我想编写一个评估两个传感器的传感器数据的应用程序。两个传感器均在Package个对象中发送数据，这些对象被拆分为Frame个对象。 Package本质上是Tuple<Timestamp, Data[]>，Frame是Tuple<Timestamp, Data>。然后，我需要始终使用两个来源中带有最早时间戳的Frame。
所以基本上我的对象流是
Package -(1:n)-> Frame \
}-pair synchronized-> Tuple<Frame, Frame>
Package -(1:n)-> Frame /
假设每个Package包含2个或3个值（真实度：5-7）和以1递增的整数时间戳（真实度：〜200Hz =>〜5ms增量）。为简单起见，“数据”仅为timestamp * 100。
Packages (timestamp, values[])
Source 1:
{(19, [1700, 1800, 1900]), (22, [2000, 2100, 2200]), (26, [2500, 2600]),
(29, [2700, 2800, 2900]), ...}
Source 2:
{(17, [1500, 1600, 1700]), (19, [1800, 1900]), (21, [2000, 2100]),
(26, [2400, 2500, 2600]), ...}
在(1:n)步骤之后：
Frames (timestamp, value)
Source 1:
{(17, 1700), (18, 1800), (19, 1900), (20, 2000), (21, 2100),
(22, 2200), (25, 2500), (26, 2600), (27, 2700), (28, 2800),
(29, 2900), ...}
Source 2:
{(15, 1500), (16, 1600), (17, 1700), (18, 1800), (19, 1900),
(20, 2000), (21, 2100), (24, 2400), (25, 2500), (26, 2600), ...}
在pair synchronized步骤之后：
Merged tuples (timestamp, source1, source2)
{(15, null, 1500), (16, null, 1600), (17, 1700, 1700), (18, 1800, 1800),
(19, 1900, 1900), (20, 2000, 2000), (21, 2100, 2100), (22, 2200, null),
(24, null, 2400), (25, 2500, 2500), (26, 2600, 2600), ...}
请注意，由于两个来源的 none 均未发送值，因此缺少时间戳23。那只是副作用。我可以放一个空的元组，没关系。元组是(27, 2700, 2700)还是((27, 2700), (27, 2700))也无关紧要，即。 e。 Tuple<Timestamp, Data, Data>或Tuple<Frame, Frame>。
如果文档正确，我很确定(1:n)部分应该是TransformManyBlock<Package, Frame>。
但是我应该在哪个区块中使用 pair synchronized ？？起初，我以为JoinBlock<Frame, Frame>是我想要的东西，但似乎只是将两个元素按索引配对。但是，由于既不能确保两个管道都以相同的时间戳开始，也不能确保两个管道都将始终产生稳定的连续时间戳流（因为有时传输时会丢失几帧的包），所以这不是一种选择。因此，我需要的更多是“ MergeBlock”，它可以决定将两个输入流的哪个元素下一步传播到输出（如果有）。
我认为自己必须自己写类似的东西。但是我很难编写能正确处理两个ISourceBlock变量和一个ITargetBlock变量的代码。我基本上早就被困住了：
private void MergeSynchronized(
ISourceBlock<Frame> source1,
ISourceBlock<Frame> source2,
ITargetBlock<Tuple<Frame, Frame>> target)
{
var frame1 = source1.Receive();
var frame2 = source2.Receive();
//Loop {
// Depending on the timestamp [mis]match,
// either pair frame1+frame2 or frame1+null or null+frame2, and
// replace whichever frame(s) was/were propagated already
// with the next frame from the respective pipeline
//}
}
我甚至对这份草案都不敢肯定：方法应该是async以便我可以使用var frame1 = await source1.ReceiveAsnyc();吗？循环的条件是什么？在哪里以及如何检查完成情况？如何解决明显的问题，即我的代码意味着我必须等到流中的间隙结束才能意识到存在间隙？
我考虑过的替代方法是在管道中添加一个额外的块，以确保每个传感器将足够的“前哨帧”放入管道中，以便始终将每个管道中的第一个对齐，将正确的两个对齐。我猜测这是一种TransformManyBlock，它读取一个Frame，将“期望的”时间戳与实际时间戳进行比较，然后为缺少的时间戳插入前哨帧，直到该帧的时间戳再次正确。
还是pair synchronized部分才是TPL Dataflow对象停止并启动已经与Data部分一起工作的实际代码的地方？
答案 0 :(得分：3)
TPL DataFlow API的问题在于，所有内容都是内部/私有和/或密封的。这样就没有太多扩展API的可能性了。
无论如何，实现一个新的SynchronizedJoinBlock类可能是一个好主意。实际的业务逻辑位于GetMessagesRecursive方法中：
public sealed class SynchronizedJoinBlock<T1, T2>
: IReceivableSourceBlock<Tuple<T1, T2>>
{
private readonly Func<T1, T2, int> _compareFunction;
private readonly Queue<T1> _target1Messages;
private readonly Queue<T2> _target2Messages;
private readonly TransformManyBlock<T1, Tuple<T1, T2>> _target1;
private readonly TransformManyBlock<T2, Tuple<T1, T2>> _target2;
