How to combine partially pre-cached MemoryStream with FileStream?

Question

For a time-critical media presentation application, it is important that media files be presented right at the instance when the user selects it. Those files reside in a truly humongous directory structure, comprised of thousands of media files.

Clearly, caching the media files in a MemoryStream is the way to go; however, due to the sheer amount of files, it’s not feasible to cache each file entirely. Instead, my idea is to pre-cache a certain buffer of each file, and once the file is presented, play from that cache until the rest of the file is loaded from the hard disk.

What I don’t see is how to “concatenate” both the MemoryStream and the FileStream so as to provide a seamless playback experience. I’m not very strong in data streams (yet), and I see several problems:

• How does one keep track of the current read position within the MemoryStream and provide that to the FileStream without the MemoryStream reading more than that?
• How does one switch from one stream to the other without having both streams either partially overlap each other or creating a “playback break”?
• If using a queue of streams (as suggested in How do I concatenate two System.Io.Stream instances into one?), how can I specify that the second stream has to be ready for read access instantaneously after the first stream is done? Here, in particular, I don’t see how the MemoryStream would help at all, since the FileStream, as second one in the queue, would only begin accessing the hard disk once it’s actually used.
• Is it really a feasible approach to have literally hundreds, if not thousands of open streams at once?

Note that I don’t need write access—reading is fully sufficient for the problem at hand. Also, this question is similar to Composite Stream Wrapper providing partial MemoryStream and full original Stream, but the solution provided there is a bug fix for Windows Phone 8 that doesn’t apply in my case.

I’d very much like to widen my rather limited understanding of this, so any help is greatly appreciated.

Answer 1

I would suggest something like the following solution:

• Inherit your own CachableFileStream from FileStream
• Implement a very simple Cache which uses a data structure you prefer (like a Queue)
• Allow Preloading data into the internal cache
• Allow Reloading data into the internal cache
• Modify the original Read behaviour in a way, that your cache is used

To give you an idea of my idea I would suggest some implementation like the following one:

The usage could be like that:

CachableFileStream cachedStream = new CachableFileStream(...)
{
    PreloadSize = 8192,
    ReloadSize = 4096,
};

// Force preloading data into the cache
cachedStream.Preload();

...
cachedStream.Read(buffer, 0, buffer.Length);
...

Warning: The code below is neither correctly tested nor ideal - this shall just give you an idea!

The CachableFileStream class:

using System;
using System.IO;
using System.Threading.Tasks;

/// <summary>
/// Represents a filestream with cache.
/// </summary>
public class CachableFileStream : FileStream
{
    private Cache<byte> cache;
    private int preloadSize;
    private int reloadSize;

    /// <summary>
    /// Gets or sets the amount of bytes to be preloaded.
    /// </summary>
    public int PreloadSize
    {
        get
        {
            return this.preloadSize;
        }

        set
        {
            if (value <= 0)
                throw new ArgumentOutOfRangeException(nameof(value), "The specified preload size must not be smaller than or equal to zero.");

            this.preloadSize = value;
        }
    }

    /// <summary>
    /// Gets or sets the amount of bytes to be reloaded.
    /// </summary>
    public int ReloadSize
    {
        get
        {
            return this.reloadSize;
        }

        set
        {
            if (value <= 0)
                throw new ArgumentOutOfRangeException(nameof(value), "The specified reload size must not be smaller than or equal to zero.");

