X-Git-Url: https://gitweb.michael.orlitzky.com/?a=blobdiff_plain;f=src%2FTSN%2FXML%2FGameInfo.hs;h=3062f5a2b79a1d826a649ccc4c5cba7c9b98fe08;hb=32147474ba5c91452eeb532381f63e88c257a982;hp=b13edd78af9c28c8e917fe3e2719a02291b5e6c4;hpb=7ab10646746d18c3d5c00d4c68e78807197aa560;p=dead%2Fhtsn-import.git diff --git a/src/TSN/XML/GameInfo.hs b/src/TSN/XML/GameInfo.hs index b13edd7..3062f5a 100644 --- a/src/TSN/XML/GameInfo.hs +++ b/src/TSN/XML/GameInfo.hs @@ -25,12 +25,12 @@ import Data.String.Utils ( replace ) import Data.Time.Clock ( UTCTime ) import Database.Groundhog ( countAll, - migrate, - runMigration, - silentMigrationLogger ) -import Database.Groundhog.Generic ( runDbConn ) + insert_, + migrate ) +import Database.Groundhog.Generic ( runDbConn, runMigrationSilent ) import Database.Groundhog.Sqlite ( withSqliteConn ) import Database.Groundhog.TH ( + defaultCodegenConfig, groundhog, mkPersist ) import Test.Tasty ( TestTree, testGroup ) @@ -39,20 +39,18 @@ import Text.XML.HXT.Core ( XmlTree ) import Text.XML.HXT.DOM.ShowXml ( xshow ) -- Local imports. -import TSN.Codegen ( tsn_codegen_config ) import TSN.DbImport ( DbImport(..), ImportResult(..), run_dbmigrate ) import TSN.Parse ( + ParseError, + parse_game_id, parse_message, + parse_schedule_id, parse_xmlfid, parse_xml_time_stamp ) -import TSN.XmlImport ( XmlImport(..) ) -import Xml ( - FromXml(..), - ToDb(..), - unsafe_read_document ) +import Xml ( unsafe_read_document ) -- | The DTDs for everything that we consider \"Game Info.\" @@ -84,85 +82,78 @@ dtds = "WorldBaseballPreviewXML.dtd" ] --- | XML representation of a GameInfo \. +-- | This serves as both the database and XML representation of a +-- GameInfo \. -- -data Message = - Message { - xml_dtd :: String, - xml_xml_file_id :: Int, - xml_time_stamp :: UTCTime, - xml_xml :: String } +-- The 'game_id' and 'schedule_id' fields are foreign keys, but they +-- key into multiple tables and key on records which may not exist +-- when we import the GameInfo document. We therefore don't declare +-- them as foreign keys; i.e. we don't require them to point +-- anywhere in particular. But if they do, that's nice. +-- +data GameInfo = + GameInfo { + dtd :: String, + xml_file_id :: Int, + game_id :: Maybe Int, -- ^ These are optional because they are missing + -- from at least the MLB_Matchup_XML.dtd documents. + -- They provide foreign keys into any tables storing + -- games with their IDs. + + schedule_id :: Maybe Int, -- ^ Optional key into any table storing a + -- schedule along with its ID. We've noticed + -- them missing in e.g. recapxml.dtd documents. + time_stamp :: UTCTime, + xml :: String } deriving (Eq, Show) --- | Attempt to parse a 'Message' from an 'XmlTree'. If we cannot, +-- | Attempt to parse a 'GameInfo' from an 'XmlTree'. If we cannot, -- we fail with an error message. -- -parse_xml :: String -> XmlTree -> Either String Message +parse_xml :: String -> XmlTree -> Either ParseError GameInfo parse_xml dtdname xmltree = do xmlfid <- parse_xmlfid xmltree + game_id <- parse_game_id xmltree + schedule_id <- parse_schedule_id xmltree timestamp <- parse_xml_time_stamp xmltree message <- parse_message xmltree - return $ Message dtdname (fromInteger xmlfid) timestamp (xshow [message]) - - --- | Database representation of a 'Message'. --- -data GameInfo = - GameInfo { - db_dtd :: String, - db_xml_file_id :: Int, - db_time_stamp :: UTCTime, - db_xml :: String } - + return $ GameInfo + dtdname + xmlfid + game_id + schedule_id + timestamp + (xshow [message]) -instance ToDb Message where - -- | The database analogue of a 'Message' is an 'GameInfo'. - type Db Message = GameInfo - -instance FromXml Message where - -- | The XML to DB conversion is trivial here. - -- - from_xml Message{..} = GameInfo { - db_dtd = xml_dtd, - db_xml_file_id = xml_xml_file_id, - db_time_stamp = xml_time_stamp, - db_xml = xml_xml } - - --- | This allows us to insert the XML representation 'Message' --- directly. -- -instance XmlImport Message - - --- --- Database code +-- * Database code -- -instance DbImport Message where +instance DbImport GameInfo where dbmigrate _ = run_dbmigrate $ migrate (undefined :: GameInfo) - -- | We import a 'Message' by inserting the whole thing at + -- | We import a 'GameInfo' by inserting the whole thing at -- once. Nothing fancy going on here. dbimport msg = do - insert_xml_ msg + insert_ msg return ImportSucceeded -- | The database schema for GameInfo is trivial; all we need is for -- the XML_File_ID to be unique. -- -mkPersist tsn_codegen_config [groundhog| +mkPersist defaultCodegenConfig [groundhog| - entity: GameInfo + dbName: game_info constructors: - name: GameInfo uniques: - name: unique_game_info type: constraint # Prevent multiple imports of the same message. - fields: [db_xml_file_id] + fields: [xml_file_id] |] @@ -176,10 +167,36 @@ game_info_tests :: TestTree game_info_tests = testGroup "GameInfo tests" - [ test_parse_xml_succeeds, + [ test_accessors, + test_parse_xml_succeeds, test_dbimport_succeeds ] +-- | Make sure the accessors work and that we can parse one file. Ok, +-- so the real point of this is to make the unused fields (dtd, xml, +-- ...) warning go away without having to mangle the groundhog code. +-- +test_accessors :: TestTree +test_accessors = testCase "we can access a parsed game_info" $ do + xmltree <- unsafe_read_document "test/xml/gameinfo/recapxml.xml" + let Right t = parse_xml "recapxml.dtd" xmltree + let a1 = dtd t + let ex1 = "recapxml.dtd" + let a2 = xml_file_id t + let ex2 = 21201550 + let a3 = show $ time_stamp t + let ex3 = "2014-05-31 15:13:00 UTC" + let a4 = game_id t + let ex4 = Just 39978 + let a5 = schedule_id t + let ex5 = Just 39978 + let a6 = take 9 (xml t) + let ex6 = "" + let actual = (a1,a2,a3,a4,a5,a6) + let expected = (ex1,ex2,ex3,ex4,ex5,ex6) + actual @?= expected + + -- | Sample XML documents for GameInfo types. -- game_info_test_files :: [FilePath] @@ -215,7 +232,7 @@ test_dbimport_succeeds = testCase "dbimport succeeds" $ do xmltrees <- mapM unsafe_read_document game_info_test_files let msgs = rights $ map (parse_xml "dummy") xmltrees actual <- withSqliteConn ":memory:" $ runDbConn $ do - runMigration silentMigrationLogger $ do + runMigrationSilent $ migrate (undefined :: GameInfo) mapM_ dbimport msgs countAll (undefined :: GameInfo)