X-Git-Url: http://gitweb.michael.orlitzky.com/?a=blobdiff_plain;f=src%2FTSN%2FXML%2FNews.hs;h=54169b982b00d278717d89e2a1042bde14cc6faf;hb=f0425854304197ab5ad47293b27b2e0b188cb844;hp=f401d41cb2dbb67d139395a6089c36f09ac9a23d;hpb=2f6ea95c2b3e19545c09ed5ff81d90bb3b120c16;p=dead%2Fhtsn-import.git diff --git a/src/TSN/XML/News.hs b/src/TSN/XML/News.hs index f401d41..54169b9 100644 --- a/src/TSN/XML/News.hs +++ b/src/TSN/XML/News.hs @@ -1,3 +1,4 @@ +{-# LANGUAGE DeriveGeneric #-} {-# LANGUAGE DeriveDataTypeable #-} {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE GADTs #-} @@ -17,9 +18,8 @@ module TSN.XML.News ( news_tests, -- * WARNING: these are private but exported to silence warnings News_LocationConstructor(..), - News_NewsTeamConstructor(..), - NewsConstructor(..), - NewsTeamConstructor(..) ) + News_TeamConstructor(..), + NewsConstructor(..) ) where -- System imports. @@ -28,20 +28,19 @@ import Data.Time.Clock ( UTCTime ) import Data.List.Utils ( join, split ) import Data.Tuple.Curry ( uncurryN ) import Data.Typeable ( Typeable ) +import qualified Data.Vector.HFixed as H ( HVector, convert ) import Database.Groundhog ( countAll, deleteAll, insert_, - migrate, - runMigration, - silentMigrationLogger ) + migrate ) import Database.Groundhog.Core ( DefaultKey ) -import Database.Groundhog.Generic ( runDbConn ) +import Database.Groundhog.Generic ( runDbConn, runMigrationSilent ) import Database.Groundhog.Sqlite ( withSqliteConn ) import Database.Groundhog.TH ( - defaultCodegenConfig, groundhog, mkPersist ) +import qualified GHC.Generics as GHC ( Generic ) import Test.Tasty ( TestTree, testGroup ) import Test.Tasty.HUnit ( (@?=), testCase ) import Text.XML.HXT.Core ( @@ -72,15 +71,16 @@ import TSN.Codegen ( tsn_db_field_namer ) -- Used in a test import TSN.Database ( insert_or_select ) import TSN.DbImport ( DbImport(..), ImportResult(..), run_dbmigrate ) -import TSN.Picklers ( xp_time_stamp ) import TSN.Location ( Location(..), pickle_location ) +import TSN.Picklers ( xp_attr_option, xp_time_stamp ) +import TSN.Team ( Team(..) ) import TSN.XmlImport ( XmlImport(..) ) import Xml ( FromXml(..), ToDb(..), pickle_unpickle, unpickleable, - unsafe_read_document, + unsafe_read_invalid_document, unsafe_unpickle ) @@ -101,11 +101,17 @@ dtd = "newsxml.dtd" -- prefix here so that the two names don't collide on \"id\" when -- Groundhog is creating its fields using our field namer. -- +-- The leading underscores prevent unused field warnings. +-- data MsgId = MsgId { - db_msg_id :: Int, - db_event_id :: Maybe Int } - deriving (Data, Eq, Show, Typeable) + _db_msg_id :: Int, + _db_event_id :: Maybe Int } + deriving (Data, Eq, GHC.Generic, Show, Typeable) + +-- | For 'H.convert'. +-- +instance H.HVector MsgId -- | The XML representation of a news item (\). @@ -118,14 +124,19 @@ data Message = xml_category :: String, xml_sport :: String, xml_url :: Maybe String, - xml_teams :: [NewsTeam], + xml_teams :: [NewsTeamXml], xml_locations :: [Location], xml_sms :: Maybe String, xml_editor :: Maybe String, xml_text :: Maybe String, -- Text and continue seem to show up in pairs, xml_continue :: Maybe String, -- either both present or both missing. xml_time_stamp :: UTCTime } - deriving (Eq, Show) + deriving (Eq, GHC.Generic, Show) + + +-- | For 'H.convert'. +-- +instance H.HVector Message -- | The database representation of a news item. We drop several @@ -158,6 +169,8 @@ instance FromXml Message where -- | We use a record wildcard so GHC doesn't complain that we never -- used the field names. -- + -- To convert, we drop some fields. + -- from_xml Message{..