-- | Minimal XML functionality needed to parse each document's
--   XML_File_ID.
--
module TSN.Xml (
  parse_xmlfid,
  xml_prologue,
  xml_tests )
where

import Data.Maybe ( listToMaybe, mapMaybe )
import Test.Tasty ( TestTree, testGroup )
import Test.Tasty.HUnit ( (@?=), Assertion, testCase )
import Text.Read ( readMaybe )
import Text.XML.HXT.Core (
  (>>>),
  (/>),
  getChildren,
  getText,
  hasName,
  runLA,
  xreadDoc )

-- | A tiny parser written in HXT to extract the "XML_File_ID" element
--   from a document.
parse_xmlfid :: String -> Maybe Integer
parse_xmlfid =
  listToMaybe . mapMaybe readMaybe . parse
  where
    parse :: String -> [String]
    parse =
      runLA (xreadDoc
               >>> hasName "message"
               />  hasName "XML_File_ID"
               >>> getChildren
               >>> getText)


-- | The opening "tag" for the XML prologue.
xml_prologue :: String
xml_prologue = "<?xml "


-- * Tasty Tests
xml_tests :: TestTree
xml_tests =
  testGroup
    "XML tests"
    [ xml_file_id_tests ]


xml_file_id_tests :: TestTree
xml_file_id_tests =
  testCase "XML_File_ID is parsed correctly" $ do
    let xmlfids = ["19908216", "19908216", "19908245", "19908246", "19908247"]
    mapM_ check xmlfids
  where
    check :: String -> Assertion
    check xmlfid = do
      xml <- readFile ("test/xml/" ++ xmlfid ++ ".xml")
      let actual = parse_xmlfid xml
      let expected = readMaybe xmlfid
      actual @?= expected