/// <summary>
/// Checks whether the supplied question is a duplicate of another post.
/// </summary>
/// <param name="q">The questions to check.</param>
/// <param name="matcherAlgo">The matching algorithm to use.</param>
/// <returns>
/// The match results from the specified matcher algo
/// if the question is a duplicate, otherwise null.
/// </returns>
public MatchResults IsDupe(Question q, MatcherAlgo matcherAlgo)
{
if (q == null) throw new ArgumentNullException(nameof(q));
//var testQ = PostFetcher.GetQuestion("https://stackoverflow.com/questions/15735570");
var matcher = matchers.Single(x => x.Algo == matcherAlgo);
var res = matcher.Match(q);
// Any number between 0 and 1.
// 0.5 seems like a good start.
if (res.MatchScore > 0.5)
return res;
else
return null;
}
public MatchResults Match(Question q)
{
if (q == null) throw new ArgumentNullException(nameof(q));
var questionSearchRes = SeApiAccessor.Search(q.Title, q.Tags);
var bow = new BagOfWordsBeta();
foreach (var question in questionSearchRes)
{
if (question.ID == q.ID) continue;
var tf = pos.GetModel(question.Body);
bow.AddDocument(question.ID, tf);
}
var qModel = pos.GetModel(q.Body);
var match = bow.GetSimilarity(qModel, 1);
return new MatchResults
{
LinkToDupeTarget = $"http://stackoverflow.com/q/{match.Keys.First()}",
MatchScore = match.Values.First()
};
}