} = News { db_xml_file_id = xml_xml_file_id, db_mid = xml_mid, db_sport = xml_sport, @@ -173,32 +186,49 @@ instance FromXml Message where instance XmlImport Message --- * NewsTeam +-- * NewsTeamXml --- | The database/XML type for teams as they show up in the news. We --- can't reuse the representation from "TSN.Team" because they --- require a team id. We wouldn't want to make the team ID optional --- and then insert a team with no id, only to find the same team --- later with an id and be unable to update the record. (We could --- add the update logic, but it would be more trouble than it's --- worth.) +-- | The XML type for teams as they show up in the news. We can't +-- reuse the representation from "TSN.Team" because our name doesn't +-- appear optional and we have no abbreviation. -- -data NewsTeam = - NewsTeam { team_name :: String } - deriving (Eq, Show) +data NewsTeamXml = + NewsTeamXml { xml_team_id :: String, + xml_team_name :: String } + deriving (Eq, GHC.Generic, Show) +-- | For 'H.convert'. +-- +instance H.HVector NewsTeamXml --- * News_NewsTeam --- | Mapping between News records and NewsTeam records in the --- database. We don't name the fields because we don't use the names --- explicitly; that means we have to give them nice database names --- via groundhog. +instance ToDb NewsTeamXml where + -- | The database representation of 'NewsTeamXml' is 'Team'. + type Db NewsTeamXml = Team + +-- | Convert the XML representation 'NewsTeamXml' to the database +-- representation 'Team'. -- -data News_NewsTeam = News_NewsTeam - (DefaultKey News) - (DefaultKey NewsTeam) +instance FromXml NewsTeamXml where + from_xml NewsTeamXml{..} = + Team { team_id = xml_team_id, + abbreviation = Nothing, + name = Just xml_team_name } + +-- | Allow us to import 'NewsTeamXml' directly. +-- +instance XmlImport NewsTeamXml + + +-- * News_Team + +-- | Mapping between News records and Team records in the database. We +-- don't name the fields because we don't use the names explicitly; +-- that means we have to give them nice database names via +-- groundhog. +-- +data News_Team = News_Team (DefaultKey News) (DefaultKey Team) -- * News_Location @@ -248,7 +278,7 @@ has_only_single_sms xmltree = -- -- | Define 'dbmigrate' and 'dbimport' for 'Message's. The import is --- slightly non-generic because of our 'News_NewsTeam' and +-- slightly non-generic because of our 'News_Team' and -- 'News_Location' join tables. -- instance DbImport Message where @@ -256,23 +286,23 @@ instance DbImport Message where run_dbmigrate $ do migrate (undefined :: Location) migrate (undefined :: News) - migrate (undefined :: NewsTeam) - migrate (undefined :: News_NewsTeam) + migrate (undefined :: Team) + migrate (undefined :: News_Team) migrate (undefined :: News_Location) dbimport message = do -- Insert the message and acquire its primary key (unique ID) news_id <- insert_xml message - -- Now insert the teams. We use insert_or_select because we know - -- that most teams will already exist, and we want to get back the - -- id for the existing team when there's a collision. - nt_ids <- mapM insert_or_select (xml_teams message) + -- Now insert the teams. We use insert_xml_or_select because we + -- know that most teams will already exist, and we want to get + -- back the id for the existing team when there's a collision. + team_ids <- mapM insert_xml_or_select (xml_teams message) -- Now that the teams have been inserted, create - -- news__news_team records mapping beween the two. - let news_news_teams = map (News_NewsTeam news_id) nt_ids - mapM_ insert_ news_news_teams + -- news__team records mapping beween the two. + let news_teams = map (News_Team news_id) team_ids + mapM_ insert_ news_teams -- Do all of that over again for the Locations. loc_ids <- mapM insert_or_select (xml_locations message) @@ -282,21 +312,6 @@ instance DbImport Message