I would like to implement a certain type of throttling of events in reactive-banana. It should work such that an event is not let through if arrives at less then delta seconds from the last event that passed through. If it is not let through then it is stored and is fired after delta seconds from the last fired event.
Below is a program that implements this for lists of time stamped numbers. Would it be possible to translate this to reactive-banana ?
Also, in reactive-banana how do I fire an event x seconds after some other event comes in ?
module Main where
import Data.List
-- 1 second throtling
-- logic is to never output a value before 1 second has passed since last value was outputed.
main :: IO()
main = print $ test [ (0.0, 1.0), (1.1, 2.0), (1.5,3.0), (1.7,4.0), (2.2, 5.0) ]
--should output [ (0.0, 1.0), (1.1, 2.0), (2.1,4.0), (3.1, 5.0) ]
test :: [(Double,Double)] -> [(Double,Double)]
test list = g v (concat xs)
where
(v, xs) = mapAccumL f (-50,Nothing) list
g (t, Just x) ys = ys ++ [ (t+1,x) ]
g _ ys = ys
f (lasttime, Just holdvalue) (t,x) = if t > (lasttime+1) then
if t > (lasttime + 2) then
( (t, Nothing), [ (lasttime+1,holdvalue), (t,x)] )
else ( (lasttime+1, Just x) , [ (lasttime+1,holdvalue) ] )
else
( (lasttime, Just x), [] )
f (lasttime, Nothing) (t,x) = if t > (lasttime+1) then
( (t,Nothing) , [ (t, x ) ] )
else ( (lasttime, Just x), [] )
As of reactive-banana-0.6, it is definitely possible to implement the functionality you desire, but it is a little involved.
Basically, you have use an external framework like wxHaskell to create a timer, which you can then use to schedule events. The Wave.hs example demonstrates how to do that.
At the moment, I have opted to not include a notion of time in the reactive-banana library itself. The reason is simply that different external framework have timers of different resolution or quality, there is no one-size that fits it all.
I do intend to add common helper functions that deal with time and timers to the library itself, but I still need to find a good way to make it generic over different timers and figure out which guarantees I can provide.
Ok, I managed to implement what I described in my question. I'm not so happy that IO is needed to control the timer via reactimate. I wonder if it would be possible to have a throttle with signature throttle::Event t a -> Int -> Event t a ...
ps: I'm very novice in Haskell so the code could probably a lot more compact or elegant.
{-----------------------------------------------------------------------------
------------------------------------------------------------------------------}
{-# LANGUAGE ScopedTypeVariables #-} -- allows "forall t. NetworkDescription t"
import Graphics.UI.WX hiding (Event)
import Reactive.Banana
import Reactive.Banana.WX
import Data.Time
{-----------------------------------------------------------------------------
Main
------------------------------------------------------------------------------}
data ThrottledValue a = FireStoredValue a | FireNowAndStartTimer a| HoldIt a | Stopped deriving Show
data ThrottledEvent a = TimerEvent | RealEvent a deriving Show
main = start $ do
f <- frame [text := "Countercesss"]
sl1 <- hslider f False 0 100 []
sl2 <- hslider f False 0 100 []
set f [ layout := column 0 [widget sl1, widget sl2] ]
t <- timer f []
set t [ enabled := False ]
let networkDescription :: forall t. NetworkDescription t ()
networkDescription = do
slEv <- event0 sl1 command
tick <- event0 t command
slB <- behavior sl1 selection
let (throttledEv, reactimates) = throttle (slB <# slEv) tick t 100
reactimates
reactimate $ fmap (\x -> set sl2 [selection := x]) throttledEv
net <- compile networkDescription
actuate net
throttle::Event t a -> Event t () -> Timer -> Int -> (Event t a, NetworkDescription t () )
throttle ev tick timer dt = (throttledEv, reactimates)
where
all = union (fmap (\x-> RealEvent x) ev) (fmap (\x -> TimerEvent) tick)
result = accumE Stopped $ fmap h all
where
h (RealEvent x) Stopped = FireNowAndStartTimer x
h TimerEvent Stopped = Stopped
h (RealEvent x) (FireNowAndStartTimer _) = HoldIt x
h TimerEvent (FireNowAndStartTimer _) = Stopped
h (RealEvent x) (HoldIt _) = HoldIt x
