Futese

Futese

'Futese' adalah singkatan dari fu ll te xt se Arch.

Ini adalah mesin pencari teks lengkap yang dapat dipertahankan dalam memori yang sederhana dalam kurang dari 1000 baris kode C#:

• Bagian indeks menandai string dan menambahkan kunci yang sesuai ke indeks
• Tokenizer indeks dapat disesuaikan. Secara default ia menghapus diakritik dan menyimpannya saja
• Bagian kueri dapat mencari frasa. Itu mendukung dan (bla blih), atau (bla | blih) dan bukan (bla - blih) dengan tokenizernya sendiri
• Indeks dapat disimpan ke aliran (atau file) dan dimuat ulang
• Vesi yang aman-utas menggunakan kunci ( ThreadSafeIndex<T> ) dan versi bersamaan menggunakan ConcurrentDictionary ( ConcurrentIndex<T> ) disediakan. ConcurrentIndex<T> menggunakan lebih banyak memori.
• Anda dapat indeks menggunakan versi non thread-aman, simpan dan memuatnya dengan versi yang aman
• Seluruh kode juga tersedia sebagai satu file .cs: futese.cs

 // create an index with string keys
var index = new Index<string>();

// add keys and phrases
index.Add("a", "This is a simple phrase");
index.Add("b", "And this one is another phrase a bit longer");
index.Add("c", "The last phrase (this one) contains french (with diacritics) like 'réveillez-vous à l'heure!'");

// search
SimpleTest(index);

// persist to a file
var fileName = "test.fts";
index.Save(fileName);

// load from a file
var newIndex = new Index<string>();  // or new ThreadSafe<string>() for example
newIndex.Load(fileName);

// search again
SimpleTest(newIndex);

newIndex.KeysCount.Should().Be(index.KeysCount);

 static void SimpleTest(Index<string> index)
{
    string[] result;
    result = index.Search("this").Distinct().ToArray();
    result.Should().BeEquivalentTo(["a", "b", "c"]);

    result = index.Search("this is").Distinct().ToArray();
    result.Should().BeEquivalentTo(["a", "b"]);

    result = index.Search("simple | with").Distinct().ToArray();
    result.Should().BeEquivalentTo(["a", "c"]);

    result = index.Search("that").Distinct().ToArray();
    result.Should().BeEquivalentTo([]);

    result = index.Search("the").Distinct().ToArray();
    result.Should().BeEquivalentTo(["c"]);

    result = index.Search("rev").Distinct().ToArray();
    result.Should().BeEquivalentTo(["c"]);

    result = index.Search("-one").Distinct().ToArray();
    result.Should().BeEquivalentTo(["a"]);

    result = index.Search("-this | last").Distinct().ToArray();
    result.Should().BeEquivalentTo([]);
}

Anda juga dapat menggunakan kunci yang kompleks, misalnya kelas Customer ini dapat digunakan sebagai kunci, dan datanya akan bertahan juga:

 // a key must be IParsable and should generally implement IEquatable<T>
private sealed class Customer(int id, string firstName, string lastName, int age) :
    IParsable<Customer>, IEquatable<Customer>
{
    public int Id => id;
    public string FirstName => firstName;
    public string LastName => lastName;
    public int Age => age;

    // used when saved to a stream.
    // by default, the index will use object.ToString() to persist the key,
    // but you can also implement IStringable.ToString()
    public override string ToString() => id + "t" + firstName + "t" + lastName + "t" + age;

    // use id as the real key
    public override int GetHashCode() => id.GetHashCode();
    public override bool Equals(object? obj) => Equals(obj as Customer);
    public bool Equals(Customer? other) => other != null && other.Id == id;

    // used when loading from a stream
    public static Customer Parse(string s, IFormatProvider? provider)
    {
        var split = s.Split('t');
        return new Customer(int.Parse(split[0]), split[1], split[2], int.Parse(split[3]));
    }

    // not called by futese which expects Parse to succeed
    public static bool TryParse([NotNullWhen(true)] string? s, IFormatProvider? provider, [MaybeNullWhen(false)] out Customer result) => throw new NotImplementedException();
}

Dan inilah cara Anda menggunakannya:

 var index = new Index<Customer>();

index.Add(new(0, "alice", "hunting-bobby-crown", 25));
index.Add(new(1, "bob", "albert-down", 32));
index.Add(new(2, "carl", "ctrl-alt", 15));

// search
TestWithObjects(index);

// persist to a file
var fileName = "customer.fts";
index.Save(fileName);

// load from a file
var newIndex = new Index<Customer>();
newIndex.Load(fileName);

// search again
TestWithObjects(newIndex);

 static void TestWithObjects(Index<Customer> index)
{
    Customer[] result;
    result = index.Search("al").Distinct().ToArray();
    result.Should().OnlyContain(c => c.FirstName == "alice" || c.FirstName == "bob" || c.FirstName == "carl");

    result = index.Search("b").Distinct().ToArray();
    result.Should().OnlyContain(c => c.FirstName == "alice" || c.FirstName == "bob");

    result = index.Search("a -c").Distinct().ToArray();
    result.Should().OnlyContain(c => c.FirstName == "bob");

    result = index.Search("a c").Distinct().ToArray();
    result.Should().OnlyContain(c => c.FirstName == "alice" || c.FirstName == "carl");

    result = index.Search("a d").Distinct().ToArray();
    result.Should().OnlyContain(c => c.FirstName == "bob");

    result = index.Search("hunting a").Distinct().ToArray();
    result.Should().OnlyContain(c => c.FirstName == "alice");
}