private readonly BatchedJoinBlock<Tuple<T1, T2>, Tuple<T1, T2>> _batchedJoinBlock;
private readonly TransformManyBlock<Tuple<IList<Tuple<T1, T2>>, IList<Tuple<T1, T2>>>, Tuple<T1, T2>> _transformManyBlock;
public ITargetBlock<T1> Target1 => _target1;
public ITargetBlock<T2> Target2 => _target2;
public Task Completion => _transformManyBlock.Completion;
public SynchronizedJoinBlock(Func<T1, T2, int> compareFunction)
{
_compareFunction = compareFunction
?? throw new ArgumentNullException(nameof(compareFunction));
_batchedJoinBlock = new BatchedJoinBlock<Tuple<T1, T2>, Tuple<T1, T2>>(1);
_target1Messages = new Queue<T1>();
_target2Messages = new Queue<T2>();
var syncObject = new object();
Func<ICollection<Tuple<T1, T2>>> getMessagesFunction = () =>
{
lock (syncObject)
{
if (_target1Messages.Count > 0 && _target2Messages.Count > 0)
{
return GetMessagesRecursive(_target1Messages.Peek(), _target2Messages.Peek()).ToArray();
}
else
{
return new Tuple<T1, T2>[0];
}
}
};
_target1 = new TransformManyBlock<T1, Tuple<T1, T2>>((element) =>
{
_target1Messages.Enqueue(element);
return getMessagesFunction();
});
_target1.LinkTo(_batchedJoinBlock.Target1, new DataflowLinkOptions() { PropagateCompletion = true });
_target2 = new TransformManyBlock<T2, Tuple<T1, T2>>((element) =>
{
_target2Messages.Enqueue(element);
return getMessagesFunction();
});
_target2.LinkTo(_batchedJoinBlock.Target2, new DataflowLinkOptions() { PropagateCompletion = true });
_transformManyBlock = new TransformManyBlock<Tuple<IList<Tuple<T1, T2>>, IList<Tuple<T1, T2>>>, Tuple<T1, T2>>(
element => element.Item1.Concat(element.Item2)
);
_batchedJoinBlock.LinkTo(_transformManyBlock, new DataflowLinkOptions() { PropagateCompletion = true });
}
private IEnumerable<Tuple<T1, T2>> GetMessagesRecursive(T1 value1, T2 value2)
{
int result = _compareFunction(value1, value2);
if (result == 0)
{
yield return Tuple.Create(_target1Messages.Dequeue(), _target2Messages.Dequeue());
}
else if (result < 0)
{
yield return Tuple.Create(_target1Messages.Dequeue(), default(T2));
if (_target1Messages.Count > 0)
{
foreach (var item in GetMessagesRecursive(_target1Messages.Peek(), value2))
{
yield return item;
}
}
}
else
{
yield return Tuple.Create(default(T1), _target2Messages.Dequeue());
if (_target2Messages.Count > 0)
{
foreach (var item in GetMessagesRecursive(value1, _target2Messages.Peek()))
{
yield return item;
}
}
}
}
public void Complete()
{
_target1.Complete();
_target2.Complete();
}
Tuple<T1, T2> ISourceBlock<Tuple<T1, T2>>.ConsumeMessage(
DataflowMessageHeader messageHeader,
ITargetBlock<Tuple<T1, T2>> target, out bool messageConsumed)
{
return ((ISourceBlock<Tuple<T1, T2>>)_transformManyBlock)
.ConsumeMessage(messageHeader, target, out messageConsumed);
}
void IDataflowBlock.Fault(Exception exception)
{
((IDataflowBlock)_transformManyBlock).Fault(exception);
}
public IDisposable LinkTo(ITargetBlock<Tuple<T1, T2>> target,
DataflowLinkOptions linkOptions)
{
return _transformManyBlock.LinkTo(target, linkOptions);
