Add flakes

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>

addressbook.cabal

@@ -41,6 +41,8 @@ library
                      , vector
                      , containers
                      , filepath
+                     , parallel
+                     , unix
   hs-source-dirs:      src
   default-language:    Haskell2010
   ghc-options:         -Wall
@@ -55,7 +57,7 @@ executable addressbook
                      , text
   hs-source-dirs:      app
   default-language:    Haskell2010
-  ghc-options:         -Wall -threaded
+  ghc-options:         -Wall -threaded -eventlog
 
 test-suite addressbook-test
   import:              deps

default.nix

@@ -1,7 +1,7 @@
 { mkDerivation, attoparsec, base, bytestring, conduit
 , conduit-extra, containers, criterion, filepath, hedgehog
 , hedgehog-corpus, HUnit, lens, lib, mtl, optparse-applicative
-, tasty, tasty-hedgehog, tasty-hunit, text, vector
+, parallel, tasty, tasty-hedgehog, tasty-hunit, text, unix, vector
 }:
 mkDerivation {
   pname = "addressbook";
@@ -11,10 +11,11 @@ mkDerivation {
   isExecutable = true;
   libraryHaskellDepends = [
     attoparsec base bytestring conduit conduit-extra containers
-    filepath lens mtl text vector
+    filepath lens mtl parallel text unix vector
   ];
   executableHaskellDepends = [
-    base criterion hedgehog-corpus optparse-applicative text
+    base bytestring containers criterion hedgehog-corpus
+    optparse-applicative text
   ];
   testHaskellDepends = [
     base bytestring conduit conduit-extra containers hedgehog

flake.lock

@@ -0,0 +1,62 @@
+{
+  "nodes": {
+    "easy-hls": {
+      "inputs": {
+        "nixpkgs": [
+          "nixpkgs"
+        ]
+      },
+      "locked": {
+        "lastModified": 1637250802,
+        "narHash": "sha256-/crlHEVB148PGQLZCsHOR9L5qgvCAfRSocIoKgmMAhA=",
+        "owner": "jkachmar",
+        "repo": "easy-hls-nix",
+        "rev": "7c123399ef8a67dc0e505d9cf7f2c7f64f1cd847",
+        "type": "github"
+      },
+      "original": {
+        "owner": "jkachmar",
+        "repo": "easy-hls-nix",
+        "type": "github"
+      }
+    },
+    "flake-utils": {
+      "locked": {
+        "lastModified": 1637014545,
+        "narHash": "sha256-26IZAc5yzlD9FlDT54io1oqG/bBoyka+FJk5guaX4x4=",
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "rev": "bba5dcc8e0b20ab664967ad83d24d64cb64ec4f4",
+        "type": "github"
+      },
+      "original": {
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "type": "github"
+      }
+    },
+    "nixpkgs": {
+      "locked": {
+        "lastModified": 1637312849,
+        "narHash": "sha256-OhVZopkyryEfLyPwcXk2IQsdi80lj6TY1YFoMNZ4hCQ=",
+        "owner": "NixOS",
+        "repo": "nixpkgs",
+        "rev": "e4806bb4416f88c20f8be0b8ef9b5b09ff9022a6",
+        "type": "github"
+      },
+      "original": {
+        "id": "nixpkgs",
+        "type": "indirect"
+      }
+    },
+    "root": {
+      "inputs": {
+        "easy-hls": "easy-hls",
+        "flake-utils": "flake-utils",
+        "nixpkgs": "nixpkgs"
+      }
+    }
+  },
+  "root": "root",
+  "version": 7
+}