h (TimerEvent) (HoldIt y) = FireStoredValue y
h (RealEvent x) (FireStoredValue _) = HoldIt x
h (TimerEvent) (FireStoredValue _) = Stopped
start (FireStoredValue a) = Just $ resetTimer timer dt
start (FireNowAndStartTimer a) = Just $ resetTimer timer dt
start _ = Nothing
stop Stopped = Just $ stopTimer timer
stop _ = Nothing
reactimates = do
reactimate $ filterJust $ fmap stop result
reactimate $ filterJust $ fmap start result
filterFired (FireStoredValue a) = Just a
filterFired (FireNowAndStartTimer a) = Just a
filterFired _ = Nothing
throttledEv = filterJust $ fmap filterFired result
startTimer t dt = set t [ enabled := True, interval := dt ]
stopTimer t = set t [ enabled := False ]
resetTimer t dt = stopTimer t >> startTimer t dt
Related
Duplicating this from github as per #HeinrichApfelmus's suggestion:
This may be just a usage error on my part, but I am noticing a strange phenomenon when trying to set up conditional visibility/layout for dynamically created UI elements (in WX of course). As somewhat of a toy-example, I tried to create a widget that created StaticText elements on the fly and allowed the user to "browse" through these elements through '<' '>' buttons.
The problem I am noting is that all labels are invisible until a new one is created, at which point the current widget in focus becomes visible. Whether this is a bug or just a paradigm I am misusing, or a subtlety with reactive frameworks, I am unsure as to how to resolve this. Here is the code I have at this point, which exhibits the problem:
{-# LANGUAGE RecursiveDo #-}
module Test.Adder where
import Reactive.Banana
import Reactive.Banana.WX
import Graphics.UI.WX.Attributes
import Graphics.UI.WX hiding (Event, newEvent, empty, Identity)
import Graphics.UI.WXCore hiding (Event, Timer, empty, Identity, newEvent)
import Graphics.UI.WXCore.Frame
-- | Combine Unit-Events
anyEvent :: [Event ()] -> Event ()
anyEvent = foldl1 (unionWith (\_ _ -> ()))
-- | Unsugared if-then-else function
if_ :: Bool -> a -> a -> a
if_ True x _ = x
if_ False _ y = y
-- | Apply a function to the value at an index, or return a default value
-- if the index is out of range
(!?) :: (a -> b) -> b -> Int -> ([a] -> b)
(f!? ~y) n xs
| n < 0 = y
| otherwise = case drop n xs of
x:_ -> f x
[] -> y
main :: IO ()
main = start test
create :: Window w -> Int -> Behavior Int -> Event Int -> Event () -> MomentIO (StaticText ())
create t i bi ei eRef = do
let tx = replicate i '\t' ++ show i
x <- liftIO $ staticText t [ text := tx ]
let beq = (==i) <$> bi
let eMe = filterE (==i) ei
sink x [ visible :== beq ]
reactimate (refresh x <$ anyEvent [ eRef, () <$ eMe ])
return x
test :: IO ()
test = do
f <- frame [text := "Test"]
add <- button f [ text := "+" ]
prv <- button f [ text := "<" ]
cur <- staticText f []
nxt <- button f [ text := ">" ]
tab <- panel f [ clientSize := sz 200 300 ]
deb <- staticText f []
ref <- button f [ text := "refresh" ]
let networkDescription :: MomentIO ()
networkDescription = mdo
eAdd <- event0 add command
eRef <- event0 ref command
let bNotFirst = (>0) <$> bCur
bNotLast = (<) <$> bCur <*> bNext
sink prv [ enabled :== bNotFirst ]
sink cur [ text :== show <$> bCur ]
sink nxt [ enabled :== bNotLast ]
ePrev <- event0 prv command
eNext <- event0 nxt command
let eDelta :: Enum n => Event (n -> n)
eDelta = unions [ pred <$ whenE bNotFirst ePrev
, succ <$ whenE bNotLast eNext ]
eChange = flip ($) <$> bCur <#> eDelta
bCur <- stepper 0 $ eChange
(eIndex, bCount) <- mapAccum 0 ((\x -> (x, succ x)) <$ eAdd)
let bView = (\n i -> if_ (n==0) (0) i) <$> bCount <*> bCur
bNext = pred <$> bCount
eCreate = (\n -> create tab n bView eChange $ anyEvent [eRef,eAdd]) <$> eIndex
reCreate <- execute eCreate
bItemer <- accumB id $ flip (.) . (:) <$> reCreate
let bItems = ($[]) <$> bItemer
bThis = (widget!?(nullLayouts!!0)) <$> bCur <*> bItems
sink tab [ layout :== bThis ]
liftIO $ set f [ layout := column 5 [ margin 10 $ row 5 [ widget add
, widget prv
, widget cur
, widget nxt
, widget ref
]
, fill $ widget tab
]
]
network <- compile networkDescription
actuate network
>
The program below works if run with runhaskell or if compiled but not with -O2. If compiled with -O2 it seems to hang.
