diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index ea4f028..26f9626 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -9,8 +9,76 @@ on: - "**" jobs: - build: - name: Build + crawler: + name: Crawler Dry Run + runs-on: ubuntu-latest + strategy: + matrix: + python-version: ["3.11", "3.12", "3.13-dev"] + + steps: + - uses: actions/checkout@v4 + + - name: Setup Python ${{ matrix.python-version }} + uses: actions/setup-python@v4 + with: + python-version: ${{ matrix.python-version }} + + - name: Install Python dependencies + run: | + python -m pip install --upgrade pip + if [ -f requirements.txt ]; then pip install -r requirements.txt; fi + + - name: Run crawler script + shell: bash + run: python crawler.py + + parser: + name: Parser Dry Run + runs-on: ubuntu-latest + + strategy: + matrix: + ghc: ["9.2.8", "9.6", "latest"] + cabal: ["3.10.1.0", "3.10", "latest"] + exclude: + - ghc: latest + cabal: 3.10.1.0 + + steps: + - uses: actions/checkout@v4 + + - name: Setup Haskell ${{ matrix.ghc }} with Cabal ${{ matrix.cabal }} + uses: haskell-actions/setup@v2 + with: + ghc-version: ${{ matrix.ghc }} + cabal-version: ${{ matrix.cabal }} + + - name: Configure the build + run: | + cabal configure --enable-tests --enable-benchmarks --disable-documentation + cabal build --dry-run + + - name: Build + run: cabal build all + + - name: Run tests + run: cabal test all + + - name: Check cabal file + run: cabal check + + - name: Run parser + run: cabal run + + - name: Checkout files + run: | + pwd + ls -la + ls -la docs + + hlint: + name: HLint runs-on: ubuntu-latest defaults: @@ -18,24 +86,16 @@ jobs: working-directory: . steps: - - uses: actions/checkout@v3 + - uses: actions/checkout@v4 - - name: Setup Go - uses: actions/setup-go@v4 + - name: "Set up HLint" + uses: haskell-actions/hlint-setup@v2 + + - name: "Run HLint" + uses: haskell-actions/hlint-run@v2 with: - go-version-file: go.mod - - - name: Go Build - run: go build -v ./... - - - name: Go Test - run: go test ./... -shuffle on -count 10 -coverprofile cover.out - - - name: Go Vet - run: go test -v ./... -vet all - - - name: Coverage - run: go tool cover -func cover.out + path: parser/ + fail-on: warning golint: name: Golint @@ -53,21 +113,21 @@ jobs: pylint: name: Pylint runs-on: ubuntu-latest - strategy: - matrix: - python-version: [ "3.11", "3.12" ] steps: - - uses: actions/checkout@v3 - - name: Set up Python ${{ matrix.python-version }} + - uses: actions/checkout@v4 + + - name: Set up Python Environment uses: actions/setup-python@v4 with: - python-version: ${{ matrix.python-version }} - - name: Install dependencies + python-version: 3.12 + + - name: Install Python dependencies run: | python -m pip install --upgrade pip pip install pylint if [ -f requirements.txt ]; then pip install -r requirements.txt; fi - - name: Analysing the code with pylint + + - name: Analyzing the code with pylint run: | pylint $(git ls-files '*.py') diff --git a/parser.cabal b/parser.cabal index 5a9b445..1ea68dc 100644 --- a/parser.cabal +++ b/parser.cabal @@ -12,6 +12,10 @@ cabal-version: 3.0 -- -- The name of the package. name: parser +category: Data +synopsis: Chinese Holidays Calendar Parser and Generator +description: Calendar of Public Holidays in China + 中国大陆节假日日历订阅 -- The package version. -- See the Haskell package versioning policy (PVP) for