flake.nix

@@ -0,0 +1,53 @@
+{
+  description = "addressbook";
+
+  inputs = {
+    easy-hls = {
+      url = "github:jkachmar/easy-hls-nix";
+      inputs.nixpkgs.follows = "nixpkgs";
+    };
+    flake-utils = {
+      url = "github:numtide/flake-utils";
+    };
+  };
+
+  outputs = { self, nixpkgs, flake-utils, easy-hls }:
+    { overlay = final: prev: {
+        haskellPackages = prev.haskellPackages.override ( old: {
+          overrides = final.lib.composeExtensions (old.overrides or (_: _: {})) (f: p: {
+            addressbook = f.callPackage ./. {};
+          });
+        });
+      };
+    }
+    //
+    flake-utils.lib.eachSystem ["x86_64-linux" "x86_64-darwin"] ( system:
+      let
+        pkgs = import nixpkgs { inherit system; overlays = [ self.overlay ]; };
+        hp = pkgs.haskellPackages;
+        hls = (easy-hls.withGhcs [ hp.ghc ]).${system};
+      in
+      rec {
+
+        packages = { inherit (hp) addressbook; };
+
+        defaultPackage = packages.addressbook;
+        apps.addressbook = {
+          type = "app";
+          program = "${hp.addressbook}/bin/addressbook";
+        };
+        devShell = hp.shellFor {
+          packages = h: [h.addressbook];
+          withHoogle = true;
+          buildInputs = with pkgs; [
+            entr
+            cabal-install
+            hp.hlint
+            stylish-haskell
+            ghcid
+            hls
+          ];
+        };
+      }
+    );
+}

src/Control/Addressbook/Streaming.hs

@@ -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 +20,54 @@ 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
+import Control.Monad (unless)
+import System.Posix (touchFile)
 
 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
+  let datFile = datDir </> ".addressbook.dat"
-    x <- runConduit stream
+  touchFile datFile
-    runConduit (CL.sourceList (Map.keys x) .| C.map (<> "\n") .| CB.sinkFileCautious (datDir </> ".addressbook.dat"))
+  original <- Set.fromList . map LBS.toStrict . lbsLines <$> LBS.readFile datFile
+  xs <- LBS.getContents >>= stream
+  let set = original `Set.union` F.fold (parMap rseq F.fold (chunks 200 xs))
+  unless (original == set) $
+    runResourceT $
+      runConduit $
+           CL.sourceList (Set.elems set)
+        .| C.map (<> "\n")
+        .| CB.sinkFileCautious datFile
   where
     separate = \case
       From x -> [x]
       To xs -> F.toList xs
+    -- A set of (locally) unique addresses. Composes with parMap
+    lbsLines = LBS.split (fromIntegral $ ord '\n')
+    stream :: LBS.ByteString -> IO [Set ByteString]
     stream =
-      CB.sourceHandle stdin
+        traverse (unsafeInterleaveIO . parse . LBC.unpack)
-        .| CT.decode CT.utf8
+      . filter (not . LBS.null)
-        .| CT.lines
+      . lbsLines
-        .| C.map T.unpack
+    parse path =
-        .| combine
+      runResourceT $
-        .| C.concatMap separate
+        runConduit $
-        .| CT.encode CT.utf8
+          CB.sourceFile path .| parseEmail .| C.concatMap separate .| C.foldMap Set.singleton
-        .| C.foldMap (`Map.singleton` ())

src/Data/Email.hs

@@ -5,12 +5,9 @@ import Data.Email.Header
 
 import Conduit
 import qualified Data.Conduit.Combinators as C
-import qualified Data.Conduit.Text as CT
 
 import Data.ByteString
        (ByteString)
 
 parseEmail :: (MonadUnliftIO m, MonadThrow m, Monad m) => ConduitM ByteString Header m ()
-parseEmail = catchC (CT.decode CT.utf8) err .| CT.lines .| C.concatMap decode
+parseEmail = C.linesUnboundedAscii .| C.concatMap decode
-  where
-    err e = liftIO (print @CT.TextException e) >> yield ""

src/Data/Email/Header.hs

@@ -1,55 +1,63 @@
+{-# LANGUAGE OverloadedStrings #-}
+{-# LANGUAGE ApplicativeDo #-}
 module Data.Email.Header where
 
-import Data.Text
-       (Text)
-import qualified Data.Text as T
-
 import qualified Data.Foldable as F
 
-import Data.Attoparsec.Text
+import Data.Attoparsec.ByteString.Char8
 
 import Data.Vector
        (Vector)
 import qualified Data.Vector as V
 
-import Data.Char
-       (isSpace)
-
 import Control.Applicative
        ((<|>))
+import Data.ByteString (ByteString)
+import qualified Data.ByteString.Char8 as BC
 
 data Header
-  = From !Text
+  = From !ByteString
-  | To !(Vector Text)
+  | To !(Vector ByteString)
   deriving (Show, Eq)
 
