Question

I have to extract summary from newspaper article . The summary is extracted based on given Keyword and according to below mentioned rules .

1. Summary should be of 200 characters.

2. Start printing from that sentence in article as soon as keyword
   appears in that sentence and print upto 200 characters

3. If the matching sentence occurs towards ending of article such that
   summary is coming out to be less than 200 characters , then move
   back from matching sentence towards previous sentences uptill
   finally 200 charcters containing matching sentence are printed
   finally.

What I have done untill now is …

var regex = new Regex(keyword+@"(.{0,200})");

foreach (Match match in regex.Matches(input))
{
    var result = match.Groups[1].Value;
    Console.WriteLine(result);

    // work with the result
}

The above code successfully reaches the first matching sentence but starts printing AFTER the keyword upto 200 characters rather than beginning of matching sentence.

Also there is no backtracking if end of article is reached before 200 characters are printed.

Please guide me how should I proceed . Even if somebody doesn’t know complete solution , PLEASE do help me out in sub parts of question .

Answer 1

var nextIndex = input.IndexOf(keyword);

while (nextIndex != -1)
{
    var index = nextIndex;
    // To start the 200chars result from right after the keyword, do instead:
    // var index = nextIndex + keyword.Length;

    // If you want to stop after you reached the end of the text once:
    // var end = false;

    if (index + 200 >= input.Length)
    {
        index = input.Length - 200;

        // If you want to stop after you reached the end of the text once:
        // var end = true;
    }

    var result = index < 0 ? input : input.Substring(index, 200);

    Console.WriteLine(result);

    // If you want to stop after you reached the end of the text once:
    // if (end) { break; }

    nextIndex = input.IndexOf(keyword, nextIndex + 1);
}

And if you want to search to be case insensitive, just add StringComparison.OrdinalIgnoreCase as another parameter in both IndexOfs.