1 {-# LANGUAGE DoAndIfThenElse #-}
6 import qualified Data.Map as Map (lookup)
7 import Data.Time (getCurrentTime)
8 import qualified Data.ByteString.Lazy as B (ByteString, hPut)
9 import Data.String.Utils (split, strip)
10 import qualified Data.Map as Map (Map, empty, insert)
11 import Data.Maybe (catMaybes, fromJust, isNothing)
12 import Prelude hiding (readFile)
13 import System.IO (Handle, hClose, hFlush)
14 import Test.HUnit (Assertion, assertEqual)
15 import Test.Framework (Test, testGroup)
16 import Test.Framework.Providers.HUnit (testCase)
23 import Text.XML.HXT.Core (
42 import Text.HandsomeSoup (css, parseHtml)
44 import Configuration (Cfg)
69 -- | An LWN page with one article on it.
70 ArticlePage { article :: Article } |
72 -- | An LWN page with more than one article on it. These require
73 -- different parsing and display functions than the single-article
75 FullPage { headline :: String,
76 articles :: [Article] }
79 instance XHTML Page where
80 to_xhtml (ArticlePage a) =
81 "<?xml version=\"1.0\" encoding=\"utf-8\" ?>" ++
82 "<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.1//EN\"" ++
83 "\"http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd\">" ++
85 " <meta http-equiv=\"Content-Type\"" ++
86 " content=\"application/xhtml+xml; charset=utf-8\" />" ++
87 " <title>" ++ (show $ LWN.Article.title a) ++ "</title>" ++
94 to_xhtml (FullPage hl as) =
95 "<?xml version=\"1.0\" encoding=\"utf-8\" ?>" ++
96 "<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.1//EN\"" ++
97 "\"http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd\">" ++
99 " <meta http-equiv=\"Content-Type\"" ++
100 " content=\"application/xhtml+xml; charset=utf-8\" />" ++
101 " <title>" ++ hl ++ "</title>" ++
105 "<h1>" ++ hl ++ "</h1>" ++
106 (concatMap to_xhtml as) ++
113 page_from_url :: Cfg -> URL -> IO (Maybe Page)
114 page_from_url cfg url = do
115 contents <- get_article_contents cfg url
116 case (xml_from_contents contents) of
117 Just html -> parse cfg html
118 Nothing -> return Nothing
122 insert_full_stories :: (ArrowXml a) => StoryMap -> a XmlTree XmlTree
123 insert_full_stories story_map =
124 processTopDown (article_xml `when` full_story_paragraph)
126 lookup_func :: (ArrowXml a) => URL -> a XmlTree XmlTree
128 case Map.lookup href story_map of
129 -- Leave it alone if we don't have the full story.
133 article_xml :: (ArrowXml a) => a XmlTree XmlTree
137 (this /> full_story_link >>> getAttrValue "href")
139 replace_remote_img_srcs :: (ArrowXml a) => ImageMap -> a XmlTree XmlTree
140 replace_remote_img_srcs image_map =
141 processTopDown (make_srcs_local `when` is_image)
143 -- old_src -> new_src
144 change_src_func :: String -> String
145 change_src_func old_src =
146 case Map.lookup old_src image_map of
147 -- Leave it alone if we don't have the file locally
151 change_src :: (ArrowXml a) => a XmlTree XmlTree
153 changeAttrValue change_src_func
155 make_srcs_local :: (ArrowXml a) => a XmlTree XmlTree
157 processAttrl $ (change_src `when` (hasName "src"))
162 -- Should be called *after* preprocessing.
163 download_images :: IOSArrow XmlTree XmlTree -> IO ImageMap
164 download_images xml = do
165 image_urls <- runX $ xml >>> image_srcs
166 download_image_urls image_urls
170 type StoryMap = Map.Map URL Article
172 -- These come *before* preprocessing.
173 download_full_story_urls :: Cfg -> [URL] -> IO StoryMap
174 download_full_story_urls cfg story_urls = do
175 pages <- mapM (page_from_url cfg) story_urls
176 let pairs = zip story_urls pages
177 return $ foldl my_insert empty_map pairs
179 empty_map = Map.empty :: StoryMap
181 my_insert :: StoryMap -> (URL, Maybe Page) -> StoryMap
182 my_insert dict (k, Just (ArticlePage v)) = Map.insert k v dict
183 my_insert dict (_, _) = dict
186 download_full_stories :: Cfg -> IOSArrow XmlTree XmlTree -> IO StoryMap
187 download_full_stories cfg xml = do
188 story_urls <- runX $ xml >>> full_story_urls
189 download_full_story_urls cfg story_urls
192 parse :: Cfg -> IOSArrow XmlTree XmlTree -> IO (Maybe Page)
194 story_map <- download_full_stories cfg xml
195 let fs_xml = xml >>> insert_full_stories story_map
197 let clean_xml = fs_xml >>> preprocess
198 image_map <- download_images clean_xml
199 let local_xml = clean_xml >>> replace_remote_img_srcs image_map
201 appr <- ap_parse local_xml
202 fppr <- fp_parse local_xml
204 if (isNothing appr) then
211 parse_headline :: IOSArrow XmlTree XmlTree -> IO (Maybe String)
212 parse_headline xml = do
213 let element_filter = xml >>> css "div.PageHeadline h1"
214 let element_text_filter = element_filter /> getText
215 element_text <- runX element_text_filter
218 [x] -> Just $ strip x
220 _ -> error "Found more than one headline."