}
void ISourceBlock<Tuple<T1, T2>>.ReleaseReservation(
DataflowMessageHeader messageHeader, ITargetBlock<Tuple<T1, T2>> target)
{
((ISourceBlock<Tuple<T1, T2>>)_transformManyBlock)
.ReleaseReservation(messageHeader, target);
}
bool ISourceBlock<Tuple<T1, T2>>.ReserveMessage(
DataflowMessageHeader messageHeader, ITargetBlock<Tuple<T1, T2>> target)
{
return ((ISourceBlock<Tuple<T1, T2>>)_transformManyBlock)
.ReserveMessage(messageHeader, target);
}
public bool TryReceive(Predicate<Tuple<T1, T2>> filter, out Tuple<T1, T2> item)
{
return _transformManyBlock.TryReceive(filter, out item);
}
public bool TryReceiveAll(out IList<Tuple<T1, T2>> items)
{
return _transformManyBlock.TryReceiveAll(out items);
}
}
答案 1 :(得分：1)
这里是SynchronizedJoinBlock块的实现，与Hardy Hobeck的answer中介绍的块类似。当输入块Target1和Target2被标记为完成时，这一小节将处理一些较小的细节，例如取消，处理异常以及处理剩余项的传播。同样，合并逻辑不涉及递归，这应该使递归性能更好（希望我没有测量它），并且不易受到堆栈溢出异常的影响。偏差很小：输出是ValueTuple<T1, T2>而不是Tuple<T1, T2>（目的是减少分配）。
public sealed class SynchronizedJoinBlock<T1, T2> : IReceivableSourceBlock<(T1, T2)>
{
private readonly Func<T1, T2, int> _comparison;
private readonly Queue<T1> _queue1 = new Queue<T1>();
private readonly Queue<T2> _queue2 = new Queue<T2>();
private readonly ActionBlock<T1> _input1;
private readonly ActionBlock<T2> _input2;
private readonly BufferBlock<(T1, T2)> _output;
private readonly object _locker = new object();
public SynchronizedJoinBlock(Func<T1, T2, int> comparison,
CancellationToken cancellationToken = default)
{
_comparison = comparison ?? throw new ArgumentNullException(nameof(comparison));
// Create the three internal blocks
var options = new ExecutionDataflowBlockOptions()
{
CancellationToken = cancellationToken
};
_input1 = new ActionBlock<T1>(Add1, options);
_input2 = new ActionBlock<T2>(Add2, options);
_output = new BufferBlock<(T1, T2)>(options);
// Link the input blocks with the output block
var inputTasks = new Task[] { _input1.Completion, _input2.Completion };
Task.WhenAny(inputTasks).Unwrap().ContinueWith(t =>
{
// If ANY input block fails, then the whole block has failed
((IDataflowBlock)_output).Fault(t.Exception.InnerException);
if (!_input1.Completion.IsCompleted) _input1.Complete();
if (!_input2.Completion.IsCompleted) _input2.Complete();
ClearQueues();
}, default, TaskContinuationOptions.OnlyOnFaulted |
TaskContinuationOptions.RunContinuationsAsynchronously,
TaskScheduler.Default);
Task.WhenAll(inputTasks).ContinueWith(t =>
{
// If ALL input blocks succeeded, then the whole block has succeeded
try
{
if (!t.IsCanceled) PostRemaining(); // Post what's left
}
catch (Exception ex)
{
((IDataflowBlock)_output).Fault(ex);
}
_output.Complete();
ClearQueues();
}, default, TaskContinuationOptions.NotOnFaulted |
TaskContinuationOptions.RunContinuationsAsynchronously,
TaskScheduler.Default);
}
public ITargetBlock<T1> Target1 => _input1;
public ITargetBlock<T2> Target2 => _input2;
public Task Completion => _output.Completion;
private void Add1(T1 value1)
{
lock (_locker)
{
_queue1.Enqueue(value1);
