Super ugly inline-js parsing

src/SEARX.hs

@@ -21,6 +21,8 @@ import SEARX.Client
 
 import Data.Foldable
 import qualified Data.Text as T
+import Data.Aeson
+import GHC.Generics
 
 {-
 debug purpose, use getMetadataWith
@@ -37,19 +39,39 @@ getMetadataWith2 q = do
             // Title
             article.close();
 -}
-parseWebsite :: T.Text -> IO Int
+
+data Article =
+  Article {title :: Maybe T.Text
+          ,byline :: Maybe T.Text
+          ,dir :: Maybe T.Text
+          ,content :: Maybe T.Text
+          ,textContent :: Maybe T.Text
+          ,length :: Maybe Int
+          ,excerpt :: Maybe T.Text
+          ,siteName :: Maybe T.Text
+          }
+  deriving (Generic, Show, FromJSON, ToJSON)
+
+parseWebsite :: T.Text -> IO Article
 parseWebsite url =
   withSystemTempDirectory "" $ \tmpdir -> do
     withCurrentDirectory tmpdir $
       traverse_
         callCommand
-        ["npm init --yes", "npm install --save node-readability"]
+        ["npm init --yes", "npm install --save jsdom request-promise-native",
+        "git clone https://github.com/mozilla/readability.git"]
     withJSSession
       defJSSessionOpts {nodeWorkDir = Just tmpdir}
       [block|
-          return 12;
-            });
-            |]
+        var {JSDOM} = require('jsdom');
+        var rp = require('request-promise-native');
+        var Readability = require('./readability');
+        const val = await rp($url);
+        const doc = new JSDOM(val, {url: $url})
+        const reader = new Readability(doc.window.document);
+        const article = reader.parse();
+        return article;
+      |]
 
 specConcatEith :: Semigroup a => Either b a -> Either b a -> Either b a
 specConcatEith (Left _) b = b