I'm using GHC 7.10.2.
I've changed the min/max iterations to 10 and 20 respectively. It will
generate anywhere from 20 to 100 MB of output into the file test.out.
Run time is about 15 - 60 secs.
Program Explanation
Below is a multi-threaded program that has a pool of workers and a manager. The workers generate traces to be used in plotting a Buddhabrot, put it in a queue, and a manager periodically empties the queue and writes the data to disk. When a certain amount of data has been generated, the program stops.
But when the program runs the manager thread only does one check, and then it gets stuck (the worker threads are still running). However, if I remove the part where the manager thread writes to file, then everything seems to work. I just don't understand why...
import Control.Concurrent
import Control.Concurrent.Async
import Control.Concurrent.STM
import Control.Monad
( forever
, unless
)
import Control.Monad.Loops
import System.IO
import System.Random
import qualified Data.Binary as B
import qualified Data.ByteString.Lazy as BS
type Coord = (Double, Double)
type Trace = [Coord]
-- | Represents a rectangle in the complex plane, bounded by a lower left
-- coordinate and an upper right coordinate.
data Plane
= Plane { ll :: Coord, ur :: Coord }
deriving (Show)
-- | Adds two coordinates.
(+.) :: Coord -> Coord -> Coord
(r1, i1) +. (r2, i2) = (r1 + r2, i1 + i2)
-- | Multiplies two coordinates.
(*.) :: Coord -> Coord -> Coord
(r1, i1) *. (r2, i2) = (r1*r2 - i1*i2, r1*i2 + r2*i1)
-- | Computes the square of a coordinate.
square :: Coord -> Coord
square (r, i) = (r*r - i*i, 2*r*i)
-- | Distance from origin to a given coordinate.
distFromOrigin :: Coord -> Double
distFromOrigin (r, i) = r*r + i*i
-- | A structure for passing data to the worker threads.
data WorkerData
= WorkerData { wdMinIt :: Int
, wdMaxIt :: Int
, wdTraceQueue :: TQueue Trace
-- ^ A queue of traces to be written to disk.
}
-- | A structure for passing data to the manager thread.
data ManagerData
= ManagerData { mdOutHandle :: Handle
-- ^ Handle to the output file.
, mdNumTraces :: Integer
-- ^ Number of traces to gather.
, mdTraceQueue :: TQueue Trace
-- ^ A queue of traces to be written to disk.
}
-- | Encodes an entity to binary bytestring.
encode :: B.Binary a => a -> BS.ByteString
encode = B.encode
-- | Writes a lazy bytestring to file.
writeToFile :: Handle -> BS.ByteString -> IO ()
writeToFile = BS.hPut
mkManagerData :: TQueue Trace -> IO ManagerData
mkManagerData t_queue =
do let out_f = "test.out"
out_h <- openBinaryFile out_f WriteMode
let num_t = 1000
return $ ManagerData { mdOutHandle = out_h
, mdNumTraces = num_t
, mdTraceQueue = t_queue
}
mkWorkerData :: TQueue Trace -> IO WorkerData
mkWorkerData t_queue =
do let min_it = 10 -- 1000
max_it = 20 -- 10000
return $ WorkerData { wdMinIt = min_it
, wdMaxIt = max_it
, wdTraceQueue = t_queue
}
-- | The actions to be performed by the manager thread.
runManager :: ManagerData -> IO ()
runManager m_data =
do execute 0
return ()
where execute count =
do new_traces <- purgeTQueue $ mdTraceQueue m_data
let new_count = count + (toInteger $ length new_traces)
putStrLn $ "Found " ++ (show $ new_count) ++ " traces so far. "
if length new_traces > 0
then do putStrLn $ "Writing new traces to file..."
