X-Git-Url: http://gitweb.michael.orlitzky.com/?p=dead%2Flwn-epub.git;a=blobdiff_plain;f=src%2FMain.hs;h=981a70531d717c1c7bef8f3fa874d26e943140eb;hp=a3728d289004783a09045ab739ea8a30edbd22b7;hb=4220827f62d772d7edcbdcc1c2f13d6c2eb5f534;hpb=b591d5f5c3c47d253c24144beae1edf8648cd94b diff --git a/src/Main.hs b/src/Main.hs index a3728d2..981a705 100644 --- a/src/Main.hs +++ b/src/Main.hs @@ -1,18 +1,101 @@ +{-# LANGUAGE DoAndIfThenElse #-} 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), + hPutStrLn, + openBinaryFile, + stderr, + stdout) +import System.IO.UTF8 (readFile) +import Text.XML.HXT.Core ( + IOStateArrow, + XmlTree) + +import CommandLine (show_help) +import Configuration (Cfg(..), get_cfg, use_account) +import LWN.Article (real_article_path) +import LWN.HTTP (get_page, log_in, make_cookie_jar) +import LWN.Page (epublish, parse) +import LWN.XHTML (parse_lwn) + + + +-- | 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 :: Cfg -> IO (Maybe (IOStateArrow s b XmlTree)) +get_xml_from_article cfg = do + my_article <- real_article_path (article cfg) + is_file <- doesFileExist my_article + case is_file of + True -> do + contents <- readFile my_article + return $ Just $ parse_lwn contents + False -> do + -- Download the URL and try to parse it. + if use_account cfg then do + -- use_account would be false if these fromJusts would fail. + cj <- make_cookie_jar + li_result <- log_in cj + (fromJust $ username cfg) + (fromJust $ password cfg) + + case li_result of + Left err -> do + let msg = "Failed to log in. " ++ err + hPutStrLn stderr msg + Right response_body -> do + hPutStrLn stderr response_body + + html <- get_page (Just cj) my_article + + case html of + Left err -> do + let msg = "Failed to retrieve page. " ++ err + hPutStrLn stderr msg + return Nothing + Right h -> return $ Just $ parse_lwn h + else do + html <- get_page Nothing my_article + case html of + Left err -> do + let msg = "Failed to retrieve page. " ++ err + hPutStrLn stderr msg + return Nothing + Right h -> return $ Just $ parse_lwn h + + +-- | 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 -import Epublishable -import LWN.ArticlePage 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 cfg) + maybe_html <- get_xml_from_article cfg + + 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 ()