223 parse_byline :: IOSArrow XmlTree XmlTree -> IO (Maybe String)
224 parse_byline xml = do
225 let element_filter = xml >>> css "div.FeatureByLine"
226 let element_text_filter = element_filter /> getText
227 element_text <- runX element_text_filter
230 [x] -> Just $ strip x
232 _ -> error "Found more than one article byline."
238 ap_parse :: IOSArrow XmlTree XmlTree -> IO (Maybe Page)
240 arts <- ap_parse_articles xml
242 [x] -> return $ Just $ ArticlePage x
246 ap_parse_body :: IOSArrow XmlTree XmlTree -> IO (Maybe String)
247 ap_parse_body xml = do
248 let element_filter = xml >>> css "div.ArticleText"
249 let element_html_filter = xshow element_filter
250 element_html <- runX element_html_filter
251 return $ case element_html of
254 _ -> error "Found more than one article."
257 ap_parse_articles :: IOSArrow XmlTree XmlTree -> IO [Article]
258 ap_parse_articles xml = do
259 parsed_headline <- parse_headline xml
260 parsed_byline <- parse_byline xml
261 parsed_body <- ap_parse_body xml
263 putStrLn $ fromJust parsed_headline
265 if (isNothing parsed_headline) || (isNothing parsed_body)
268 let title' = Title $ fromJust parsed_headline
269 let byline' = Byline parsed_byline
270 let body' = BodyHtml $ fromJust parsed_body
272 return $ [Article title' byline' body']
280 fp_parse :: IOSArrow XmlTree XmlTree -> IO (Maybe Page)
282 hl <- parse_headline xml
283 parsed_articles <- fp_parse_articles xml
284 case parsed_articles of
286 x -> return $ Just $ FullPage (fromJust hl) x
290 fp_parse_article_title :: IOSArrow XmlTree XmlTree -> IO (Maybe String)
291 fp_parse_article_title xml = do
292 let element_filter = xml >>> css "h2.SummaryHL"
293 let element_text_filter = element_filter //> getText
294 element_text <- runX element_text_filter
295 return $ case element_text of
296 [x] -> Just $ strip x
298 _ -> error "Found more than one article title."
303 fp_parse_article_body :: IOSArrow XmlTree XmlTree -> IO (Maybe String)
304 fp_parse_article_body xml = do
305 -- First, delete the article title and byline.
306 let clean_xml' = xml >>> remove_title >>> remove_byline
307 -- The only child of the body element should be a div.lwn-article
308 -- since we wrapped the article's HTML in that.
309 let clean_xml = clean_xml' >>> css "body" >>> getChildren
310 clean_html <- runX . xshow $ clean_xml
311 return $ case clean_html of
314 _ -> error "Found more than one article body."
316 fp_parse_article :: IOSArrow XmlTree XmlTree -> IO (Maybe Article)
317 fp_parse_article xml = do
318 parsed_article_title <- fp_parse_article_title xml
319 parsed_article_byline <- parse_byline xml
320 parsed_article_body <- fp_parse_article_body xml
322 if (isNothing parsed_article_title) || (isNothing parsed_article_body)
326 let title' = Title $ fromJust parsed_article_title
327 let byline' = Byline parsed_article_byline
328 let body' = BodyHtml $ fromJust parsed_article_body
329 return $ Just $ Article title' byline' body'
331 parse_html_article :: String -> IO (Maybe Article)
332 parse_html_article html = do
333 let xml = parseHtml $ wrap_in_body_div html
337 -- | In the full page, all of the article titles and bodies are
338 -- wrapped in one big div.ArticleText.
339 parse_bodies :: IOSArrow XmlTree XmlTree -> IOSArrow XmlTree XmlTree
341 xml >>> css "div.ArticleText"
344 fp_parse_articles :: IOSArrow XmlTree XmlTree -> IO [Article]
345 fp_parse_articles xml = do
346 bodies <- runX . xshow $ parse_bodies xml
347 let article_separator = "<h2 class=\"SummaryHL\">"
348 let split_articles'' = split article_separator (concat bodies)
349 -- The first element will contain the crap before the first <h2...>.
