Begin working on determining the article argument type.

[dead/lwn-epub.git] / src / Main.hs

diff --git a/src/Main.hs b/src/Main.hs
index a3728d289004783a09045ab739ea8a30edbd22b7..655f1bef40c0680bda2c98e2d4ae9bc6f06c2cd8 100644 (file)
--- a/src/Main.hs
+++ b/src/Main.hs
@@ -1,18 +1,80 @@
+{-# LANGUAGE ScopedTypeVariables, RecordWildCards #-}
 module Main
 where
 
-import Data.Maybe (fromJust)
-import Text.HandsomeSoup (parseHtml)
-import System.Time (ClockTime( TOD ), getClockTime)
+import Prelude hiding (readFile)
+import System.Directory(doesFileExist)
+import System.IO (
+  Handle,
+  IOMode (WriteMode),
+  openBinaryFile,
+  stdout
+  )
+import System.IO.UTF8 (readFile)
+import Text.XML.HXT.Core
 
-import Epublishable
-import LWN.ArticlePage
+import CommandLine (Args(..), apply_args, show_help)
+import LWN.Page
+
+
+my_read_opts :: SysConfigList
+my_read_opts = [ withValidate  no,
+                 withParseHTML yes,
+                 withWarnings no ]
+
+-- | My version of HandsomeSoup's parseHTML.
+my_read :: String -> IOStateArrow s b XmlTree
+my_read = readString my_read_opts
+
+-- | Try to parse the given article using HXT. We try a few different
+--   methods; if none of them work, we return 'Nothing'.
+get_xml_from_article :: String -> IO (Maybe (IOStateArrow s b XmlTree))
+get_xml_from_article s = do
+  article <- real_article_path s
+  is_file <- doesFileExist article
+  case is_file of
+    True -> do
+      contents <- readFile article
+      return $ Just $ my_read contents
+    False -> do
+      -- Download the URL and try to parse it.
+      return Nothing
+
+-- | If we're given an empty path, return a handle to
+--   'stdout'. Otherwise, open the given file and return a read/write
+--   handle to that.
+get_output_handle :: FilePath -> IO Handle
+get_output_handle path =
+  if (null path) then
+    return stdout
+  else
+    openBinaryFile path WriteMode
+
+
+-- | Convert the given article to either a URL or a filesystem
+--   path. If the given article exists on the filesystem, we assume
+--   it's a file. Otherwise, we check to see if it's a URL. Failing
+--   that, we try to construct a URL from what we're given and do our
+--   best.
+real_article_path :: String -> IO String
+real_article_path = return . id
 
 main :: IO ()
 main = do
-  article_html <- readFile "test/fixtures/501317-article.html"
-  ioap <- parse $ parseHtml article_html
-  let article_page :: ArticlePage = fromJust $ ioap
-  (TOD t _) <- getClockTime 
-  epublish article_page "out.epub" t
+  Args{..} <- apply_args
+  output_handle <- get_output_handle output
+  maybe_html <- get_xml_from_article article
+
+  case maybe_html of
+
+    Just html -> do
+      result <- parse html
+      case result of
+        Just stuff  -> epublish stuff output_handle
+        Nothing     -> return ()
+
+    Nothing -> do
+      _ <- show_help
+      return ()
+
   putStrLn "Done."