_ <- mapM (writeToFile (mdOutHandle m_data))
(map encode new_traces)
putStr "Done"
else return ()
putStrLn ""
unless (new_count >= mdNumTraces m_data) $
do threadDelay (1000 * 1000) -- Sleep 1s
execute new_count
-- | The actions to be performed by a worker thread.
runWorker :: WorkerData -> IO ()
runWorker w_data =
forever $
do c <- randomCoord
case computeTrace c (wdMinIt w_data) (wdMaxIt w_data) of
Just t -> atomically $ writeTQueue (wdTraceQueue w_data) t
Nothing -> return ()
-- | Reads all values from a given 'TQueue'. If any other thread reads from the
-- same 'TQueue' during the execution of this function, then this function may
-- deadlock.
purgeTQueue :: Show a => TQueue a -> IO [a]
purgeTQueue q =
whileJust (atomically $ tryReadTQueue q)
(return . id)
-- | Generates a random coordinate to trace.
randomCoord :: IO Coord
randomCoord =
do x <- randomRIO (-2.102613, 1.200613)
y <- randomRIO (-1.237710, 1.239710)
return (x, y)
-- | Computes a trace, using the classical Mandelbrot function, for a given
-- coordinate and minimum and maximum iteration count. If the length of the
-- trace is less than the minimum iteration count, or exceeds the maximum
-- iteration count, 'Nothing' is returned.
computeTrace
:: Coord
-> Int
-- ^ Minimum iteration count.
-> Int
-- ^ Maximum iteration count.
-> Maybe Trace
computeTrace c0 min_it max_it =
if isUsefulCoord c0
then let step c = square c +. c0
computeIt c it = if it < max_it
then computeIt (step c) (it + 1)
else it
computeTr [] = error "computeTr: empty list"
computeTr (c:cs) = if length cs < max_it
then computeTr (step c:(c:cs))
else (c:cs)
num_it = computeIt c0 0
in if num_it >= min_it && num_it <= max_it
then Just $ reverse $ computeTr [c0]
else Nothing
else Nothing
-- | Checks if a given coordinate is useful by checking if it belongs in the
-- cardioid or period-2 bulb of the Mandelbrot.
isUsefulCoord :: Coord -> Bool
isUsefulCoord (x, y) =
let t1 = x - 1/4
p = sqrt (t1*t1 + y*y)
is_in_cardioid = x < p - 2*p*p + 1/4
t2 = x + 1
is_in_bulb = t2*t2 + y*y < 1/16
in not is_in_cardioid && not is_in_bulb
main :: IO ()
main =
do t_queue <- newTQueueIO
m_data <- mkManagerData t_queue
w_data <- mkWorkerData t_queue
let num_workers = 1
workers <- mapM async (replicate num_workers (runWorker w_data))
runManager m_data
_ <- mapM cancel workers
_ <- mapM waitCatch workers
putStrLn "Tracing finished"
Why It Fails
After reviewing the answers below, I finally realized why it doesn't work as intended. The program does not hang, but the time it takes for the manager thread to encode a single trace is in the order of tens of seconds (and when encoded it consumes several megabytes)! This means that even if there are some tens of traces in the queue when exhausted -- on my machine the workers manage to produce about 250 traces before the queue is exhausted by the manger thread -- it will take forever before the next exhaust.
Hence it matters little what solution I choose unless the work of the manager thread is greatly reduced. For that, I will have to abandon my idea of dumping each individual trace to file and instead process it once generated.
The problem is two-fold:
(1) The manager thread doesn't process any
Traces until it has exhausted the queue.
