Text 0.4.0 countvectorizer loadasync

Lodestar.Text 0.4.0. This page is frozen at that release. Read the current documentation for what main says now. A link to a decision or a migration page follows main, and leaves the archive.

CountVectorizer.LoadAsync

Read a fitted vectorizer back without blocking the caller.

public static Task<CountVectorizer> LoadAsync(Stream source, ArtifactLoadOptions options = null, CancellationToken cancellationToken = default)

Parameters — source is a readable stream, left open. options bounds what will be accepted. cancellationToken cancels the read.

Returns — Task<CountVectorizer>, completing with a fitted vectorizer.

Exceptions — ArgumentNullException for a null source. InvalidDataException for content that is not a saved vectorizer or that exceeds a bound. OperationCanceledException when cancelled.

Example — the asynchronous half of the round trip.

using Lodestar.Text.Vectorization;

static async Task<string> FirstFeatureAsync()
{
    var original = new CountVectorizer();
    original.Fit(["the cat eats", "the dog eats"]);

    using var buffer = new MemoryStream();
    await original.SaveAsync(buffer);
    buffer.Position = 0;

    CountVectorizer restored = await CountVectorizer.LoadAsync(buffer);
    return restored.GetFeatureNames()[0];
}

string first = FirstFeatureAsync().GetAwaiter().GetResult();  // => cat

The GetAwaiter().GetResult() is only what lets a synchronous example drive an async one; in async code the call is simply await.

Remarks — there is no LoadAsync(string path) overload, for the same reason there is no SaveAsync(string): opening the file with useAsync: true is the caller's decision, and doing it here would hide whether the asynchrony reaches the disk.

The bounds in options are checked as the content is read, not after, so an oversized file is refused before it is allocated rather than afterwards.

Applies to — net10.0, netstandard2.0.

See also — CountVectorizer.Load, CountVectorizer.SaveAsync, ArtifactLoadOptions.