Don't download full stories if we aren't going to use them.

[dead/lwn-epub.git] / src / LWN / Page.hs

diff --git a/src/LWN/Page.hs b/src/LWN/Page.hs
index f0ada7cd13cfff0707904272f93fa2b4b354f9e3..3dccf466fea98f49a8ee42dcf44a2cdebe66a42c 100644 (file)
--- a/src/LWN/Page.hs
+++ b/src/LWN/Page.hs
@@ -3,6 +3,7 @@
 module LWN.Page
 where
 
+import Control.Concurrent.ParallelIO (parallel)
 import qualified Data.Map as Map (lookup)
 import Data.Time (getCurrentTime)
 import qualified Data.ByteString.Lazy as B (ByteString, hPut)
@@ -41,7 +42,7 @@ import Text.XML.HXT.Core (
   when)
 import Text.HandsomeSoup (css, parseHtml)
 
-import Configuration (Cfg)
+import Configuration (Cfg, full_stories)
 import LWN.Article
 import LWN.HTTP (
   ImageMap,
@@ -133,7 +134,7 @@ insert_full_stories story_map =
     article_xml :: (ArrowXml a) => a XmlTree XmlTree
     article_xml =
       lookup_func
-      $<
+      $< -- From HXT's Control.Arrow.ArrowList
       (this /> full_story_link >>> getAttrValue "href")
                    
 replace_remote_img_srcs :: (ArrowXml a) => ImageMap -> a XmlTree XmlTree
@@ -172,7 +173,7 @@ type StoryMap = Map.Map URL Article
 -- These come *before* preprocessing.
 download_full_story_urls :: Cfg -> [URL] -> IO StoryMap
 download_full_story_urls cfg story_urls = do
-  pages <- mapM (page_from_url cfg) story_urls
+  pages <- parallel $ map (page_from_url cfg) story_urls
   let pairs = zip story_urls pages
   return $ foldl my_insert empty_map pairs
   where
@@ -191,8 +192,11 @@ download_full_stories cfg xml = do
 
 parse :: Cfg -> IOSArrow XmlTree XmlTree -> IO (Maybe Page)
 parse cfg xml = do
-  story_map <- download_full_stories cfg xml
-  let fs_xml = xml >>> insert_full_stories story_map
+  fs_xml <- if (full_stories cfg) then do
+             story_map <- download_full_stories cfg xml
+             return $ xml >>> insert_full_stories story_map
+           else do
+             return xml
 
   let clean_xml = fs_xml >>> preprocess
   image_map <- download_images clean_xml
@@ -260,8 +264,6 @@ ap_parse_articles xml = do
   parsed_byline   <- parse_byline xml
   parsed_body     <- ap_parse_body xml
 
-  putStrLn $ fromJust parsed_headline
-
   if (isNothing parsed_headline) || (isNothing parsed_body)
   then return []
   else do
@@ -352,7 +354,6 @@ fp_parse_articles xml = do
                           []    -> []
   -- Put the separator back, it was lost during the split.
   let split_articles = map (article_separator ++) split_articles'
-  --_ <- mapM print_article split_articles
   real_articles <- mapM parse_html_article split_articles
   let just_articles = catMaybes real_articles
   return just_articles