where return ImportSucceeded --- These types don't have special XML representations or field name --- collisions so we use the defaultCodegenConfig and give their --- fields nice simple names. -mkPersist defaultCodegenConfig [groundhog| -- entity: NewsTeam - dbName: news_teams - constructors: - - name: NewsTeam - uniques: - - name: unique_news_teams - type: constraint - fields: [team_name] - -|] - -- These types have fields with e.g. db_ and xml_ prefixes, so we -- use our own codegen to peel those off before naming the columns. @@ -317,22 +332,22 @@ mkPersist tsn_codegen_config [groundhog| - embedded: MsgId fields: - - name: db_msg_id + - name: _db_msg_id dbName: msg_id - - name: db_event_id + - name: _db_event_id dbName: event_id -- entity: News_NewsTeam - dbName: news__news_teams +- entity: News_Team + dbName: news__teams constructors: - - name: News_NewsTeam + - name: News_Team fields: - - name: news_NewsTeam0 # Default created by mkNormalFieldName + - name: news_Team0 # Default created by mkNormalFieldName dbName: news_id reference: onDelete: cascade - - name: news_NewsTeam1 # Default created by mkNormalFieldName - dbName: news_teams_id + - name: news_Team1 # Default created by mkNormalFieldName + dbName: teams_id reference: onDelete: cascade @@ -356,18 +371,16 @@ mkPersist tsn_codegen_config [groundhog| -- XML Picklers -- --- | Convert a 'NewsTeam' to/from XML. +-- | Convert a 'NewsTeamXml' to/from XML. -- -pickle_news_team :: PU NewsTeam +pickle_news_team :: PU NewsTeamXml pickle_news_team = xpElem "team" $ - xpWrap (from_string, to_string) xpText + xpWrap (from_tuple, H.convert) $ + xpPair (xpAttr "id" xpText) + xpText -- team name where - to_string :: NewsTeam -> String - to_string = team_name - - from_string :: String -> NewsTeam - from_string = NewsTeam + from_tuple = uncurry NewsTeamXml -- | Convert a 'MsgId' to/from XML. @@ -375,11 +388,10 @@ pickle_news_team = pickle_msg_id :: PU MsgId pickle_msg_id = xpElem "msg_id" $ - xpWrap (from_tuple, to_tuple) $ - xpPair xpInt (xpAttr "EventId" (xpOption xpInt)) + xpWrap (from_tuple, H.convert) $ + xpPair xpInt (xpAttr "EventId" xp_attr_option) where from_tuple = uncurryN MsgId - to_tuple m = (db_msg_id m, db_event_id m) @@ -388,7 +400,7 @@ pickle_msg_id = pickle_message :: PU Message pickle_message = xpElem "message" $ - xpWrap (from_tuple, to_tuple) $ + xpWrap (from_tuple, H.convert) $ xp13Tuple (xpElem "XML_File_ID" xpInt) (xpElem "heading" xpText) pickle_msg_id @@ -404,19 +416,7 @@ pickle_message = (xpElem "time_stamp" xp_time_stamp) where from_tuple = uncurryN Message - to_tuple m = (xml_xml_file_id m, -- Verbose, - xml_heading m, -- but - xml_mid m, -- eliminates - xml_category m, -- GHC - xml_sport m, -- warnings - xml_url m, -- . - xml_teams m, -- . - xml_locations m, -- . - xml_sms m, - xml_editor m, - xml_text m, - xml_continue m, - xml_time_stamp m) + -- | We combine all of the \ elements into one 'String' -- while unpickling and do the reverse while pickling. @@ -539,11 +539,11 @@ test_on_delete_cascade = testGroup "cascading delete tests" news <- unsafe_unpickle path pickle_message let a = undefined :: Location let b = undefined :: News - let c = undefined :: NewsTeam - let d = undefined :: News_NewsTeam + let c = undefined :: Team + let d = undefined :: News_Team let e = undefined :: News_Location actual <- withSqliteConn ":memory:" $ runDbConn $ do - runMigration silentMigrationLogger $ do + runMigrationSilent $ do migrate a migrate b migrate c @@ -574,6 +574,6 @@ test_sms_detected_correctly = False ] where check path desc expected = testCase desc $ do - xmltree <- unsafe_read_document path + xmltree <- unsafe_read_invalid_document path let actual = has_only_single_sms xmltree actual @?= expected