(2) The worker thread can add elements to the queue very, very quickly.
This results in a race that the manager thread rarely wins. [ This also explains the observed behavior with -O2 - the optimization just made the worker thread faster. ]
Adding some debugging code shows that the worker can add
items to the queue in excess of 100K Traces per second.
Moreover, even though the manager is only interested in
writing out the first 1000 Traces, the worker doesn't
stop at this limit. So, under certain circumstances,
the manager is never able to exit this loop:
purgeTQueue q =
whileJust (atomically $ tryReadTQueue q)
(return . id)
The simplest way to fix the code is to have the
manager thread use readTQueue to read and process just one
item off the queue at a time. This will also block
the manager thread when the queue us empty obviating
the need to the manager thread to periodically sleep.
Changing purgeTQueue to:
purgeTQueue = do item <- atomically $ readTQueue (mdTraceQueue m_data)
return [item]
and removing the threadDelay from runManager fixes the problem.
Example code available in the Lib4.hs module at: https://github.com/erantapaa/mandel
To practice concurrent programming, I wrote the following (suboptimal) program, which repeatedly calculates the first prime bigger than whatever the user inputs:
import Control.Concurrent
import Control.Concurrent.Chan
import Control.Monad (forever)
primeAtLeast n = -- Some pure code that looks up the first prime at least as big as n
outputPrimeAtLeast n = putStrLn $ show $ (n, primeAtLeast n)
main = do
chan <- newChan
worker <- forkIO $ forever $ readChan chan >>= outputPrimeAtLeast
forever $ (readLn :: (IO Int)) >>= (writeChan chan)
killThread worker
I want to have a worker thread in the background that does the actual calculation and outputs (n, primeAtLeast n) as soon as it's finished.
What it's doing now: As soon as I enter a number n, it immediately outputs (n,, returns the control to the main thread, calculates primeAtLeast n in the background and outputs the second half primeAtLeast n) as soon as it's finished.
So is putStrLn not atomic? Or where is the problem?
Try this:
outputPrimeAtLeast n = let p = primeAtLeast n in p `seq` putStrLn $ show (n, p)
The above forces the computation of the prime before the putStrLn is run.
Further, you may use print instead of putStrLn . show:
outputPrimeAtLeast n = let p = primeAtLeast n in p `seq` print (n, p)
Alternatively, you may use a putStrLn function which forces every single character before starting printing anything.
strictPutStrLn :: Show a => a -> IO ()
strictPutStrLn x = let str = show x in str `listSeq` putStrLn str
listSeq :: [a] -> b -> b
listSeq [] w = w
listSeq (x:xs) w = x `seq` listSeq xs w
Given the following:
integralB :: Num a => Behavior t a -> Behavior t a -- definite integral of a behaviour
eJump :: Event t a -- tells the player to jump
bYAccel = pure 4000 -- y acceleration
bYVel = integralB bYAccel -- y velocity
bY = integralB bYVel -- y position
How do I make the player jump (probably by setting its y velocity) when a jump event arrives?
You'll need to be able to apply an impulse to the Y velocity for the jump. From your own answer, you've come up with a way to do so by summing all the impulses from the jumps and adding them to the integral of the acceleration.
Your acceleration is also constant. If you don't want the player falling constantly, you'd need something like:
bYAccel = (ifB airborne) 4000 0
airborne = fmap (>0) bY
ifB :: Behavior t Bool -> a -> a -> Behavior t a
ifB boolBehavior yes no = fmap (\bool -> if bool then yes else no) boolBehavior
One possible reason the height of your jumps varies is you aren't resetting the velocity when the player lands. If you have rules that hold the player above some position (like the floor), and are somehow stopping acceleration when the player hits the floor, you will also need to set the velocity to 0 if it is in the direction of the floor. (If you also set it to 0 when it's not in the direction of the floor, the player can never get the velocity to leave the ground.)
The reason this would cause erratic jumping heights is that the final velocity when the player lands will be close to the impulse you applied for them to take off. Using your numbers, if a jump started with a velocity of -5000, and ended with a velocity of 4800, the next jump will add an impulse of -5000, taking the jump to a starting velocity of only -200. That might have an ending velocity of 300, so the next jump will be an almost full -4700 jump.
