public void Add_Words_To_Dictionary()
{
//Arrange
Dictionary <string, int> myWordDictionary = new Dictionary <string, int>();
List <string> myWordList = new List <string>()
{
"the", "railway", "children", "the", "children", "father", "the"
};
Dictionary <string, int> myExpectedWordDictionary = new Dictionary <string, int>();
myExpectedWordDictionary.Add("the", 3);
myExpectedWordDictionary.Add("children", 2);
myExpectedWordDictionary.Add("railway", 1);
myExpectedWordDictionary.Add("father", 1);
var myWordCountUtils = new WordCountUtils();
//Act
//Loop through words in word list
myWordCountUtils.WordCountDictionary(myWordList, myWordDictionary);
//Assert
CollectionAssert.AreEquivalent(myExpectedWordDictionary, myWordDictionary);
CollectionAssert.AreEqual(myExpectedWordDictionary, myWordDictionary);
}
/// <summary>
/// Get word count from input based on options selected
///
/// </summary>
/// <param name="input">Input can be either URL or Text</param>
/// <param name="isFilterStopWord">set to true will filterout pre-defined stop word</param>
/// <param name="options"> <list type="bullet"> <item> 0: calculates number of occurences of each word on the page </item> <item> 1: calculates number of occurences of each word on the page meta tags. </item> <item> 2: calculates number of external links in the text.</item></list > </param>
/// <param name="sortByColumn"><para>1: sort by word </para><para> 1: sort by count</para></param>
/// <param name="sortOrder"><para>0:ASC</para><para> 1: DESC</para></param>
public WordCountResponse Get(string input, bool isFilterStopWord, OptionsEnum options, int sortByColumn = 2, SortEnum sortOrder = SortEnum.DESC)
{
string content = "";
Dictionary <string, int> result = new Dictionary <string, int>();
Dictionary <string, int> resultSorted = new Dictionary <string, int>();
if (!isFilterStopWord)
{
stopWord = new List <string>();
}
try
{
bool isUri = Uri.IsWellFormedUriString(input, UriKind.RelativeOrAbsolute);
content = (isUri) ? WordCountUtils.LoadContentFromURL(input) : input;
switch (options)
{
case OptionsEnum.COUNTLINKS:
{
result = WordCountUtils.GetLinkCount(content);
break;
}
case OptionsEnum.COUNTWORDINMETA:
{
result = WordCountUtils.GetMetaTextCount(content, stopWord);
break;
}
case OptionsEnum.COUNTWORD:
default:
{
result = WordCountUtils.GetWordCount(content, stopWord);
break;
}
}
resultSorted = WordCountUtils.SortResult(result, sortByColumn, sortOrder);
}
catch (Exception ex)
{
//log it
throw new HttpResponseException(new HttpResponseMessage(HttpStatusCode.InternalServerError)
{
Content = new StringContent(ex.Message),
ReasonPhrase = "Something went wrong"
});
}
res.Source = input;
res.IsFilterStopWord = isFilterStopWord;
res.SelectedOption = options;
res.result = resultSorted;
return(res);
}
public void SplitString_By_Space_And_RemoveEmptyEntries()
{
//Arrange
string inputText = "the railway children they ";
string[] outputArray = { "the", "railway", "children", "they" };
var myWordCountUtils = new WordCountUtils();
//Act
var myWordList = myWordCountUtils.SplitWords(inputText);
//List<string> wordList = inputText.Split(new []{' '},StringSplitOptions.RemoveEmptyEntries).ToList();
//Assert
CollectionAssert.AreEqual(outputArray, myWordList);
}
public void Remove_Punctuation()
{
//Arrange
string inputText = "CHILDREN!\r\n\r\nBy E. () [] ^ _ - ;\r\n\r\n\r\n I. The beginning of things.\r\n";
string expectedOutputText = "CHILDREN\r\n\r\nBy E ^ \r\n\r\n\r\n I The beginning of things\r\n";
string outputText = "";
var myWordCountUtils = new WordCountUtils();
//Act
//outputText = new string(inputText.Where(c => !char.IsPunctuation(c)).ToArray());
outputText = myWordCountUtils.RemovePunctuation(inputText);
//Assert
Assert.AreEqual(expectedOutputText, outputText);
}
public void Remove_RomanNumerals_FollowedBy_FullStop()
{
//Arrange
// string pattern = @"(?=[MDCLXVI])M*(C[MD]|D?C{0,3})(X[CL]|L?X{0,3})(I[XV]|V?I{0,3}[.])";
string inputText = "I. The beginning of things.II. Peter's coal-mine. I don't suppose they had...";
string expectedOutput = " The beginning of things. Peter's coal-mine. I don't suppose they had...";
var myWordCountUtils = new WordCountUtils();
var myTextBook = new TextBook();
//Act
//inputText = System.Text.RegularExpressions.Regex.Replace(inputText, pattern, " ");
inputText = myWordCountUtils.RegexCleaner(inputText, myTextBook.Pattern);
//Assert
Assert.AreEqual(expectedOutput, inputText);
}
public void Remove_Specific_Characters()
{
//Arrange
string inputText = "CHILDREN\r\n\r\nBy E ^ \r\n\r\n\r\n I The beginning of things\r\n";
// Define characters to strip from the input and do it
string[] stripChars = { "^", "0", "1", "2", "3", "4", "5", "6", "7", "8", "9", "\n", "\t", "\r" };
string expectedOutputText = "CHILDREN By E I The beginning of things ";
var myWordCountUtils = new WordCountUtils();
//Act
inputText = myWordCountUtils.CleanDistinctCharacters(inputText, stripChars);
//Assert
Assert.AreEqual(expectedOutputText, inputText);
}
//Clear output and instantiate objects
private void resetValues()
{
txtOutput.Clear();//Clear output content
myTextBook = new TextBook();
myWordCountUtils = new WordCountUtils();
}