diff --git a/demo-ints/Main.hs b/demo-ints/Main.hs new file mode 100644 --- /dev/null +++ b/demo-ints/Main.hs @@ -0,0 +1,71 @@ +module Main where + +import System.Random (RandomGen, mkStdGen, randomR) + +import Data.IntSet (IntSet) +import qualified Data.IntSet as IntSet +import qualified Data.PerfectHash.Construction as Construction +import qualified Data.PerfectHash.Hashing as Hashing +import qualified Data.PerfectHash.Lookup as Lookup +import qualified Data.Vector.Unboxed as Vector +import Exercise (Atom (Atom)) +import qualified Exercise + + +valueCount = 500000 + +randomRange = (0, Hashing.mask32bits) + + +data RandIntAccum t = RandIntAccum + t -- ^ random number generator + Int -- ^ max count + IntSet -- ^ accumulated unique random numbers + + +-- | Since computing the size of the set is O(N), we +-- maintain the count separately. +getUniqueRandomIntegers :: RandomGen t => RandIntAccum t -> IntSet +getUniqueRandomIntegers (RandIntAccum std_gen count current_set) = + + if count == 0 + then current_set + else getUniqueRandomIntegers newstate + + where + (next_int, next_std_gen) = randomR randomRange std_gen + + a = RandIntAccum next_std_gen + newstate = if IntSet.member next_int current_set + then a count current_set + else a (count - 1) (IntSet.insert next_int current_set) + + +intMapTuples :: [(Atom Int, Int)] +intMapTuples = zip (map Atom random_ints) [1..] + where + seed_value = RandIntAccum (mkStdGen 0) valueCount IntSet.empty + random_ints = IntSet.toList $ getUniqueRandomIntegers seed_value + + +main = do + + putStrLn $ unwords ["Keys size:", show $ length intMapTuples] + + let lookup_table = Construction.createMinimalPerfectHash intMapTuples + + putStrLn $ unwords [ + "Finished computing lookup table with" + , show $ Lookup.size lookup_table + , "entries." + ] + + let direct_mapping_nonces = Vector.filter (< 0) $ Lookup.nonces lookup_table + + putStrLn $ unwords [ + "There were" + , show $ Vector.length direct_mapping_nonces + , "lookup entries with direct mappings." + ] + + Exercise.eitherExit $ Exercise.testLookups lookup_table intMapTuples diff --git a/demo-strings/Main.hs b/demo-strings/Main.hs new file mode 100644 --- /dev/null +++ b/demo-strings/Main.hs @@ -0,0 +1,33 @@ +module Main where + +import Control.Monad (when) + +import qualified Data.PerfectHash.Construction as Construction +import qualified Data.PerfectHash.Lookup as Lookup +import qualified Exercise + + +enableDebug = False + +dictionaryPath = "/usr/share/dict/words" + + +main = do + + word_index_tuples <- Exercise.wordsFromFile dictionaryPath + + putStrLn $ unwords ["Words size:", show $ length word_index_tuples] + + let lookup_table = Construction.createMinimalPerfectHash word_index_tuples + + putStrLn $ unwords [ + "Finished computing lookup table with" + , show $ Lookup.size lookup_table + , "entries." + ] + + when enableDebug $ do + putStrLn $ unwords ["Vector G:", show $ Lookup.nonces lookup_table] + putStrLn $ unwords ["Vector V:", show $ Lookup.values lookup_table] + + Exercise.eitherExit $ Exercise.testLookups lookup_table word_index_tuples diff --git a/demo/Ints/Main.hs b/demo/Ints/Main.hs deleted file mode 100644 --- a/demo/Ints/Main.hs +++ /dev/null @@ -1,71 +0,0 @@ -module Main where - -import System.Random (RandomGen, mkStdGen, randomR) - -import Data.IntSet (IntSet) -import qualified Data.IntSet as IntSet -import qualified Data.PerfectHash.Construction as Construction -import qualified Data.PerfectHash.Hashing as Hashing -import qualified Data.PerfectHash.Lookup as Lookup -import qualified Data.Vector.Unboxed as Vector -import Exercise (Atom (Atom)) -import qualified Exercise - - -valueCount = 500000 - -randomRange = (0, Hashing.mask32bits) - - -data RandIntAccum t = RandIntAccum - t -- ^ random number generator - Int -- ^ max count - IntSet -- ^ accumulated unique random numbers - - --- | Since computing the size of the set is O(N), we --- maintain