X-Git-Url: http://gitweb.michael.orlitzky.com/?a=blobdiff_plain;f=src%2FLWN%2FPage.hs;h=cdcd1a602e1add49f729f5a5f7263eebae2e69a5;hb=b054d19e638a4983f88243e09f1aa474ae405998;hp=a4a56c222a8ebd1533fff5e9773efbf2dfe22acf;hpb=006b05fc061c0ce2fa0ce8b8c7f9361b0d64bb43;p=dead%2Flwn-epub.git diff --git a/src/LWN/Page.hs b/src/LWN/Page.hs index a4a56c2..cdcd1a6 100644 --- a/src/LWN/Page.hs +++ b/src/LWN/Page.hs @@ -1,41 +1,74 @@ +{-# LANGUAGE DoAndIfThenElse #-} + module LWN.Page where -import Text.Pandoc +import Control.Concurrent.ParallelIO (parallel) +import qualified Data.Map as Map (lookup) import Data.Time (getCurrentTime) -import System.IO (Handle) import qualified Data.ByteString.Lazy as B (ByteString, hPut) import Data.String.Utils (split, strip) +import qualified Data.Map as Map (Map, empty, insert) import Data.Maybe (catMaybes, fromJust, isNothing) -import Data.Tree.NTree.TypeDefs (NTree) +import Prelude hiding (readFile) +import System.IO (Handle, hClose, hFlush) import Test.HUnit (Assertion, assertEqual) import Test.Framework (Test, testGroup) import Test.Framework.Providers.HUnit (testCase) +import Text.Pandoc ( + defaultParserState, + defaultWriterOptions, + readHtml, + writeEPUB, + writerEPUBMetadata, + writerUserDataDir) +import Text.Pandoc.Shared ( readDataFile ) import Text.XML.HXT.Core ( ArrowXml, IOSArrow, XmlTree, - XNode, + ($<), (>>>), (/>), (//>), + changeAttrValue, + getAttrValue, getChildren, getText, - hasAttrValue, hasName, - mkName, none, processAttrl, processTopDown, + this, runX, - setElemName, xshow, - when - ) + when) import Text.HandsomeSoup (css, parseHtml) +import Configuration (Cfg, full_stories) import LWN.Article -import XHTML +import LWN.HTTP ( + ImageMap, + download_image_urls, + get_article_contents) +import LWN.URI (URL) +import LWN.XHTML ( + XHTML, + full_story_urls, + image_srcs, + full_story_link, + full_story_paragraph, + is_image, + preprocess, + remove_byline, + remove_full_story_paragraphs, + remove_title, + to_xhtml, + to_xml, + xml_from_contents) + + + data Page = -- | An LWN page with one article on it. @@ -59,9 +92,7 @@ instance XHTML Page where "