module LWN.Page
where
-import qualified Data.Map as Map
+import qualified Data.Map as Map (lookup)
import Data.Time (getCurrentTime)
-import System.IO (Handle, hClose, hFlush)
import qualified Data.ByteString.Lazy as B (ByteString, hPut)
import Data.String.Utils (split, strip)
import Data.Maybe (catMaybes, fromJust, isNothing)
+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
+import Text.Pandoc (
+ defaultParserState,
+ defaultWriterOptions,
+ readHtml,
+ writeEPUB,
+ writerEPUBMetadata)
import Text.XML.HXT.Core (
ArrowXml,
IOSArrow,
(/>),
(//>),
changeAttrValue,
- getAttrValue,
getChildren,
getText,
- hasAttrValue,
hasName,
- isElem,
- mkName,
- none,
processAttrl,
processTopDown,
runX,
- setElemName,
xshow,
- when
- )
+ when)
import Text.HandsomeSoup (css, parseHtml)
+import Configuration (Cfg)
import LWN.Article
-import LWN.HTTP (save_image)
-import LWN.URI (URL, try_make_absolute_url)
-import Misc (contains)
-import XHTML
-
--- Map absolute image URLs to local system file paths where the image
--- referenced by the URL is stored.
-type ImageMap = Map.Map URL FilePath
-
-download_image_urls :: [URL] -> IO ImageMap
-download_image_urls image_urls = do
- files <- mapM save_image image_urls
- let pairs = zip image_urls files
- return $ foldl my_insert empty_map pairs
- where
- empty_map = Map.empty :: ImageMap
+import LWN.HTTP (
+ ImageMap,
+ download_image_urls,
+ get_article_contents)
+import LWN.URI (URL)
+import LWN.XHTML (
+ XHTML,
+ image_srcs,
+ is_image,
+ preprocess,
+ remove_byline,
+ remove_title,
+ to_xhtml,
+ xml_from_contents)
- my_insert :: ImageMap -> (URL, Maybe FilePath) -> ImageMap
- my_insert dict (_, Nothing) = dict
- my_insert dict (k, Just v) = Map.insert k v dict
--- Should be called *after* preprocessing.
-download_images :: IOSArrow XmlTree XmlTree -> IO ImageMap
-download_images xml = do
- image_urls <- runX $ xml >>> image_srcs
- download_image_urls image_urls
data Page =
" <title>" ++ (show $ LWN.Article.title a) ++ "</title>" ++
"</head>" ++
"<body>" ++
- "<div>" ++
(to_xhtml a) ++
- "</div>" ++
"</body>" ++
"</html>"
-is_link :: (ArrowXml a) => a XmlTree XmlTree
-is_link =
- isElem >>> hasName "a"
-
+page_from_url :: Cfg -> URL -> IO (Maybe Page)
+page_from_url cfg url = do
+ contents <- get_article_contents cfg url
+ case (xml_from_contents contents) of
+ Just html -> parse html
+ Nothing -> return Nothing
-remove_comment_links :: (ArrowXml a) => a XmlTree XmlTree
-remove_comment_links =
- processTopDown $ kill_comments `when` is_link
- where
- is_comment_link =
- hasAttrValue "href" (contains "#Comments")
- kill_comments =
- none `when` is_comment_link
-
-replace_links_with_spans :: (ArrowXml a) => a XmlTree XmlTree
-replace_links_with_spans =
- processTopDown $ (make_span >>> remove_attrs) `when` is_link
- where
- make_span = setElemName $ mkName "span"
- remove_attrs = processAttrl none
+-- Should be called *after* preprocessing.
+download_images :: IOSArrow XmlTree XmlTree -> IO ImageMap
+download_images xml = do
+ image_urls <- runX $ xml >>> image_srcs
+ download_image_urls image_urls
--- | Preprocessing common to both page types.
-preprocess :: (ArrowXml a) => a XmlTree XmlTree
-preprocess =
- make_image_srcs_absolute
- >>>
- remove_comment_links
- >>>
- replace_links_with_spans
replace_remote_img_srcs :: (ArrowXml a) => ImageMap -> a XmlTree XmlTree
return $
if (isNothing appr) then
fppr
- else
+ else
appr
--
ap_parse :: IOSArrow XmlTree XmlTree -> IO (Maybe Page)
ap_parse xml = do
- arts <- ap_parse_articles xml
+ arts <- ap_parse_articles xml
case arts of
[x] -> return $ Just $ ArticlePage x
_ -> return Nothing
-
+
ap_parse_body :: IOSArrow XmlTree XmlTree -> IO (Maybe String)
ap_parse_body xml = do
let element_filter = xml >>> css "div.ArticleText"
if (isNothing parsed_headline) || (isNothing parsed_body)
then return []
- else do
+ else do
let title' = Title $ fromJust parsed_headline
let byline' = Byline parsed_byline
let body' = BodyHtml $ fromJust parsed_body
-
+
return $ [Article title' byline' body']
-is_title :: (ArrowXml a) => a XmlTree XmlTree
-is_title =
- (hasName "h2")
- >>>
- (hasAttrValue "class" (== "SummaryHL"))
-
-
-is_byline :: (ArrowXml a) => a XmlTree XmlTree
-is_byline =
- (hasName "div")
- >>>
- (hasAttrValue "class" (== "FeatureByLine"))
-
-
-is_image :: (ArrowXml a) => a XmlTree XmlTree
-is_image = isElem >>> hasName "img"
-
-remove_title :: (ArrowXml a) => a XmlTree XmlTree
-remove_title =
- processTopDown ((none) `when` is_title)
-
-
-remove_byline :: (ArrowXml a) => a XmlTree XmlTree
-remove_byline =
- processTopDown ((none) `when` is_byline)
-
-
fp_parse_article_body :: IOSArrow XmlTree XmlTree -> IO (Maybe String)
fp_parse_article_body xml = do
let xml = parseHtml $ wrap_in_body_div html
fp_parse_article xml
-
+
-- | In the full page, all of the article titles and bodies are
-- wrapped in one big div.ArticleText.
parse_bodies :: IOSArrow XmlTree XmlTree -> IOSArrow XmlTree XmlTree
read_html = readHtml defaultParserState
-
---
--- Misc
---
-
-image_srcs :: (ArrowXml a) => a XmlTree URL
-image_srcs =
- css "img"
- >>>
- getAttrValue "src"
-
-make_image_srcs_absolute :: (ArrowXml a) => a XmlTree XmlTree
-make_image_srcs_absolute =
- processTopDown (make_srcs_absolute `when` is_image)
- where
- change_src :: (ArrowXml a) => a XmlTree XmlTree
- change_src =
- changeAttrValue try_make_absolute_url
-
- make_srcs_absolute :: (ArrowXml a) => a XmlTree XmlTree
- make_srcs_absolute =
- processAttrl $ change_src `when` hasName "src"
-
-
--
-- Tests
--