the count separately. -getUniqueRandomIntegers :: RandomGen t => RandIntAccum t -> IntSet -getUniqueRandomIntegers (RandIntAccum std_gen count current_set) = - - if count == 0 - then current_set - else getUniqueRandomIntegers newstate - - where - (next_int, next_std_gen) = randomR randomRange std_gen - - a = RandIntAccum next_std_gen - newstate = if IntSet.member next_int current_set - then a count current_set - else a (count - 1) (IntSet.insert next_int current_set) - - -intMapTuples :: [(Atom Int, Int)] -intMapTuples = zip (map Atom random_ints) [1..] - where - seed_value = RandIntAccum (mkStdGen 0) valueCount IntSet.empty - random_ints = IntSet.toList $ getUniqueRandomIntegers seed_value - - -main = do - - putStrLn $ unwords ["Keys size:", show $ length intMapTuples] - - let lookup_table = Construction.createMinimalPerfectHash intMapTuples - - putStrLn $ unwords [ - "Finished computing lookup table with" - , show $ Lookup.size lookup_table - , "entries." - ] - - let direct_mapping_nonces = Vector.filter (< 0) $ Lookup.nonces lookup_table - - putStrLn $ unwords [ - "There were" - , show $ Vector.length direct_mapping_nonces - , "lookup entries with direct mappings." - ] - - Exercise.eitherExit $ Exercise.testLookups lookup_table intMapTuples diff --git a/demo/Strings/Main.hs b/demo/Strings/Main.hs deleted file mode 100644 --- a/demo/Strings/Main.hs +++ /dev/null @@ -1,33 +0,0 @@ -module Main where - -import Control.Monad (when) - -import qualified Data.PerfectHash.Construction as Construction -import qualified Data.PerfectHash.Lookup as Lookup -import qualified Exercise - - -enableDebug = False - -dictionaryPath = "/usr/share/dict/words" - - -main = do - - word_index_tuples <- Exercise.wordsFromFile dictionaryPath - - putStrLn $ unwords ["Words size:", show $ length word_index_tuples] - - let lookup_table = Construction.createMinimalPerfectHash word_index_tuples - - putStrLn $ unwords [ - "Finished computing lookup table with" - , show $ Lookup.size lookup_table - , "entries." - ] - - when enableDebug $ do - putStrLn $ unwords ["Vector G:", show $ Lookup.nonces lookup_table] - putStrLn $ unwords ["Vector V:", show $ Lookup.values lookup_table] - - Exercise.eitherExit $ Exercise.testLookups lookup_table word_index_tuples diff --git a/perfect-hash-generator.cabal b/perfect-hash-generator.cabal --- a/perfect-hash-generator.cabal +++ b/perfect-hash-generator.cabal @@ -3,7 +3,7 @@ -- see: https://github.com/sol/hpack name: perfect-hash-generator -version: 0.1.0.3 +version: 0.1.0.4 synopsis: Perfect minimal hashing implementation in native Haskell description: A for a set @S@ is a hash function that maps distinct elements in @S@ to a set of integers, with __no collisions__. A is a perfect hash function that maps @n@ keys to @n@ __consecutive__ integers, e.g. the numbers from @0@ to @n-1@. . @@ -71,9 +71,9 @@ default-language: Haskell2010 executable hash-perfectly-ints-demo - main-is: Ints/Main.hs + main-is: Main.hs hs-source-dirs: - demo + demo-ints test ghc-options: -fwarn-tabs -W build-depends: @@ -86,15 +86,13 @@ , hashable , containers other-modules: - Strings.Main Exercise - Main default-language: Haskell2010 executable hash-perfectly-strings-demo - main-is: Strings/Main.hs + main-is: Main.hs hs-source-dirs: - demo + demo-strings test ghc-options: -fwarn-tabs -W build-depends: @@ -106,9 +104,7 @@ , vector , hashable other-modules: - Ints.Main Exercise - Main default-language: Haskell2010 test-suite regression-tests diff --git a/src/Data/PerfectHash/Lookup.hs b/src/Data/PerfectHash/Lookup.hs --- a/src/Data/PerfectHash/Lookup.hs +++ b/src/Data/PerfectHash/Lookup.hs @@ -1,7 +1,7 @@ {-# OPTIONS_HADDOCK prune #-} -- | Note that what is referred to as a \"nonce\" in this library may be --- equivalently described as a \"salt\" by some. +-- referred to by some as a \"salt\". module Data.PerfectHash.Lookup ( LookupTable (LookupTable) , nonces