350 let split_articles' = case split_articles'' of
351 (_:_) -> tail split_articles''
353 -- Put the separator back, it was lost during the split.
354 let split_articles = map (article_separator ++) split_articles'
355 --_ <- mapM print_article split_articles
356 real_articles <- mapM parse_html_article split_articles
357 let just_articles = catMaybes real_articles
361 -- | This makes it easy to select otherwise-random chunks of html
363 wrap_in_body_div :: String -> String
365 "<body><div class=\"lwn-article\">" ++ s ++ "</div></body>"
371 -- Epublishable stuff
374 title :: Page -> String
375 title (ArticlePage a) = getTitle $ LWN.Article.title a
376 title (FullPage hl _) = hl
379 metadata :: Page -> IO String
381 date <- getCurrentTime
383 "<dc:creator>http://lwn.net/</dc:creator>\n" ++
384 "<dc:date>" ++ (show date) ++ "</dc:date>\n" ++
385 "<dc:language>en-US</dc:language>\n" ++
386 "<dc:rights>Copyright Eklektix, Inc.</dc:rights>\n" ++
387 "<dc:title>" ++ (LWN.Page.title obj) ++ "</dc:title>\n"
390 epublish :: Page -> Handle -> IO ()
391 epublish obj handle = do
392 let xhtml = to_xhtml obj
394 epub <- xhtml_to_epub epmd xhtml
399 xhtml_to_epub :: String -> String -> IO B.ByteString
401 write_epub . read_html
403 my_writer_options = defaultWriterOptions { writerEPUBMetadata = epmd }
404 write_epub = writeEPUB Nothing [] my_writer_options
405 read_html = readHtml defaultParserState
412 test_preprocess_links :: Assertion
413 test_preprocess_links = do
414 actual_xml' <- runX $ input_xml >>> preprocess >>> css "body"
415 let actual_xml = actual_xml' !! 0
417 expected_xml' <- runX $ expected_xml'' >>> css "body"
418 let expected_xml = expected_xml' !! 0
421 "Links are replaced with spans"
425 input_html = "<body><a href=\"#\">Hello, world!</a></body>"
426 input_xml = parseHtml input_html
427 expected_html = "<body><span>Hello, world!</span></body>"
428 expected_xml'' = parseHtml expected_html
431 test_absolve_images :: Assertion
432 test_absolve_images = do
433 actual_xml' <- runX $ input_xml >>> preprocess >>> css "body"
434 let actual_xml = actual_xml' !! 0
436 expected_xml' <- runX $ expected_xml'' >>> css "body"
437 let expected_xml = expected_xml' !! 0
440 "Image srcs are made absolute"
446 "<img src=\"/images/2012/example.jpg\" />" ++
448 input_xml = parseHtml input_html
451 "<img src=\"https://lwn.net/images/2012/example.jpg\" />" ++
453 expected_xml'' = parseHtml expected_html
456 test_comments_removed :: Assertion
457 test_comments_removed = do
458 actual_xml' <- runX $ input_xml >>> preprocess >>> css "body"
459 let actual_xml = actual_xml' !! 0
461 expected_xml' <- runX $ expected_xml'' >>> css "body"
462 let expected_xml = expected_xml' !! 0
465 "Comment links are removed"
471 "<a href=\"/Articles/501490/#Comments\">Comments (6 posted)</a>" ++
473 input_xml = parseHtml input_html
475 expected_html = "<body><p></p></body>"
476 expected_xml'' = parseHtml expected_html
479 test_full_story_urls_parsed :: Assertion
480 test_full_story_urls_parsed = do
481 actual <- runX $ actual'
484 "Full Story URLs are parsed"
488 expected = ["/Articles/500738/", "/Articles/501837/"]
492 "<a href=\"/Articles/500738/\">Full Story</a> ",
493 "(<a href=\"/Articles/500738/#Comments\">comments: 49</a>)",
495 "<a href=\"/Articles/501837/\">Full Story</a> ",
496 "(<a href=\"/Articles/501837/#Comments\">comments: none</a>)",
499 xml = parseHtml full_story_html
500 actual' = xml >>> full_story_urls
504 testGroup "Page Tests" [
505 testCase "Links are replaced with spans" test_preprocess_links,
506 testCase "Image srcs are made absolute" test_absolve_images,
507 testCase "Comment links are removed" test_comments_removed,
508 testCase "Full Story URLs are parsed" test_full_story_urls_parsed ]