Here's a complete working example. It uses the gloss library for input and display. The gameDefinition corresponds to the components introduced in your question. integrateDeltas is equivalent to your integralB, but produces events that are impulses, which are easy to generate in a clocked framework like gloss, and easy to use mixed with other events that cause impulses, like jumping.
{-# LANGUAGE RankNTypes #-}
module Main where
import Reactive.Banana
import Reactive.Banana.Frameworks.AddHandler
import Reactive.Banana.Frameworks
import Data.IORef
import qualified Graphics.Gloss.Interface.IO.Game as Gloss
gameDefinition :: GlossGameEvents t -> Behavior t Gloss.Picture
gameDefinition events = renderBehavior
where
bY = accumB 0 (fmap sumIfPositive yShifts)
yShifts = integrateDeltas bYVel
bYVel = accumB 0 yVelChanges
yVelChanges = apply ((ifB airborne) (+) sumIfPositive) yVelShifts
yVelShifts = union (integrateDeltas bYAccel) (fmap (const 3) eJump)
bYAccel = (ifB airborne) (-10) 0
airborne = fmap (>0) bY
eJump = filterE isKeyEvent (event events)
integrateDeltas = integrateDeltaByTimeStep (timeStep events)
renderBehavior = (liftA3 render) bY bYVel bYAccel
render y yVel yAccel =
Gloss.Pictures [
Gloss.Translate 0 (20+y*100) (Gloss.Circle 20),
Gloss.Translate (-50) (-20) (readableText (show y)),
Gloss.Translate (-50) (-40) (readableText (show yVel)),
Gloss.Translate (-50) (-60) (readableText (show yAccel))
]
readableText = (Gloss.Scale 0.1 0.1) . Gloss.Text
-- Utilities
sumIfPositive :: (Ord n, Num n) => n -> n -> n
sumIfPositive x y = max 0 (x + y)
ifB :: Behavior t Bool -> a -> a -> Behavior t a
ifB boolBehavior yes no = fmap (\bool -> if bool then yes else no) boolBehavior
integrateDeltaByTimeStep :: (Num n) => Event t n -> Behavior t n -> Event t n
integrateDeltaByTimeStep timeStep derivative = apply (fmap (*) derivative) timeStep
isKeyEvent :: Gloss.Event -> Bool
isKeyEvent (Gloss.EventKey _ _ _ _) = True
isKeyEvent _ = False
-- Main loop to run it
main :: IO ()
main = do
reactiveGame (Gloss.InWindow "Reactive Game Example" (400, 400) (10, 10))
Gloss.white
100
gameDefinition
-- Reactive gloss game
data GlossGameEvents t = GlossGameEvents {
event :: Event t Gloss.Event,
timeStep :: Event t Float
}
makeReactiveGameNetwork :: Frameworks t
=> IORef Gloss.Picture
-> AddHandler Gloss.Event
-> AddHandler Float
-> (forall t. GlossGameEvents t -> Behavior t Gloss.Picture)
-> Moment t ()
makeReactiveGameNetwork latestFrame glossEvent glossTime game = do
eventEvent <- fromAddHandler glossEvent
timeStepEvent <- fromAddHandler glossTime
let
events = GlossGameEvents { event = eventEvent, timeStep = timeStepEvent }
pictureBehavior = game events
pictureChanges <- changes pictureBehavior
reactimate (fmap (writeIORef latestFrame) pictureChanges)
reactiveGame :: Gloss.Display
-> Gloss.Color
-> Int
-> (forall t. GlossGameEvents t -> Behavior t Gloss.Picture)
-> IO ()
reactiveGame display color steps game = do
latestFrame <- newIORef Gloss.Blank
(glossEvent, fireGlossEvent) <- newAddHandler
(glossTime, addGlossTime) <- newAddHandler
network <- compile (makeReactiveGameNetwork latestFrame glossEvent glossTime game)
actuate network
Gloss.playIO
display
color
steps
()
(\world -> readIORef latestFrame)
(\event world -> fireGlossEvent event)
(\time world -> addGlossTime time)
In this example, bY checks for collision with a floor at 0 by accumulating the impulses, but constraining the accumulated value to be above 0.