-decode :: Text -> Either String Header
+decode :: ByteString -> Either String Header
 decode = parseOnly parseHeader
-  where
+{-# INLINE decode #-}
+
 parseHeader :: Parser Header
 parseHeader = parseFrom <|> parseTo
+{-# INLINE parseHeader #-}
+
 parseFrom :: Parser Header
 parseFrom = From <$> (string "From:" *> emptySpace *> (bracketEmail <|> email))
+
 parseTo :: Parser Header
 parseTo = To <$> (string "To:" *> emptySpace *> emails)
-    emptySpace = many' space
+
-    emails :: Parser (Vector Text)
+emptySpace :: Parser ()
+emptySpace = () <$ many' space
+
+emails :: Parser (Vector ByteString)
 emails = V.fromList <$> (bracketEmail <|> email) `sepBy` char ','
-    bracketEmail :: Parser Text
+
+bracketEmail :: Parser ByteString
 bracketEmail = do
   _ <- manyTill anyChar (char '<')
   email
-    email :: Parser Text
+{-# INLINE bracketEmail #-}
+
+email :: Parser ByteString
 email = do
   _ <- many' space
-      name <- T.pack <$> many' (satisfy (\c -> not (isSpace c) && c /= '@'))
+  name <- BC.pack <$> many' (satisfy (\c -> not (isSpace c) && c /= '@'))
   _ <- char '@'
-      rest <- T.pack <$> many' (satisfy (\c -> not (isSpace c) && c /= ',' && c /= '>'))
+  rest <- BC.pack <$> many' (satisfy (\c -> not (isSpace c) && c /= ',' && c /= '>' && c /= '<'))
   _ <- many' (notChar ',')
   pure (name <> "@" <> rest)
+{-# INLINE email #-}
 
 
-encode :: Header -> Text
+encode :: Header -> ByteString
 encode = \case
   From addr -> "From: " <> addr
-  To addrs -> "To: " <> T.intercalate ", " (F.toList addrs)
+  To addrs -> "To: " <> BC.intercalate ", " (F.toList addrs)

test/Test/Data/Email.hs

@@ -31,4 +31,7 @@ tests = testGroup "Data.Email"
   [ testCase "Can parse a sample email" $ do
       got <- parseToList sample
       got @?= [ From "me@example.com", To ["you@example.com"]]
+  , testCase "Combined mailto" $ do
+      let got = decode "To: James Doe<james.doe@example.com<mailto:james.doe@example.com>>"
+      got @?= Right (To ["james.doe@example.com"])
   ]

test/Test/Data/Email/Header.hs

@@ -10,11 +10,12 @@ import qualified Hedgehog.Corpus as Corpus
 import qualified Hedgehog.Gen as Gen
 import qualified Hedgehog.Range as Range
 
-import Data.Text
+import Data.ByteString
-import qualified Data.Text as T
+import qualified Data.ByteString as T
 import qualified Data.Vector as V
 
 import Data.Email.Header
+import qualified Data.ByteString.Char8 as BC
 
 genHeader :: Gen Header
 genHeader = Gen.choice
@@ -22,17 +23,17 @@ genHeader = Gen.choice
   , To . V.fromList <$> Gen.list (Range.linear 0 10) genEmail
   ]
 
-genEmail :: Gen Text
+genEmail :: Gen ByteString
 genEmail = do
   name <- Gen.element Corpus.simpsons
   domain <- Gen.element Corpus.cooking
   tld <- Gen.element ["com","fi","org"]
   pure $ name <> "@" <> domain <> "." <> tld
 
-wrapped :: Char -> Text -> Char -> Text
+wrapped :: Char -> ByteString -> Char -> ByteString
-wrapped l x r = T.singleton l <> x <> T.singleton r
+wrapped l x r = BC.singleton l <> x <> BC.singleton r
 
-genComment :: Gen Text
+genComment :: Gen ByteString
 genComment = do
   x <- Gen.element Corpus.simpsons
   Gen.element [x, wrapped '"' x '"']