            this.reloadSize = value;
        }
    }

    /// <summary>
    /// Initializes a new instance of the <see cref="CachableFileStream"/> class with the specified path and creation mode.
    /// </summary>
    /// <param name="path">A relative or absolute path for the file that the current CachableFileStream object will encapsulate</param>
    /// <param name="mode">A constant that determines how to open or create the file.</param>
    /// <exception cref="System.ArgumentException">
    /// Path is an empty string (""), contains only white space, or contains one or more invalid characters.
    /// -or- path refers to a non-file device, such as "con:", "com1:", "lpt1:", etc. in an NTFS environment.
    /// </exception>
    /// <exception cref="System.NotSupportedException">
    /// Path refers to a non-file device, such as "con:", "com1:", "lpt1:", etc. in a non-NTFS environment.
    /// </exception>
    /// <exception cref="System.ArgumentNullException">
    /// Path is null.
    /// </exception>
    /// <exception cref="System.Security.SecurityException">
    /// The caller does not have the required permission.
    /// </exception>
    /// <exception cref="System.IO.FileNotFoundException">
    /// The file cannot be found, such as when mode is FileMode.Truncate or FileMode.Open, and the file specified by path does not exist.
    /// The file must already exist in these modes.
    /// </exception>
    /// <exception cref="System.IO.IOException">
    /// An I/O error, such as specifying FileMode.CreateNew when the file specified by path already exists, occurred.-or-The stream has been closed.
    /// </exception>
    /// <exception cref="System.IO.DirectoryNotFoundException">
    /// The specified path is invalid, such as being on an unmapped drive.
    /// </exception>
    /// <exception cref="System.IO.PathTooLongException">
    /// The specified path, file name, or both exceed the system-defined maximum length.
    /// For example, on Windows-based platforms, paths must be less than 248 characters, and file names must be less than 260 characters.
    /// </exception>
    /// <exception cref="System.ArgumentOutOfRangeException">
    /// Mode contains an invalid value
    /// </exception>
    public CachableFileStream(string path, FileMode mode) : base(path, mode)
    {
        this.cache = new Cache<byte>();
        this.cache.CacheIsRunningLow += CacheIsRunningLow;
    }

    /// <summary>
    /// Reads a block of bytes from the stream and writes the data in a given buffer.
    /// </summary>
    /// <param name="array">
    /// When this method returns, contains the specified byte array with the values between
    /// offset and (offset + count - 1) replaced by the bytes read from the current source.
    /// </param>
    /// <param name="offset">The byte offset in array at which the read bytes will be placed.</param>
    /// <param name="count">The maximum number of bytes to read.</param>
    /// <returns>
    /// The total number of bytes read into the buffer. This might be less than the number
    /// of bytes requested if that number of bytes are not currently available, or zero
    /// if the end of the stream is reached.
    /// </returns>
    /// <exception cref="System.ArgumentNullException">
    /// Array is null.
    /// </exception>
    /// <exception cref="System.ArgumentOutOfRangeException">
    /// Offset or count is negative.
    /// </exception>
    /// <exception cref="System.NotSupportedException">
    /// The stream does not support reading.
    /// </exception>
    /// <exception cref="System.IO.IOException">
    /// An I/O error occurred.
    /// </exception>
    /// <exception cref="System.ArgumentException">
    /// Offset and count describe an invalid range in array.
    /// </exception>
    /// <exception cref="System.ObjectDisposedException">
    /// Methods were called after the stream was closed.
    /// </exception>
    public override int Read(byte[] array, int offset, int count)
    {
        int readBytesFromCache;

        for (readBytesFromCache = 0; readBytesFromCache < count; readBytesFromCache++)
        {
            if (this.cache.Size == 0)
                break;

            array[offset + readBytesFromCache] = this.cache.Read();
        }

        if (readBytesFromCache < count)
            readBytesFromCache += base.Read(array, offset + readBytesFromCache, count - readBytesFromCache);

        return readBytesFromCache;
    }

    /// <summary>
    /// Preload data into the cache.
    /// </summary>
    public void Preload()
    {
        this.LoadBytesFromStreamIntoCache(this.PreloadSize);
    }

    /// <summary>
    /// Reload data into the cache.
    /// </summary>
    public void Reload()
    {
        this.LoadBytesFromStreamIntoCache(this.ReloadSize);
    }