standards @@ -20,7 +24,7 @@ name: parser -- PVP summary: +-+------- breaking API changes -- | | +----- non-breaking API additions -- | | | +--- code changes with no API change -version: 0.1.0.0 +version: 0.1.1.0 -- A short (one-line) description of the package. -- synopsis: @@ -53,7 +57,7 @@ build-type: Simple common warnings ghc-options: -Wall -executable temp +executable main -- Import common warning flags. import: warnings @@ -67,10 +71,33 @@ executable temp -- other-extensions: -- Other library packages from which modules are imported. - build-depends: base ^>=4.16.4.0, directory, filepath, split, time, uuid + build-depends: base >= 4.16.4.0 && < 4.20, + directory >= 1.3.6 && < 1.4, + filepath >= 1.4.2 && < 1.5, + split >= 0.2.4 && < 0.3, + time >= 1.11.1 && < 1.12, + uuid >= 1.3.15 && < 1.4 -- Directories containing source files. hs-source-dirs: parser -- Base language which the package is written in. default-language: Haskell2010 + +test-suite test + type: exitcode-stdio-1.0 + main-is: Test.hs + other-modules: Main.Base Main.Input Main.Output + build-depends: base >=4.16.4.0 && <4.20, + directory >= 1.3.6 && < 1.4, + filepath >= 1.4.2 && < 1.5, + split >= 0.2.4 && < 0.3, + time >= 1.11.1 && < 1.12, + uuid >= 1.3.15 && < 1.4, + HUnit >=1.6 && <1.7 + hs-source-dirs: parser + default-language: Haskell2010 + +source-repository head + type: git + location: https://github.com/theRank/chinese-holidays-calendar diff --git a/parser/Main.hs b/parser/Main.hs index 1412fc8..ecd0dfd 100644 --- a/parser/Main.hs +++ b/parser/Main.hs @@ -38,5 +38,5 @@ debug dataByYear = mapM_ showByYear $ sortByYear data' data' = map (\(y, r, w) -> (y, sortByDate $ r ++ w)) dataByYear sortByYear = sortBy (\(y1, _) (y2, _) -> compare y1 y2) showByYear (year, dates) = do - putStrLn $ "\nYear " ++ year - mapM (putStrLn . show) dates + putStrLn $ "\nYear" <> year + mapM_ print dates diff --git a/parser/Main/Base.hs b/parser/Main/Base.hs index 641c945..5be434c 100644 --- a/parser/Main/Base.hs +++ b/parser/Main/Base.hs @@ -1,49 +1,50 @@ module Main.Base where -import Data.List (intercalate, sortBy) +import Data.Function (on) +import Data.List (sortBy) import Data.Time (UTCTime, defaultTimeLocale, formatTime) -data DateDataType = Both | Rest | Work deriving (Enum) +data DateType = Both | Rest | Work deriving (Enum) -instance Show DateDataType where +instance Show DateType where show Both = "" show Rest = "假期" show Work = "补班" -- Title of output ics file -titleDateDataType :: DateDataType -> String -titleDateDataType Both = "节假日" -titleDateDataType flag = "节假日（" ++ show flag ++ "）" +titleDateType :: DateType -> String +titleDateType Both = "中国节假日安排" +titleDateType flag = "中国节假日安排（" <> show flag <> "）" -- Index of input txt file -indexDateDataType :: DateDataType -> Int -indexDateDataType Both = 0 -indexDateDataType Rest = 1 -indexDateDataType Work = 2 +indexDateType :: DateType -> Int +indexDateType Both = 0 +indexDateType Rest = 1 +indexDateType Work = 2 data Date = Date { name :: String, time :: UTCTime, - flag :: DateDataType, + flag :: DateType, index :: Int, total :: Int } instance Show Date where show (Date name time flag index total) = - intercalate " " $ + unwords [ formatTime defaultTimeLocale "%Y-%m-%d" time, name, show flag, - show index ++ "/" ++ show total + show index <> "/" <> show total ] -constructDate :: String -> DateDataType -> (Int, Int, UTCTime) -> Date +constructDate :: String -> DateType -> (Int, Int, UTCTime) -> Date constructDate name flag (index, total, time) = Date name time flag index total sortByDate :: [Date] -> [Date] -sortByDate = sortBy (\(Date _ t1 _ _ _) (Date _ t2 _ _ _) -> compare t1 t2) +sortByDate = sortBy (compare `on` time) -filterByType :: DateDataType -> [Date] -> [Date] +filterByType :: DateType -> [Date] -> [Date] filterByType Both = id -filterByType flag = filter (\(Date _ _ f _ _) -> indexDateDataType f == indexDateDataType flag) +filterByType flag = filter (\(Date _ _ f _ _) -> show f == show flag) diff --git a/parser/Main/Input.hs b/parser/Main/Input.hs index 5c370d8..fef2a5f 100644 --- a/parser/Main/Input.hs +++ b/parser/Main/Input.hs @@ -1,5 +1,6 @@ module Main.Input where +import Data.Char (isSpace) import Data.List.Split (splitOn) import Data.Time (UTCTime, addUTCTime, defaultTimeLocale, formatTime, parseTimeOrError) import Main.Base @@ -7,7 +8,7 @@ import System.FilePath (takeBaseName) -- Join data from each year and each type join :: [(String, [Date], [Date])] -> [Date] -join dataByFile = concatMap (\(_, rest, work) -> rest ++ work) dataByFile +join = concatMap (\ (_, rest, work) -> rest ++ work) -- Parse holiday data -- Organized by year @@ -19,21 +20,21 @@ parseByFile (file, content) = (year, rest, work) work = parse year content Work -- Convert data to Date -parse :: String -> String -> DateDataType -> [Date] +parse :: String -> String -> DateType -> [Date] parse year content flag = concatMap constructor $ zip (map head raw) dates where - constructor (name, dates) = map (constructDate name flag) dates - dates = map parseDate $ map (!! indexDateDataType flag) raw + constructor (name, dates) = constructDate name flag <$> dates + dates = parseDate <$> map (!! indexDateType flag) raw raw = parseFile content -- Parse data from file -- Result: [[Name, RestDays, WorkDays]] parseFile :: String -> [[String]] -parseFile content = holidays +parseFile content = filter ((== 3) . length) $ splitOn ";" <$> rawEvents where - holidays = filter ((== 3) . length) . map (splitOn ";") $ eachData - eachData = filter (not . null) . map (head . words) $ eachLine - eachLine = filter (not . null) . map (head . splitOn "//") $ lines content + rawEvents = filter (not . null) . map (head . words) $ uncomment + uncomment = filter (not . null) . map (head . splitOn "//") $ unindent + unindent = dropWhile isSpace <$> lines content -- Expand date ranges to UTCTime list -- Support multiple date ranges separated by comma @@ -47,17 +48,18 @@ parseDate range = zip3 [1 ..] (repeat $ length dates) dates -- 1. like "2020.1.1" -- 2. like "2020.1.1-2020.1.3" parseDate' :: [String] -> [UTCTime] -parseDate' [date] = [parseTime date] +parseDate' [single] = [parseTime single] parseDate' [start, end] | start == end = parseDate' [end] | otherwise = first : parseDate' [second, end] where first = parseTime start - second = printTime $ addUTCTime 86400 first + second = printTime $ addUTCTime day first + day = 24 * 60 * 60 -- Parse date in format "2020.1.1" parseTime :: String -> UTCTime -parseTime date = parseTimeOrError True defaultTimeLocale "%Y.%-m.%-d" date :: UTCTime +parseTime = parseTimeOrError True defaultTimeLocale "%Y.%-m.