFive++;
}
else if (position <= 10)
{
totalTopTen++;
}
else if (position <= 25)
{
totalTopTwentyFive++;
}
break;
}
}
if (!found)
{
if (_SpellChecker.Suggestions.Count > 0)
{
Console.WriteLine("{0}\t{1}\t{2}\t{3}",
misSpelled, correctSpelled, "0", _SpellChecker.Suggestions.Count.ToString());
}
else
{
Console.WriteLine("{0}\t{1}\t{2}\t{3}",
misSpelled, correctSpelled, "-1", "-1");
}
}
}
else
{
Console.WriteLine("{0}\t{1}\t{2}\t{3}",
misSpelled, correctSpelled, "1", "1");
}
}
}
totalTopFive += totalFirst;
totalTopTen += totalTopFive;
totalTopTwentyFive += totalTopTen;
Console.WriteLine("Total Tested\t{0}", totalChecked);
Console.WriteLine("Total Found\t{0}\t{1}%", totalFound, ((float)totalFound / (float)totalChecked * 100f));
Console.WriteLine("First Suggestions\t{0}\t{1}%", totalFirst, ((float)totalFirst / (float)totalChecked * 100f));
Console.WriteLine("Top 5 Suggestions\t{0}\t{1}%", totalTopFive, ((float)totalTopFive / (float)totalChecked * 100f));
Console.WriteLine("Top 10 Suggestions\t{0}\t{1}%", totalTopTen, ((float)totalTopTen / (float)totalChecked * 100f));
Console.WriteLine("Top 25 Suggestions\t{0}\t{1}%", totalTopTwentyFive, ((float)totalTopTwentyFive / (float)totalChecked * 100f));
sr.Close();
fs.Close();
}
public void SpellCheck()
{
_loggingController.Log(MessageType.information, " load spell check");
SpellChecker.Text = GetSelectedTextControl().Text;
SpellChecker.SpellCheck();
}