X-Git-Url: https://gitweb.michael.orlitzky.com/?a=blobdiff_plain;f=src%2FMain.hs;h=bd5ed5c1cb07e3c17aadd9c0df946bc304040452;hb=0fb33e421979003bec08430d550314bf95504eee;hp=648ddae40a22a708d5b0edf8971bb031faae03a7;hpb=a0853ebe00e630b7d50f97949d4dd45c602bd454;p=dead%2Fhtsn-import.git diff --git a/src/Main.hs b/src/Main.hs index 648ddae..bd5ed5c 100644 --- a/src/Main.hs +++ b/src/Main.hs @@ -45,18 +45,47 @@ import Network.Services.TSN.Report ( report_info, report_error ) import TSN.DbImport ( DbImport(..), ImportResult(..) ) +import TSN.Parse ( format_parse_error ) +import qualified TSN.XML.AutoRacingDriverList as AutoRacingDriverList ( + dtd, + pickle_message ) +import qualified TSN.XML.AutoRacingResults as AutoRacingResults ( + dtd, + pickle_message ) import qualified TSN.XML.AutoRacingSchedule as AutoRacingSchedule ( dtd, pickle_message ) +import qualified TSN.XML.EarlyLine as EarlyLine ( + dtd, + pickle_message ) +import qualified TSN.XML.GameInfo as GameInfo ( dtds, parse_xml ) import qualified TSN.XML.Heartbeat as Heartbeat ( dtd, verify ) import qualified TSN.XML.Injuries as Injuries ( dtd, pickle_message ) import qualified TSN.XML.InjuriesDetail as InjuriesDetail ( dtd, pickle_message ) -import qualified TSN.XML.News as News ( dtd, pickle_message ) +import qualified TSN.XML.MLBBoxScore as MLBBoxScore ( + dtd, + pickle_message ) +import qualified TSN.XML.MLBEarlyLine as MLBEarlyLine ( + dtd, + pickle_message ) +import qualified TSN.XML.JFile as JFile ( dtd, pickle_message ) +import qualified TSN.XML.News as News ( + dtd, + has_only_single_sms, + pickle_message ) import qualified TSN.XML.Odds as Odds ( dtd, pickle_message ) +import qualified TSN.XML.ScheduleChanges as ScheduleChanges ( + dtd, + pickle_message ) import qualified TSN.XML.Scores as Scores ( dtd, pickle_message ) -import qualified TSN.XML.Weather as Weather ( dtd, pickle_message ) +import qualified TSN.XML.SportInfo as SportInfo ( dtds, parse_xml ) +import qualified TSN.XML.Weather as Weather ( + dtd, + is_type1, + pickle_message, + teams_are_normal ) import Xml ( DtdName(..), parse_opts ) @@ -87,7 +116,7 @@ import_file cfg path = do report_error $ "Unable to determine DTD for file " ++ path ++ "." return False (ImportFailed errmsg:_) -> do - report_error errmsg + report_error $ errmsg ++ " (" ++ path ++ ")" return False (ImportSkipped infomsg:_) -> do -- We processed the message but didn't import anything. Return @@ -158,43 +187,96 @@ import_file cfg path = do -- migrate_and_import m = dbmigrate m >> dbimport m + -- | The error message we return if unpickling fails. + -- + errmsg = "Could not unpickle " ++ dtd ++ "." + + -- | Try to migrate and import using the given pickler @f@; + -- if it works, return the result. Otherwise, return an + -- 'ImportFailed' along with our error message. + -- + go f = maybe + (return $ ImportFailed errmsg) + migrate_and_import + (unpickleDoc f xml) + importer - | dtd == AutoRacingSchedule.dtd = do - let m = unpickleDoc AutoRacingSchedule.pickle_message xml - maybe (return $ ImportFailed errmsg) migrate_and_import m + | dtd == AutoRacingDriverList.dtd = + go AutoRacingDriverList.pickle_message + + | dtd == AutoRacingResults.dtd = + go AutoRacingResults.pickle_message + + | dtd == AutoRacingSchedule.dtd = + go AutoRacingSchedule.pickle_message + + | dtd == EarlyLine.dtd = + go EarlyLine.pickle_message + + -- GameInfo and SportInfo appear last in the guards + | dtd == Injuries.dtd = go Injuries.pickle_message + + | dtd == InjuriesDetail.dtd = go InjuriesDetail.pickle_message + + | dtd == JFile.dtd = go JFile.pickle_message + + | dtd == MLBBoxScore.dtd = go MLBBoxScore.pickle_message - | dtd == Injuries.dtd = do - let m = unpickleDoc Injuries.pickle_message xml - maybe (return $ ImportFailed errmsg) migrate_and_import m + | dtd == MLBEarlyLine.dtd = go MLBEarlyLine.pickle_message - | dtd == InjuriesDetail.dtd = do - let m = unpickleDoc InjuriesDetail.pickle_message xml - maybe (return $ ImportFailed errmsg) migrate_and_import m + | dtd == News.dtd = + -- Some of the newsxml docs are busted in predictable ways. + -- We want them to "succeed" so that they're deleted. + -- We already know we can't parse them. + if News.has_only_single_sms xml + then go News.pickle_message + else do + let msg = "Unsupported newsxml.dtd with multiple SMS " ++ + "(" ++ path ++ ")" + return $ ImportUnsupported msg + | dtd == Odds.dtd = go Odds.pickle_message + | dtd == ScheduleChanges.dtd = go ScheduleChanges.pickle_message - | dtd == News.dtd = do - let m = unpickleDoc News.pickle_message xml - maybe (return $ ImportFailed errmsg) migrate_and_import m + | dtd == Scores.dtd = go Scores.pickle_message - | dtd == Odds.dtd = do - let m = unpickleDoc Odds.pickle_message xml - maybe (return $ ImportFailed errmsg) migrate_and_import m + -- SportInfo and GameInfo appear last in the guards + | dtd == Weather.dtd = + -- Some of the weatherxml docs are busted in predictable ways. + -- We want them to "succeed" so that they're deleted. + -- We already know we can't parse them. + if Weather.is_type1 xml + then if Weather.teams_are_normal xml + then go Weather.pickle_message + else do + let msg = "Teams in reverse order in weatherxml.dtd" ++ + " (" ++ path ++ ")" + return $ ImportUnsupported msg + else do + let msg = "Unsupported weatherxml.dtd type (" ++ path ++ ")" + return $ ImportUnsupported msg - | dtd == Scores.dtd = do - let m = unpickleDoc Scores.pickle_message xml - maybe (return $ ImportFailed errmsg) migrate_and_import m + | dtd `elem` GameInfo.dtds = do + let either_m = GameInfo.parse_xml dtd xml + case either_m of + -- This might give us a slightly better error + -- message than the default 'errmsg'. + Left err -> return $ ImportFailed (format_parse_error err) + Right m -> migrate_and_import m - | dtd == Weather.dtd = do - let m = unpickleDoc Weather.pickle_message xml - maybe (return $ ImportFailed errmsg) migrate_and_import m + | dtd `elem` SportInfo.dtds = do + let either_m = SportInfo.parse_xml dtd xml + case either_m of + -- This might give us a slightly better error + -- message than the default 'errmsg'. + Left err -> return $ ImportFailed (format_parse_error err) + Right m -> migrate_and_import m | otherwise = do let infomsg = "Unrecognized DTD in " ++ path ++ ": " ++ dtd ++ "." return $ ImportUnsupported infomsg - where - errmsg = "Could not unpickle " ++ dtd ++ "." -- | Entry point of the program. It twiddles some knobs for