X-Git-Url: http://gitweb.michael.orlitzky.com/?p=dead%2Flwn-epub.git;a=blobdiff_plain;f=src%2FMain.hs;h=47c69fa1cba9ade063d2b63bea22fe800b33a4e9;hp=a3728d289004783a09045ab739ea8a30edbd22b7;hb=ebedcdb6b1b8925dcfb5700d076f25743fac8645;hpb=b591d5f5c3c47d253c24144beae1edf8648cd94b diff --git a/src/Main.hs b/src/Main.hs index a3728d2..47c69fa 100644 --- a/src/Main.hs +++ b/src/Main.hs @@ -1,18 +1,103 @@ +{-# 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.Regex.Posix ((=~)) +import Text.XML.HXT.Core -import Epublishable -import LWN.ArticlePage +import CommandLine (show_help) +import Configuration (Cfg(..), get_cfg) +import LWN.Page +import LWN.URI (is_lwn_url, make_absolute_url) +import Misc (contains) + + +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 s = do + is_file <- doesFileExist s + return $ if is_file then s else check_cases + where + abs_current = + case make_absolute_url "current" of + Nothing -> s + Just ac -> ac + abs_s = + case make_absolute_url s of + Nothing -> s + Just as -> as + + check_cases :: String + check_cases + | is_lwn_url s = s + | s `contains` "current" = abs_current + | s =~ "^[0-9]+$" = abs_s + | otherwise = s -- Give up 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 - putStrLn "Done." + Cfg{..} <- get_cfg + 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 -> do + _ <- show_help + return () + + Nothing -> do + _ <- show_help + return ()