The trick is to combine `pipes` with the `foldl` library. In fact,
given any `Fold`, you can convert it into a scanning pipe like this:
convert :: Monad m => Fold a b -> Pipe a b m r
convert = Control.Foldl.purely Pipes.Prelude.scan
So now you have a smaller problem: write a fold for a moving average of
a stream of numbers. It turns out that there is a nice way to do this
in O(1) space if you do an exponential moving average:
http://lpaste.net/100765
You can also do an ordinary moving average, too, but that requires O(N)
space (where N is the window size). It basically involves storing the
last N elements as the fold's internal state.
The `foldl` library is easy to learn. Just read the module header here
and the documentation for the `Fold` type:
http://hackage.haskell.org/package/foldl-1.0.2/docs/Control-Foldl.html
Now, technically, you could do all of this without using the `foldl`
library at all, just by passing the folding logic directly to
`Pipes.Prelude.scan`, but by using `foldl` you make it really easy to
add additional metrics to your `Fold` while still passing over the data
just once.
On 3/5/2014 8:01 AM, Alex Rozenshteyn wrote:
I feel like I must be missing something rather basic, but I have been
trying to figure out how to use pipes to compute a moving average of a
stream of numbers. I've written code to do the EWMA, but I can't
figure out how to do the sliding window. It doesn't help that I'm a
bit of a pipes beginner.
Advice is welcome and appreciated.
Thank you.
--
You received this message because you are subscribed to the Google
Groups "Haskell Pipes" group.
To unsubscribe from this group and stop receiving emails from it, send
an email to [email protected]
<mailto:[email protected]>.
To post to this group, send email to [email protected]
<mailto:[email protected]>.
--
You received this message because you are subscribed to the Google Groups "Haskell
Pipes" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