FindAndPostMatched_Unsafe();
}
}
private void Add2(T2 value2)
{
lock (_locker)
{
_queue2.Enqueue(value2);
FindAndPostMatched_Unsafe();
}
}
private void FindAndPostMatched_Unsafe()
{
while (_queue1.Count > 0 && _queue2.Count > 0)
{
var result = _comparison(_queue1.Peek(), _queue2.Peek());
if (result < 0)
{
_output.Post((_queue1.Dequeue(), default));
}
else if (result > 0)
{
_output.Post((default, _queue2.Dequeue()));
}
else // result == 0
{
_output.Post((_queue1.Dequeue(), _queue2.Dequeue()));
}
}
}
private void PostRemaining()
{
lock (_locker)
{
while (_queue1.Count > 0)
{
_output.Post((_queue1.Dequeue(), default));
}
while (_queue2.Count > 0)
{
_output.Post((default, _queue2.Dequeue()));
}
}
}
private void ClearQueues()
{
lock (_locker)
{
_queue1.Clear();
_queue2.Clear();
}
}
public void Complete() => _output.Complete();
public void Fault(Exception exception)
=> ((IDataflowBlock)_output).Fault(exception);
public IDisposable LinkTo(ITargetBlock<(T1, T2)> target,
DataflowLinkOptions linkOptions)
=> _output.LinkTo(target, linkOptions);
public bool TryReceive(Predicate<(T1, T2)> filter, out (T1, T2) item)
=> _output.TryReceive(filter, out item);
public bool TryReceiveAll(out IList<(T1, T2)> items)
=> _output.TryReceiveAll(out items);
(T1, T2) ISourceBlock<(T1, T2)>.ConsumeMessage(
DataflowMessageHeader messageHeader, ITargetBlock<(T1, T2)> target,
out bool messageConsumed)
=> ((ISourceBlock<(T1, T2)>)_output).ConsumeMessage(
messageHeader, target, out messageConsumed);
void ISourceBlock<(T1, T2)>.ReleaseReservation(
DataflowMessageHeader messageHeader, ITargetBlock<(T1, T2)> target)
=> ((ISourceBlock<(T1, T2)>)_output).ReleaseReservation(
messageHeader, target);
bool ISourceBlock<(T1, T2)>.ReserveMessage(
DataflowMessageHeader messageHeader, ITargetBlock<(T1, T2)> target)
=> ((ISourceBlock<(T1, T2)>)_output).ReserveMessage(
messageHeader, target);
}
用法示例：
var joinBlock = new SynchronizedJoinBlock<(int, int), (int, int)>(
(x, y) => Comparer<int>.Default.Compare(x.Item1, y.Item1));
var source1 = new (int, int)[] {(17, 1700), (18, 1800), (19, 1900),
(20, 2000), (21, 2100), (22, 2200), (25, 2500), (26, 2600),
(27, 2700), (28, 2800), (29, 2900)};
var source2 = new (int, int)[] {(15, 1500), (16, 1600), (17, 1700),
(18, 1800), (19, 1900), (20, 2000), (21, 2100), (24, 2400),
(25, 2500), (26, 2600)};
Array.ForEach(source1, x => joinBlock.Target1.Post(x));
Array.ForEach(source2, x => joinBlock.Target2.Post(x));
joinBlock.Target1.Complete();
joinBlock.Target2.Complete();
while (joinBlock.OutputAvailableAsync().Result)
{
Console.WriteLine($"> Received: {joinBlock.Receive()}");
}
输出：
收到的：（（0，0），（15，1500））
收到：（（0，0），（16，1600））
收到：（（17，1700），（17，1700））
收到：（（（18，1800），（18，1800））
收到：（（（19，1900），（19，1900））
收到：（（20，2000），（20，2000））
收到：（（21，2100），（21，2100））
收到：（（22，2200），（0，0））
收到：（（0，0），（24，2400））
收到：（（25，2500），（25，2500））
收到：（（26，2600），（26，2600））
收到：（（（27，2700），（0，0）））
收到：（（28，2800），（0，0））
收到：（（（29，2900），（0，0））
假定传入数据是有序的。
该类与我前一段时间在somewhat related question中发布的JoinDependencyBlock类具有相似的结构。