The velocity, bYVel, accumulates all impulses while airborne, but only those impulses that are directed away from the floor while not airborne. If you change
yVelChanges = apply ((ifB airborne) (+) sumIfPositive) yVelShifts
to
yVelChanges = fmap (+) yVelShifts
it recreates the erratic jumping bug.
The acceleration, bYAccel, is only present while airborne.
I used a coordinate system with a +Y axis in the up direction (opposite the acceleration).
The code at the end is a small framework to hook reactive-banana up to gloss.
Solved it! I feel a little silly for not thinking of this earlier, but I just increment a counter every eJump and add that counter on to bYVel.
bJumpVel = sumB $ (-5000) <$ eJump
bYVel = (+) <$> bJumpVel <*> integralB bYAccel
-- gives the sum of the events
sumB :: Num a => Event t a -> Behavior t a
sumB e = accumB 0 $ (+) <$> e
For some reason the height of the jump always varies quite a bit, but that's probably an unrelated problem to do with my timing of things.
I won't mark this question as answered yet in case someone wants to share a better one.
I am using enumerator-0.4.10, and I need to distribute processing of
different parts of the incoming stream to different iteratees (I am
parsing a huge XML file, and different sub-trees have different
processing logic). Only a single iteratee will be active at a time
since the sub-trees don't intersect.
I wrote a simple example that filters the stream and passes the result
to one iteratee; please see below. However, with multiple nested
iteratees it seems to me that I can no longer use an enumeratee. Do I
need to write my own multi-enumeratee that holds multiple inner
iteratees? Any better ideas?
Here is my (beginner's) code for a single nested iteratee:
module Main ( main ) where
import qualified Data.Enumerator as E ( Enumeratee, Step(..), Stream(..),
checkDone, checkDoneEx, continue, enumList, joinI, run_, yield )
import Data.Enumerator ( ($$), (>>==) )
import qualified Data.Enumerator.List as EL ( consume )
-- cribbed from EL.concatMap
concatMapAccum :: Monad m => (s -> ao -> (s, [ai])) -> s ->
E.Enumeratee ao ai m b
concatMapAccum f s0 = E.checkDone (E.continue . step s0)
where
step _ k E.EOF = E.yield (E.Continue k) E.EOF
step s k (E.Chunks xs) = loop s k xs
loop s k [] = E.continue (step s k)
loop s k (x:xs) = case f s x of
(s', ais) -> k (E.Chunks $ ais) >>==
E.checkDoneEx (E.Chunks xs) (\k' -> loop s' k' xs)
passFromTo :: Monad m => ((a -> Bool), (a -> Bool)) -> Bool -> E.Enumeratee a a m b
passFromTo (from, to) pass0 =
concatMapAccum updatePass pass0
where
updatePass pass el = case (pass, from el, to el) of
(True, _, to_el) -> (not to_el, [el])
(False, True, _) -> (True, [el])
(False, False, _) -> (False, [])
main :: IO()
main = do
E.run_ (E.enumList 3 [1..20] $$
E.joinI $ passFromTo ((\e -> e == 3 || e == 13), (\e -> e == 7 || e == 17)) False $$
EL.consume) >>= print
$ ./dist/build/StatefulEnumeratee/StatefulEnumeratee
[3,4,5,6,7,13,14,15,16,17]
Yes, you need an enumeratee that passes the stream to multiple iteratees, like Data.Iteratee.sequence_ and Data.Iteratee.Parallel.psequence_ from iteratee-0.8.6. sequence_ takes a list of iteratees to run simultaneously, and handles each input chunk by mapM across that list. psequence_ takes similar arguments, but runs each input iteratee in a separate forkIO thread.
There has been some discussion on haskell-cafe and the iteratee mailing lists about these over the past year, eg: http://www.haskell.org/pipermail/haskell-cafe/2011-January/088319.html The main thing to be careful about is handling errors from the inner iteratees: in your application, if one inner iteratee fails do you want to terminate all iteratees or just that one, and [how] do you want to propagate those errors.