diff --git a/kanji.cabal b/kanji.cabal --- a/kanji.cabal +++ b/kanji.cabal @@ -1,90 +1,77 @@ --- This file has been generated from package.yaml by hpack version 0.28.2. --- --- see: https://github.com/sol/hpack --- --- hash: 18462ca584f835c151a09c99a797ee4947a3f928d05f706e3723007460f0e083 +cabal-version: 2.2 -name: kanji -version: 3.4.0 -synopsis: Perform 漢字検定 (Japan Kanji Aptitude Test) level analysis on Japanese Kanji -description: Perform 漢字検定 (Japan Kanji Aptitude Test) level analysis on Japanese Kanji. -category: Data -homepage: https://github.com/fosskers/kanji -author: Colin Woodbury -maintainer: colingw@gmail.com -copyright: 2011 - 2018 Colin Woodbury -license: BSD3 -license-file: LICENSE -build-type: Simple -cabal-version: >= 1.10 +name: kanji +version: 3.4.0.1 +synopsis: Perform 漢字検定 (Japan Kanji Aptitude Test) level analysis on Japanese Kanji +description: Perform 漢字検定 (Japan Kanji Aptitude Test) level analysis on Japanese Kanji. +category: Data +homepage: https://github.com/fosskers/kanji +author: Colin Woodbury +maintainer: colin@fosskers.ca +copyright: 2011 - 2019 Colin Woodbury +license: BSD-3-Clause +license-file: LICENSE +build-type: Simple + extra-source-files: - CHANGELOG.md - README.md + README.md + CHANGELOG.md +common commons + default-language: Haskell2010 + build-depends: + base >= 4.7 && < 5 + , aeson >= 1.2 + , containers >= 0.5.8 + , text + ghc-options: + -Wall + -Wincomplete-record-updates + -Wincomplete-uni-patterns + -Wredundant-constraints + library - hs-source-dirs: - lib - ghc-options: -fwarn-unused-imports -fwarn-unused-binds -fwarn-name-shadowing -fwarn-unused-matches -fwarn-incomplete-patterns -fwarn-incomplete-uni-patterns + import: commons + hs-source-dirs: lib build-depends: - aeson >=1.2 && <1.4 - , base >=4.7 && <5 - , containers - , deepseq + deepseq , hashable - , text exposed-modules: Data.Kanji Data.Kanji.Levels Data.Kanji.Types - default-language: Haskell2010 executable kanji + import: commons main-is: nanq.hs - hs-source-dirs: - nanq - default-extensions: OverloadedStrings - ghc-options: -fwarn-unused-imports -fwarn-unused-binds -fwarn-name-shadowing -fwarn-unused-matches -fwarn-incomplete-patterns -fwarn-incomplete-uni-patterns -O2 + hs-source-dirs: nanq + ghc-options: -O2 build-depends: - aeson >=1.2 && <1.4 - , aeson-pretty - , base >=4.7 && <5 - , containers + aeson-pretty , kanji - , microlens >=0.4 && <0.5 - , microlens-aeson >=2.2 && <2.4 - , optparse-applicative >=0.14 && <0.15 - , text + , microlens >=0.4 + , microlens-aeson >=2.2 + , optparse-applicative >=0.14 , transformers - default-language: Haskell2010 test-suite kanji-test + import: commons type: exitcode-stdio-1.0 main-is: Test.hs - hs-source-dirs: - test - ghc-options: -fwarn-unused-imports -fwarn-unused-binds -fwarn-name-shadowing -fwarn-unused-matches -fwarn-incomplete-patterns -fwarn-incomplete-uni-patterns -threaded + hs-source-dirs: test + ghc-options: -threaded -with-rtsopts=-N build-depends: - HUnit-approx >=1.1 && <1.2 - , aeson >=1.2 && <1.4 - , base >=4.7 && <5 - , containers + HUnit-approx >=1.1 , kanji - , tasty >=0.11 && <1.2 - , tasty-hunit >=0.9 && <0.11 - , text - default-language: Haskell2010 + , tasty >=0.11 + , tasty-hunit >=0.9 benchmark kanji-bench + import: commons type: exitcode-stdio-1.0 main-is: Bench.hs - hs-source-dirs: - bench - ghc-options: -fwarn-unused-imports -fwarn-unused-binds -fwarn-name-shadowing -fwarn-unused-matches -fwarn-incomplete-patterns -fwarn-incomplete-uni-patterns -threaded -O2 + hs-source-dirs: bench + ghc-options: -threaded -O2 build-depends: - aeson >=1.2 && <1.4 - , base >=4.7 && <5 - , containers - , criterion >=1.1 && <1.5 + criterion >=1.1 , kanji - , text - default-language: Haskell2010 diff --git a/lib/Data/Kanji.hs b/lib/Data/Kanji.hs --- a/lib/Data/Kanji.hs +++ b/lib/Data/Kanji.hs @@ -2,43 +2,43 @@ -- | -- Module : Data.Kanji --- Copyright : (c) Colin Woodbury, 2015 - 2018 +-- Copyright : (c) Colin Woodbury, 2015 - 2019 -- License : GPL3 --- Maintainer: Colin Woodbury +-- Maintainer: Colin Woodbury -- -- A library for analysing the density of Kanji in given texts, -- according to their "Level" classification, as defined by the -- Japan Kanji Aptitude Testing Foundation (日本漢字能力検定協会). module Data.Kanji - ( - -- * Kanji - Kanji - , kanji, _kanji - , allKanji - , isKanji, isHiragana, isKatakana - -- * Character Categories - , CharCat(..) - , category - -- * Levels - , Level(..) - , level - -- * Analysis - , percentSpread - , levelDist - , uniques - -- ** Densities - , densities - , elementaryDen - , middleDen - , highDen - ) where + ( + -- * Kanji + Kanji + , kanji, _kanji + , allKanji + , isKanji, isHiragana, isKatakana + -- * Character Categories + , CharCat(..) + , category + -- * Levels + , Level(..) + , level + -- * Analysis + , percentSpread + , levelDist + , uniques + -- ** Densities + , densities + , elementaryDen + , middleDen + , highDen + ) where import Control.Arrow hiding (second) import Data.Foldable (fold) import Data.Kanji.Levels import Data.Kanji.Types -import Data.List (sort, group) +import Data.List (group, sort) import qualified Data.Map.Strict as M import Data.Semigroup ((<>)) import qualified Data.Set as S diff --git a/lib/Data/Kanji/Levels.hs b/lib/Data/Kanji/Levels.hs --- a/lib/Data/Kanji/Levels.hs +++ b/lib/Data/Kanji/Levels.hs @@ -1,8 +1,8 @@ -- | -- Module : Data.Kanji.Levels --- Copyright : (c) Colin Woodbury, 2015 - 2018 +-- Copyright : (c) Colin Woodbury, 2015 - 2019 -- License : GPL3 --- Maintainer: Colin Woodbury +-- Maintainer: Colin Woodbury -- -- All Kanji from levels 10 to Pre-2. diff --git a/lib/Data/Kanji/Types.hs b/lib/Data/Kanji/Types.hs --- a/lib/Data/Kanji/Types.hs +++ b/lib/Data/Kanji/Types.hs @@ -1,11 +1,12 @@ -{-# LANGUAGE DeriveAnyClass, DeriveGeneric #-} -{-# LANGUAGE ViewPatterns #-} +{-# LANGUAGE DeriveAnyClass #-} +{-# LANGUAGE DeriveGeneric #-} +{-# LANGUAGE ViewPatterns #-} -- | -- Module : Data.Kanji.Types -- Copyright : (c) Colin Woodbury, 2015, 2016 -- License : GPL3 --- Maintainer: Colin Woodbury +-- Maintainer: Colin Woodbury -- -- Types for this library. While a constructor for `Kanji` is made available -- here, you should prefer the `kanji` "smart constructor" unless you know @@ -17,7 +18,7 @@ import Data.Aeson import Data.Aeson.Encoding (text) import Data.Bool (bool) -import Data.Char (ord, isLetter, isNumber, isPunctuation) +import Data.Char (isLetter, isNumber, isPunctuation, ord) import Data.Hashable import qualified Data.Text as T import GHC.Generics diff --git a/nanq/nanq.hs b/nanq/nanq.hs --- a/nanq/nanq.hs +++ b/nanq/nanq.hs @@ -1,5 +1,6 @@ -{-# LANGUAGE FlexibleContexts #-} -{-# LANGUAGE TupleSections #-} +{-# LANGUAGE FlexibleContexts #-} +{-# LANGUAGE OverloadedStrings #-} +{-# LANGUAGE TupleSections #-} module Main ( main ) where @@ -26,20 +27,22 @@ data Operation = Density | Elementary | Distribution | Splits deriving (Eq) -data Env = Env { _allKs :: [Kanji] - , _original :: Text } deriving Eq +data Env = Env { _allKs :: [Kanji], _original :: Text } deriving Eq -- | Long, Short, Help +lsh :: HasName f => String -> Char -> String -> Mod f a lsh l s h = long l <> short s <> help h flags :: Parser Flags flags = Flags <$> operations <*> (file <|> japanese) - where file = Left <$> strOption (lsh "file" 'f' "Take input from a file") - japanese = Right <$> argument str (metavar "JAPANESE") + where + file = Left <$> strOption (lsh "file" 'f' "Take input from a file") + japanese = Right <$> argument str (metavar "JAPANESE") operations :: Parser [Operation] operations = catMaybes <$> ops - where ops = traverse optional + where + ops = traverse optional [ flag' Density $ lsh "density" 'd' "Find how much of the input is made of Kanji" , flag' Elementary $ @@ -55,7 +58,8 @@ splits :: Reader Env Value splits = ob "levelSplit" . S.foldl' f mempty . S.fromList <$> asks _allKs - where f a k = (\l -> M.insertWith (++) l [k] a) $ level k + where + f a k = (\l -> M.insertWith (++) l [k] a) $ level k distribution :: Reader Env Value distribution = ob "distributions" . levelDist <$> asks _allKs @@ -70,10 +74,10 @@ -- | All operations return JSON, to be aggregated into a master Object. execOp :: Operation -> Reader Env Value -execOp Density = density -execOp Elementary = elementaryDensity +execOp Density = density +execOp Elementary = elementaryDensity execOp Distribution = distribution -execOp Splits = splits +execOp Splits = splits output :: Value -> IO () output = TLIO.putStrLn . toLazyText . encodePrettyToTextBuilder @@ -81,15 +85,18 @@ -- | Dispatch on each `Operation` given. Aggregates the resulting JSON. work :: (Env, [Operation]) -> Value work (e, os) = Object $ vals ^. each . _Object - where vals = runReader (traverse execOp os) e + where + vals = runReader (traverse execOp os) e env :: Flags -> IO (Env, [Operation]) env (Flags os inp) = case inp of Right t -> pure (e t, os) Left f -> (, os) . e <$> TIO.readFile f - where e t = Env (mapMaybe kanji $ T.unpack t) t + where + e t = Env (mapMaybe kanji $ T.unpack t) t main :: IO () main = execParser opts >>= env >>= output . work - where opts = info (helper <*> flags) - (fullDesc <> header "nanq - Kanji analysis of Japanese text") + where + opts = info (helper <*> flags) + (fullDesc <> header "nanq - Kanji analysis of Japanese text")