    /// <summary>
    /// Loads bytes from the stream into the cache.
    /// </summary>
    /// <param name="count">The number of bytes to read.</param>
    private void LoadBytesFromStreamIntoCache(int count)
    {
        byte[] buffer = new byte[count];
        int readBytes = base.Read(buffer, 0, buffer.Length);

        this.cache.AddRange(buffer, 0, readBytes);
    }

    /// <summary>
    /// Represents the event handler for the CacheIsRunningLow event.
    /// </summary>
    /// <param name="sender">The sender of the event.</param>
    /// <param name="e">Event arguments.</param>
    private void CacheIsRunningLow(object sender, EventArgs e)
    {
        this.cache.WarnIfRunningLow = false;

        new Task(() =>
        {
            Reload();
            this.cache.WarnIfRunningLow = true;
        }).Start();
    }
}

The Cache class:

using System;
using System.Collections.Concurrent;

/// <summary>
/// Represents a generic cache.
/// </summary>
/// <typeparam name="T">Defines the type of the items in the cache.</typeparam>
public class Cache<T>
{
    private ConcurrentQueue<T> queue;

    /// <summary>
    /// Is executed when the number of items within the cache run below the
    /// specified warning limit and WarnIfRunningLow is set.
    /// </summary>
    public event EventHandler CacheIsRunningLow;

    /// <summary>
    /// Gets or sets a value indicating whether the CacheIsRunningLow event shall be fired or not.
    /// </summary>
    public bool WarnIfRunningLow
    {
        get;
        set;
    }

    /// <summary>
    /// Gets or sets a value that represents the lower warning limit.
    /// </summary>
    public int LowerWarningLimit
    {
        get;
        set;
    }

    /// <summary>
    /// Gets the number of items currently stored in the cache.
    /// </summary>
    public int Size
    {
        get;
        private set;
    }

    /// <summary>
    /// Initializes a new instance of the <see cref="Cache{T}"/> class.
    /// </summary>
    public Cache()
    {
        this.queue = new ConcurrentQueue<T>();
        this.Size = 0;
        this.LowerWarningLimit = 1024;
        this.WarnIfRunningLow = true;
    }

    /// <summary>
    /// Adds an item into the cache.
    /// </summary>
    /// <param name="item">The item to be added to the cache.</param>
    public void Add(T item)
    {
        this.queue.Enqueue(item);
        this.Size++;
    }

    /// <summary>
    /// Adds the items of the specified array to the end of the cache.
    /// </summary>
    /// <param name="items">The items to be added.</param>
    public void AddRange(T[] items)
    {
        this.AddRange(items, 0, items.Length);
    }

    /// <summary>
    /// Adds the specified count of items of the specified array starting
    /// from offset to the end of the cache.
    /// </summary>
    /// <param name="items">The array that contains the items.</param>
    /// <param name="offset">The offset that shall be used.</param>
    /// <param name="count">The number of items that shall be added.</param>
    public void AddRange(T[] items, int offset, int count)
    {
        for (int i = offset; i < count; i++)
            this.Add(items[i]);
    }

    /// <summary>
    /// Reads one item from the cache.
    /// </summary>
    /// <returns>The item that has been read from the cache.</returns>
    /// <exception cref="System.InvalidOperationException">
    /// The cache is empty.
    /// </exception>
    public T Read()
    {
        T item;

        if (!this.queue.TryDequeue(out item))
            throw new InvalidOperationException("The cache is empty.");

        this.Size--;

        if (this.WarnIfRunningLow &&
            this.Size < this.LowerWarningLimit)
        {
            this.CacheIsRunningLow?.Invoke(this, EventArgs.Empty);
        }

        return item;
    }

    /// <summary>
    /// Peeks the next item from cache.
    /// </summary>
    /// <returns>The item that has been read from the cache (without deletion).</returns>
    /// <exception cref="System.InvalidOperationException">
    /// The cache is empty.
    /// </exception>
    public T Peek()
    {
        T item;

        if (!this.queue.TryPeek(out item))
            throw new InvalidOperationException("The cache is empty.");

        return item;
    }
}

I hope this helps, have fun ;-)