diff --git a/Setup.hs b/Setup.hs --- a/Setup.hs +++ b/Setup.hs @@ -4,35 +4,15 @@ -- License: BSD3 (see LICENSE) -- Author: Dino Morelli -import Control.Monad ( unless ) import Distribution.Simple import System.Cmd ( system ) -import System.Directory ( createDirectoryIfMissing ) -import System.FilePath -import System.Posix.Files ( createSymbolicLink, fileExist ) main = defaultMainWithHooks (simpleUserHooks - { postBuild = customPostBuild - , runTests = testRunner + { runTests = testRunner } ) where - -- Create symlink to the binary after build for developer - -- convenience - customPostBuild _ _ _ _ = do - let binDir = "bin" - let binName = "epubmeta" - let destPath = binDir binName - - createDirectoryIfMissing True binDir - - linkExists <- fileExist destPath - unless linkExists $ do - createSymbolicLink - (".." "dist" "build" binName binName) - destPath - -- Target for running all unit tests testRunner _ _ _ _ = do system $ "runhaskell -isrc testsuite/runtests.hs" diff --git a/epub-metadata.cabal b/epub-metadata.cabal --- a/epub-metadata.cabal +++ b/epub-metadata.cabal @@ -1,5 +1,5 @@ name: epub-metadata -version: 2.0.2 +version: 2.1.0 cabal-version: >= 1.2 build-type: Simple license: BSD3 @@ -17,7 +17,7 @@ utility to dump OPF package data to stdout in a human-readable form. category: Codec, Text -tested-with: GHC>=6.12.1 +tested-with: GHC >= 6.12.3 library exposed-modules: Codec.Epub.IO, @@ -35,8 +35,8 @@ Codec.Epub.Opf.Package.Spine, Codec.Epub.Opf.Parse hs-source-dirs: src - build-depends: base >= 3 && < 5, containers, hxt >= 9, LibZip, - mtl, regex-compat + build-depends: base >= 3 && < 5, bytestring, containers, + hxt >= 9, mtl, regex-compat, zip-archive ghc-options: -Wall executable epubmeta diff --git a/src/Codec/Epub/IO.hs b/src/Codec/Epub/IO.hs --- a/src/Codec/Epub/IO.hs +++ b/src/Codec/Epub/IO.hs @@ -5,17 +5,15 @@ {-# LANGUAGE FlexibleContexts #-} {- | Functions for doing some disk IO with ePub documents - - Note that these functions do their work by using the external - unzip utility. -} module Codec.Epub.IO - ( extractFileFromZip, opfPath ) + ( opfContents ) where -import Codec.Archive.LibZip +import Codec.Archive.Zip import Control.Arrow.ListArrows ( (>>>), deep ) import Control.Monad.Error +import qualified Data.ByteString.Lazy.Char8 as B import Text.Regex import Text.XML.HXT.Arrow.XmlArrow ( getAttrValue, hasName, isElem ) import Text.XML.HXT.Arrow.XmlState ( no, runX, withValidate ) @@ -34,37 +32,43 @@ (mkRegexWithOpts "]*>" False True)) "" -{- | Extract a file from a zipfile. --} -extractFileFromZip :: (MonadIO m, MonadError String m) - => FilePath -- ^ path to zip file - -> FilePath -- ^ path within zip file to extract - -> m String -- ^ contents of expected file -extractFileFromZip zipPath filePath = do - result <- liftIO $ catchZipError - (fmap Right $ withArchive [] zipPath $ fileContents [] filePath) - (return . Left) - - output <- either (throwError . show) return result - - return . removeEncoding . removeDoctype $ output +-- | Extract a file from a zip archive throwing an error on failure +fileFromArchive :: MonadError String m => + FilePath -> Archive -> m String +fileFromArchive filePath archive = do + let mbEntry = findEntryByPath filePath archive + maybe + (throwError $ "Unable to locate file " ++ filePath) + (return . B.unpack . fromEntry) mbEntry --- | Get the path within an ePub file to the OPF Package Document -opfPath :: (MonadError String m, MonadIO m) +-- | Get the contents of the OPF Package Document from an ePub file +opfContents :: (MonadError String m, MonadIO m) => FilePath -- ^ path to ePub zip file - -> m String -- ^ path within ePub to the OPF Package Document -opfPath zipPath = do - containerContents <- extractFileFromZip zipPath - "META-INF/container.xml" + -> m String -- ^ contents of the OPF Package Document +opfContents zipPath = do + {- We need to first extract the container.xml file + It's required to have a certain path and name in the epub + and contains the path to what we really want, the .opf file. + -} + zipFileBytes <- liftIO $ B.readFile zipPath + let archive = toArchive zipFileBytes + let containerPath = "META-INF/container.xml" + containerDoc <- fileFromArchive containerPath archive + result <- liftIO $ runX ( - readString [withValidate no] containerContents + readString [withValidate no] containerDoc >>> deep (isElem >>> hasName "rootfile") >>> getAttrValue "full-path" ) - case result of + rootPath <- case result of (p : []) -> return p - _ -> throwError - "ERROR: rootfile full-path missing from META-INF/container.xml" + _ -> throwError $ + "ERROR: rootfile full-path missing from " ++ containerPath + + -- Now that we have the path to the .opf file, extract it + rootDoc <- fileFromArchive rootPath archive + + return . removeEncoding . removeDoctype $ rootDoc diff --git a/src/Codec/Epub/Opf/Cli/Opts.hs b/src/Codec/Epub/Opf/Cli/Opts.hs --- a/src/Codec/Epub/Opf/Cli/Opts.hs +++ b/src/Codec/Epub/Opf/Cli/Opts.hs @@ -52,5 +52,5 @@ , "Options:" ] footer = init $ unlines - [ "Version 2.0.2 Dino Morelli " + [ "Version 2.1.0 Dino Morelli " ] diff --git a/src/Codec/Epub/Opf/Parse.hs b/src/Codec/Epub/Opf/Parse.hs --- a/src/Codec/Epub/Opf/Parse.hs +++ b/src/Codec/Epub/Opf/Parse.hs @@ -251,9 +251,9 @@ XML string -} parseXmlToOpf :: (MonadIO m) => String -> m [Package] -parseXmlToOpf opfContents = +parseXmlToOpf contents = liftIO $ runX ( - readString [withValidate no] opfContents + readString [withValidate no] contents >>> propagateNamespaces >>> getBookData ) @@ -263,8 +263,7 @@ parseEpubOpf :: (MonadIO m, MonadError String m) => FilePath -> m Package parseEpubOpf zipPath = do - opfContents <- extractFileFromZip zipPath =<< opfPath zipPath - result <- parseXmlToOpf opfContents + result <- opfContents zipPath >>= parseXmlToOpf case result of (em : []) -> return em