private static Article ParseDocument(string text, SummarizerArguments args)
{
Dictionary rules = Dictionary.LoadFromFile(args.DictionaryLanguage);
Article article = new Article(rules);
article.ParseText(text);
return(article);
}
private static SummarizedDocument CreateSummarizedDocument(Article article, SummarizerArguments args)
{
SummarizedDocument sumDoc = new SummarizedDocument();
sumDoc.Concepts = article.Concepts;
foreach (Sentence sentence in article.Sentences)
{
if (sentence.Selected)
{
sumDoc.Sentences.Add(sentence.OriginalSentence);
}
}
return(sumDoc);
}
internal static void Highlight(Article article, SummarizerArguments args)
{
if (args.DisplayPercent == 0 && args.DisplayLines == 0)
{
return;
}
if (args.DisplayPercent == 0)
{
//get the highest scored n lines, without reordering the list.
SelectNumberOfSentences(article, args.DisplayLines);
}
else
{
SelectSentencesByPercent(article, args.DisplayPercent);
}
}
private void button2_Click(object sender, EventArgs e)
{
string text;
string word = textBox2.Text;
WebClient web = new WebClient();
HtmlAgilityPack.HtmlDocument Htmldoc = new HtmlAgilityPack.HtmlDocument();
Process.Start("https://en.wikipedia.org/wiki/" + word);
byte[] byteArray = web.DownloadData(new Uri("https://en.wikipedia.org/wiki/" + word));
Stream stream = new MemoryStream(byteArray);
Htmldoc.Load(stream);
FileStream fs = new FileStream("D:\\htmltext.pdf", FileMode.Create, FileAccess.Write, FileShare.None);
Document pdfDoc = new Document();
PdfWriter writer = PdfWriter.GetInstance(pdfDoc, fs);
pdfDoc.Open();
foreach (HtmlNode node in Htmldoc.DocumentNode.SelectNodes("//text()"))
{
text = node.InnerText.Trim();
pdfDoc.Add(new Paragraph(text));
}
pdfDoc.Close();
if (textBox2.Text != null)
{
int sentCount = 1;
int.TryParse(textBox3.Text, out sentCount);
SummarizerArguments sumargs = new SummarizerArguments
{
DictionaryLanguage = "en",
DisplayLines = sentCount,
DisplayPercent = 0,
InputFile = @"D:\\htmltext.pdf",
};
SummarizedDocument doc = Summarizer.Summarize(sumargs);
string summary = string.Join("\r\n\r\n", doc.Sentences.ToArray());
richTextBox1.Text = summary;
}
else
{
richTextBox1.Text = "Please give the query value!!!";
}
}
public static SummarizedDocument Summarize(SummarizerArguments args)
{
if (args == null)
{
return(null);
}
Article article = null;
if (args.InputString.Length > 0 && args.InputFile.Length == 0)
{
article = ParseDocument(args.InputString, args);
}
else
{
article = ParseFile(args.InputFile, args);
}
Grader.Grade(article);
Highlighter.Highlight(article, args);
SummarizedDocument sumdoc = CreateSummarizedDocument(article, args);
return(sumdoc);
}
private void button3_Click(object sender, EventArgs e)
{
if (textBox1.Text != null)
{
int sentCount = 1;
int.TryParse(SentenceCountTextBox.Text, out sentCount);
SummarizerArguments sumargs = new SummarizerArguments
{
DictionaryLanguage = "en",
DisplayLines = sentCount,
DisplayPercent = 0,
InputFile = textBox1.Text
};
SummarizedDocument doc = Summarizer.Summarize(sumargs);
string summary = string.Join("\r\n\r\n", doc.Sentences.ToArray());
richTextBox1.Text = summary;
}
else
{
richTextBox1.Text = "Sorry there is no file!!!";
}
}
private static Article ParseFile(string fileName, SummarizerArguments args)
{
string text = LoadFile(fileName);
return(ParseDocument(text, args));
}