Make the analysis parallel (#2)

Co-authored-by: Mats Rauhala <mats.rauhala@iki.fi>
Reviewed-on: #2
Co-authored-by: Mats Rauhala <masse@rauhala.info>
Co-committed-by: Mats Rauhala <masse@rauhala.info>
This commit is contained in:
2021-10-29 22:41:46 +03:00
parent 1c4e766f92
commit 41c666fe93
6 changed files with 58 additions and 48 deletions

View File

@ -1,23 +1,18 @@
module Control.Addressbook.Streaming where
import qualified Data.Text as T
import Conduit
import qualified Data.Conduit.Binary as CB
import qualified Data.Conduit.Combinators as C
import qualified Data.Conduit.List as CL
import qualified Data.Conduit.Text as CT
import Data.Email
import Data.Email.Header
(Header(..))
import System.IO
(stdin)
import qualified Data.Foldable as F
import qualified Data.Map.Strict as Map
import Data.Maybe
(fromMaybe)
@ -25,28 +20,47 @@ import System.Environment
(lookupEnv)
import System.FilePath
((</>))
import Data.Set (Set)
import Data.ByteString (ByteString)
import qualified Data.Set as Set
import qualified Data.ByteString.Lazy as LBS
import Data.Char (ord)
import qualified Data.ByteString.Lazy.Char8 as LBC
import System.IO.Unsafe (unsafeInterleaveIO)
import Control.Parallel.Strategies (rseq, parMap)
import qualified Data.List as L
combine :: (MonadUnliftIO m, MonadResource m, MonadThrow m, MonadIO m) => ConduitM FilePath Header m ()
combine = await >>= \case
Nothing -> pure ()
Just path -> (CB.sourceFile path .| parseEmail) >> combine
chunks :: Int -> [a] -> [[a]]
chunks n = L.unfoldr $ \case
[] -> Nothing
xs -> Just (splitAt n xs)
run :: IO ()
run = do
datDir <- fromMaybe "./" <$> lookupEnv "HOME"
runResourceT $ do
x <- runConduit stream
runConduit (CL.sourceList (Map.keys x) .| C.map (<> "\n") .| CB.sinkFileCautious (datDir </> ".addressbook.dat"))
xs <- LBS.getContents >>= stream
let set = F.fold (parMap rseq F.fold (chunks 20 xs))
runResourceT $
runConduit $
CL.sourceList (Set.elems set)
.| C.map (<> "\n")
.| CB.sinkFileCautious (datDir </> ".addressbook.dat")
where
separate = \case
From x -> [x]
To xs -> F.toList xs
-- A set of (locally) unique addresses. Composes with parMap
stream :: LBS.ByteString -> IO [Set ByteString]
stream =
CB.sourceHandle stdin
.| CT.decode CT.utf8
.| CT.lines
.| C.map T.unpack
.| combine
.| C.concatMap separate
.| CT.encode CT.utf8
.| C.foldMap (`Map.singleton` ())
traverse (unsafeInterleaveIO . parse . LBC.unpack)
. filter (not . LBS.null)
. LBS.split (fromIntegral $ ord '\n')
parse path =
runResourceT $
runConduit $
CB.sourceFile path .| parseEmail .| C.concatMap separate .| C.foldMap Set.singleton