var documents = new[]
{
    "i like apples",
    "i like pears",
    "i like fruit like oranges",
    "i hate bananas"
};


// build index

var index =
    from document in documents

    // term frequency - number of times a term appears in a document
    let terms = System.Text.RegularExpressions.Regex.Split(document, @"\W+")
    let termFrequencies = terms
        .GroupBy(t => t)
        .Select(g => (term: g.Key, frequency: (double)g.Count() / terms.Length))

    // inverse document frequency - log of the ratio of the *number of documents* to the *number of documents containing the term*ƒƒ
    let uniqueTerms = terms.Distinct()
    let inverseDocumentFrequencies =
        from term in uniqueTerms
        let documentCount = documents.Count(d => d.Contains(term))
        select (term, score: Math.Log((double)documents.Length / documentCount))

    // calculate tf*idf 
    let scores =
        from tf in termFrequencies
        join idf in inverseDocumentFrequencies
            on tf.term equals idf.term
        select (tf.term, score: tf.frequency * idf.score)

    // document scores
    select (document, scores);


// search index

var results =
    from result in index
    from scores in result.scores
    where scores.term == "like"
    orderby scores.score descending
    select (result.document, scores.score);

results.Select((r, i) => $"{1 + i}. {r.document,-25} {r.score:F5}")
    .ToList()
    .ForEach(Console.WriteLine);