module Rakka.Page ( PageName , Page(..) , LanguageTag , LanguageName , encodePageName , decodePageName , pageFileName' , defaultFileName , mkPageURI , mkPageFragmentURI , mkObjectURI , mkFragmentURI , mkAuxiliaryURI , mkRakkaURI , xmlizePage , parseXmlizedPage ) where import qualified Codec.Binary.Base64 as B64 import Codec.Binary.UTF8.String import Control.Arrow import Control.Arrow.ArrowIO import Control.Arrow.ArrowList import qualified Data.ByteString.Lazy as Lazy (ByteString) import qualified Data.ByteString.Lazy as L hiding (ByteString) import Data.Char import Data.Map (Map) import qualified Data.Map as M import Data.Maybe import Data.Time import Network.HTTP.Lucu hiding (redirect) import Network.URI hiding (fragment) import Rakka.Utils import Rakka.W3CDateTime import Subversion.Types import System.FilePath.Posix import Text.XML.HXT.Arrow.XmlArrow import Text.XML.HXT.Arrow.XmlNodeSet import Text.XML.HXT.DOM.TypeDefs type PageName = String type LanguageTag = String -- See RFC 3066: http://www.ietf.org/rfc/rfc3066.txt type LanguageName = String -- i.e. "日本語" data Page = Redirection { redirName :: !PageName , redirDest :: !PageName , redirRevision :: RevNum , redirLastMod :: UTCTime } | Entity { pageName :: !PageName , pageType :: !MIMEType , pageLanguage :: !(Maybe LanguageTag) , pageFileName :: !(Maybe String) , pageIsTheme :: !Bool -- text/css 以外では無意味 , pageIsFeed :: !Bool -- text/x-rakka 以外では無意味 , pageIsLocked :: !Bool , pageIsBoring :: !Bool , pageIsBinary :: !Bool , pageRevision :: RevNum , pageLastMod :: UTCTime , pageSummary :: !(Maybe String) , pageOtherLang :: !(Map LanguageTag PageName) , pageContent :: !Lazy.ByteString } deriving (Show, Eq) -- UTF-8 に encode してから 0x20 - 0x7E の範圍を除いて URI escape する。 encodePageName :: PageName -> FilePath encodePageName = escapeURIString isSafeChar . encodeString . fixPageName where fixPageName :: PageName -> PageName fixPageName = (\ (x:xs) -> toUpper x : xs) . map (\ c -> if c == ' ' then '_' else c) isSafeChar :: Char -> Bool isSafeChar c | c == '/' = True | isReserved c = False | c > ' ' && c <= '~' = True | otherwise = False -- URI unescape して UTF-8 から decode する。 decodePageName :: FilePath -> PageName decodePageName = decodeString . unEscapeString encodeFragment :: String -> String encodeFragment = escapeURIString isSafeChar . encodeString pageFileName' :: Page -> String pageFileName' page = fromMaybe (defaultFileName (pageType page) (pageName page)) (pageFileName page) defaultFileName :: MIMEType -> PageName -> String defaultFileName pType pName = let baseName = takeFileName pName in case pType of MIMEType "text" "x-rakka" _ -> baseName <.> "rakka" MIMEType "text" "css" _ -> baseName <.> "css" _ -> baseName mkPageURI :: URI -> PageName -> URI mkPageURI baseURI name = baseURI { uriPath = foldl () "/" [uriPath baseURI, encodePageName name ++ ".html"] } mkPageFragmentURI :: URI -> PageName -> String -> URI mkPageFragmentURI baseURI name fragment = baseURI { uriPath = foldl () "/" [uriPath baseURI, encodePageName name ++ ".html"] , uriFragment = ('#' : encodeFragment fragment) } mkFragmentURI :: String -> URI mkFragmentURI fragment = nullURI { uriFragment = ('#' : encodeFragment fragment) } mkObjectURI :: URI -> PageName -> URI mkObjectURI baseURI name = mkAuxiliaryURI baseURI ["object"] name mkAuxiliaryURI :: URI -> [String] -> PageName -> URI mkAuxiliaryURI baseURI basePath name = baseURI { uriPath = foldl () "/" ([uriPath baseURI] ++ basePath ++ [encodePageName name]) } mkRakkaURI :: PageName -> URI mkRakkaURI name = URI { uriScheme = "rakka:" , uriAuthority = Nothing , uriPath = encodePageName name , uriQuery = "" , uriFragment = "" } {- -- デフォルトでない場合のみ存在 lastModified="2000-01-01T00:00:00"> blah blah... -- 存在しない場合もある -- 存在しない場合もある blah blah... SKJaHKS8JK/DH8KS43JDK2aKKaSFLLS... -} xmlizePage :: (ArrowXml a, ArrowChoice a, ArrowIO a) => a Page XmlTree xmlizePage = proc page -> do lastMod <- arrIO (utcToLocalZonedTime . pageLastMod) -< page ( eelem "/" += ( eelem "page" += sattr "name" (pageName page) += sattr "type" (show $ pageType page) += ( case pageLanguage page of Just x -> sattr "lang" x Nothing -> none ) += ( case pageFileName page of Just x -> sattr "fileName" x Nothing -> none ) += ( case pageType page of MIMEType "text" "css" _ -> sattr "isTheme" (yesOrNo $ pageIsTheme page) MIMEType "text" "x-rakka" _ -> sattr "isFeed" (yesOrNo $ pageIsFeed page) _ -> none ) += sattr "isLocked" (yesOrNo $ pageIsLocked page) += sattr "isBoring" (yesOrNo $ pageIsBoring page) += sattr "isBinary" (yesOrNo $ pageIsBinary page) += sattr "revision" (show $ pageRevision page) += sattr "lastModified" (formatW3CDateTime lastMod) += ( case pageSummary page of Just s -> eelem "summary" += txt s Nothing -> none ) += ( if M.null (pageOtherLang page) then none else selem "otherLang" [ eelem "link" += sattr "lang" lang += sattr "page" name | (lang, name) <- M.toList (pageOtherLang page) ] ) += ( if pageIsBinary page then ( eelem "binaryData" += txt (B64.encode $ L.unpack $ pageContent page) ) else ( eelem "textData" += txt (decode $ L.unpack $ pageContent page) ) ) )) -<< () parseXmlizedPage :: (ArrowXml a, ArrowChoice a) => a (PageName, XmlTree) Page parseXmlizedPage = proc (name, tree) -> do redirect <- maybeA (getXPathTreesInDoc "/page/@redirect/text()" >>> getText) -< tree case redirect of Nothing -> parseEntity -< (name, tree) Just dest -> returnA -< (Redirection { redirName = name , redirDest = dest , redirRevision = undefined , redirLastMod = undefined }) parseEntity :: (ArrowXml a, ArrowChoice a) => a (PageName, XmlTree) Page parseEntity = proc (name, tree) -> do mimeType <- (getXPathTreesInDoc "/page/@type/text()" >>> getText >>> arr read) -< tree lang <- maybeA (getXPathTreesInDoc "/page/@lang/text()" >>> getText) -< tree fileName <- maybeA (getXPathTreesInDoc "/page/@filename/text()" >>> getText) -< tree isTheme <- (withDefault (getXPathTreesInDoc "/page/@isTheme/text()" >>> getText) "no" >>> parseYesOrNo) -< tree isFeed <- (withDefault (getXPathTreesInDoc "/page/@isFeed/text()" >>> getText) "no" >>> parseYesOrNo) -< tree isLocked <- (withDefault (getXPathTreesInDoc "/page/@isLocked/text()" >>> getText) "no" >>> parseYesOrNo) -< tree isBoring <- (withDefault (getXPathTreesInDoc "/page/@isBoring/text()" >>> getText) "no" >>> parseYesOrNo) -< tree summary <- (maybeA (getXPathTreesInDoc "/page/summary/text()" >>> getText >>> deleteIfEmpty)) -< tree otherLang <- listA (getXPathTreesInDoc "/page/otherLang/link" >>> (getAttrValue0 "lang" &&& getAttrValue0 "page")) -< tree textData <- maybeA (getXPathTreesInDoc "/page/textData/text()" >>> getText) -< tree binaryData <- maybeA (getXPathTreesInDoc "/page/binaryData/text()" >>> getText) -< tree let (isBinary, content) = case (textData, binaryData) of (Just text, Nothing ) -> (False, L.pack $ encode text ) (Nothing , Just binary) -> (True , L.pack $ B64.decode binary) _ -> error "one of textData or binaryData is required" returnA -< Entity { pageName = name , pageType = mimeType , pageLanguage = lang , pageFileName = fileName , pageIsTheme = isTheme , pageIsFeed = isFeed , pageIsLocked = isLocked , pageIsBoring = isBoring , pageIsBinary = isBinary , pageRevision = undefined , pageLastMod = undefined , pageSummary = summary , pageOtherLang = M.fromList otherLang , pageContent = content }