%-d" -- Format date in format "2020.1.1" printTime :: UTCTime -> String diff --git a/parser/Main/Output.hs b/parser/Main/Output.hs index 3ee2283..f91cf23 100644 --- a/parser/Main/Output.hs +++ b/parser/Main/Output.hs @@ -2,24 +2,24 @@ module Main.Output where import Data.Time (defaultTimeLocale, formatTime, nominalDiffTimeToSeconds) import Data.Time.Clock.POSIX (utcTimeToPOSIXSeconds) -import Data.UUID (fromWords64, toString) +import Data.UUID (fromWords, toString) import Main.Base import Text.Printf (printf) -- Generate ics files -icsByType :: DateDataType -> [Date] -> String +icsByType :: DateType -> [Date] -> String icsByType flag dates = unlines [icsHead flag, icsBody, icsTail] where icsBody = unlines $ map icsEvent $ sortByDate $ filterByType flag dates -- Standard ics format for the beginning -icsHead :: DateDataType -> String +icsHead :: DateType -> String icsHead flag = unlines [ "BEGIN:VCALENDAR", "VERSION:2.0", "PRODID:-//Rank Technology//Chinese Holidays//EN", - "X-WR-CALNAME:" ++ titleDateDataType flag + "X-WR-CALNAME:" <> titleDateType flag -- "X-WR-TIMEZONE:Asia/Shanghai", ] @@ -28,15 +28,18 @@ icsEvent :: Date -> String icsEvent (Date name time flag index total) = unlines [ "BEGIN:VEVENT", - "UID:" ++ uuid, - "DTSTART;VALUE=DATE:" ++ formatTime defaultTimeLocale "%Y%m%d" time, - "SUMMARY:" ++ name ++ show flag, - "DESCRIPTION:" ++ printf "%s 第%d天/共%d天" (show flag) index total, + "UID:" <> uuid, + "DTSTART;VALUE=DATE:" <> formatTime defaultTimeLocale "%Y%m%d" time, + "SUMMARY:" <> name <> show flag, + "DESCRIPTION:" <> show flag <> printf "第%d天 / 共%d天" index total, "END:VEVENT" ] where - uuid = toString $ fromWords64 1 t - t = floor $ nominalDiffTimeToSeconds $ utcTimeToPOSIXSeconds time + uuid = toString $ fromWords a b c d + a = floor . nominalDiffTimeToSeconds . utcTimeToPOSIXSeconds $ time + b = fromIntegral $ indexDateType flag + c = fromIntegral total + d = fromIntegral index -- Standard ics format for the ending icsTail :: String diff --git a/parser/Test.hs b/parser/Test.hs new file mode 100644 index 0000000..cab2e4c --- /dev/null +++ b/parser/Test.hs @@ -0,0 +1,213 @@ +module Main where + +import Main.Input +import System.Exit (exitFailure, exitSuccess) +import Test.HUnit + +testA1 = + TestCase + ( assertEqual + "A1 - case 1. like 2020.1.1 - basic" + ["2020.1.1"] + $ map printTime + $ parseDate' ["2020.1.1"] + ) + +testA2 = + TestCase + ( assertEqual + "A2 - case 1. like 2020.1.1 - robust" + ["2021.11.2"] + (map printTime $ parseDate' ["2021.11.02"]) + ) + +testA3 = + TestCase + ( assertEqual + "A3 - case 2. like 2020.1.1-2020.1.3 - basic" + ["2022.1.1", "2022.1.2", "2022.1.3"] + (map printTime $ parseDate' ["2022.1.1", "2022.1.3"]) + ) + +testA4 = + TestCase + ( assertEqual + "A4 - case 2. like 2020.1.1-2020.1.3 - cross month" + ["2023.2.27", "2023.2.28", "2023.3.1", "2023.3.2"] + (map printTime $ parseDate' ["2023.2.27", "2023.3.2"]) + ) + +testA5 = + TestCase + ( assertEqual + "A5 - case 2. like 2020.1.1-2020.1.3 - cross month 2" + ["2024.2.28", "2024.2.29", "2024.3.1"] + (map printTime $ parseDate' ["2024.2.28", "2024.3.1"]) + ) + +testA6 = + TestCase + ( assertEqual + "A6 - case 2. like 2020.1.1-2020.1.3 - cross year" + ["2025.12.30", "2025.12.31", "2026.1.1", "2026.1.2"] + (map printTime $ parseDate' ["2025.12.30", "2026.1.2"]) + ) + +testB1 = + TestCase + ( assertEqual + "B1 - case 1. like 2020.1.1 - a1 2" + [(1, 2, "2020.1.1"), (2, 2, "2020.1.3")] + $ map (\(a, b, c) -> (a, b, printTime c)) + $ parseDate "2020.1.1,2020.1.3" + ) + +testB2 = + TestCase + ( assertEqual + "B2 - case 1. like 2020.1.1 - a1 3" + [(1, 3, "2020.1.1"), (2, 3, "2020.1.3"), (3, 3, "2020.1.5")] + $ map (\(a, b, c) -> (a, b, printTime c)) + $ parseDate "2020.1.1,2020.1.3,2020.1.5" + ) + +testB3 = + TestCase + ( assertEqual + "B3 - case 2. like 2020.1.1-2020.1.3 - a3 2" + [(1, 4, "2020.1.1"), (2, 4, "2020.1.2"), (3, 4, "2020.1.6"), (4, 4, "2020.1.7")] + $ map (\(a, b, c) -> (a, b, printTime c)) + $ parseDate "2020.1.1-2020.1.2,2020.1.6-2020.1.7" + ) + +testB4 = + TestCase + ( assertEqual + "B4 - case 2. like 2020.1.1-2020.1.3 - a1 a3" + [(1, 4, "2020.1.1"), (2, 4, "2020.1.2"), (3, 4, "2020.1.3"), (4, 4, "2020.1.6")] + $ map (\(a, b, c) -> (a, b, printTime c)) + $ parseDate "2020.1.1-2020.1.3,2020.1.6" + ) + +testB5 = + TestCase + ( assertEqual + "B5 - case 2. like 2020.1.1-2020.1.3 - a3 a1" + [(1, 5, "2019.12.6"), (2, 5, "2020.1.1"), (3, 5, "2020.1.2"), (4, 5, "2020.1.3"), (5, 5, "2021.11.11")] + $ map (\(a, b, c) -> (a, b, printTime c)) + $ parseDate "2019.12.6,2020.1.1-2020.1.3,2021.11.11" + ) + +testC1 = + TestCase + ( assertEqual + "C1 - data - basic" + [["a", "b", "c"]] + $ parseFile "a;b;c" + ) + +testC2 = + TestCase + ( assertEqual + "C2 - data - robust" + [["a12", "b345", "c678"]] + $ parseFile " a12;b345;c678 " + ) + +testC3 = + TestCase + ( assertEqual + "C3 - comments - basic" + [] + $ parseFile "//comment" + ) + +testC4 = + TestCase + ( assertEqual + "C4 - comments - robust" + [] + $ parseFile " // a;b;c 123 " + ) + +testC5 = + TestCase + ( assertEqual + "C5 - hybrid - basic" + [["a", "b", "c"]] + $ parseFile "a;b;c // d;e;f 456 " + ) + +testC6 = + TestCase + ( assertEqual + "C6 - hybrid - robust" + [["a", "b", "c"]] + $ parseFile " a;b;c 123 // d;e;f 456 " + ) + +testC7 = + TestCase + ( assertEqual + "C7 - empty - basic" + [] + $ parseFile "" + ) + +testC8 = + TestCase + ( assertEqual + "C8 - empty - robust" + [] + $ parseFile " // " + ) + +testC9 = + TestCase + ( assertEqual + "C9 - multi-line" + [["a", "b", "c"], ["d", "e", "f"]] + $ parseFile . unlines + $ [ "// INTRO", + "", + "a;b;c // dp1", + "// comment", + "d;e;f // dp2", + "END // TEST" + ] + ) + +tests :: Test +tests = + TestList + [ -- A. parseDate' + TestLabel "Test parseDate' 1" testA1, + TestLabel "Test parseDate' 2" testA2, + TestLabel "Test parseDate' 3" testA3, + TestLabel "Test parseDate' 4" testA4, + TestLabel "Test parseDate' 5" testA5, + TestLabel "Test parseDate' 6" testA6, + -- B. parseDate + TestLabel "Test parseDate 1" testB1, + TestLabel "Test parseDate 2" testB2, + TestLabel "Test parseDate 3" testB3, + TestLabel "Test parseDate 4" testB4, + TestLabel "Test parseDate 5" testB5, + -- C. parseFile + TestLabel "Test parseFile 1" testC1, + TestLabel "Test parseFile 2" testC2, + TestLabel "Test parseFile 3" testC3, + TestLabel "Test parseFile 4" testC4, + TestLabel "Test parseFile 5" testC5, + TestLabel "Test parseFile 6" testC6, + TestLabel "Test parseFile 7" testC7, + TestLabel "Test parseFile 8" testC8, + TestLabel "Test parseFile 9" testC9 + ] + +main :: IO () +main = do + counts <- runTestTT tests + if errors counts + failures counts == 0 + then exitSuccess + else exitFailure