baaa5b2 · project-m36 · git.iko.soy

git.iko.soy/project-m36

add DataFrame datatype. :showsorteddataframe <relExpr> <attrName>

commitbaaa5b231f16820abccd6f76b5017ef0de111581 authorJohnny Liao <dycan0716@gmail.com> dateTue 19 Jun 2018, 08:28 (8 years ago) parenta81e186

3 files changed+69−7

compare with master

unifiedsplit

src/bin/TutorialD/Interpreter/Base.hs+3 src/bin/TutorialD/Interpreter/RODatabaseContextOperator.hs+23−1 src/lib/ProjectM36/Base.hs+43−6

src/bin/TutorialD/Interpreter/Base.hs+3

@@ -41,6 +41,8 @@ displayOpResult (DisplayParseErrorResult mPromptLength err) = do

4141 pointyString len = T.justifyRight (len + fromIntegral errorIndent) '_' "^"

4242 maybe (pure ()) (TIO.putStrLn . pointyString) mPromptLength

4343 TIO.putStr ("ERR:" <> errString)

44+displayOpResult (DisplayDataFrameResult dFrame) = do

45+ TIO.putStrLn (showRelation $ fromDataFrame dFrame)

4446

4547spaceConsumer :: Parser ()

4648spaceConsumer = Lex.space (void spaceChar) (Lex.skipLineComment "--") (Lex.skipBlockComment "{-" "-}")

@@ -127,6 +129,7 @@ data TutorialDOperatorResult = QuitResult |

127129 DisplayResult StringType |

128130 DisplayIOResult (IO ()) |

129131 DisplayRelationResult Relation |

132+ DisplayDataFrameResult DataFrame |

130133 DisplayErrorResult StringType |

131134 DisplayParseErrorResult (Maybe PromptLength) (ParseError Char Dec) | -- Int refers to length of prompt text

132135 QuietSuccessResult

src/bin/TutorialD/Interpreter/RODatabaseContextOperator.hs+23−1

@@ -1,6 +1,8 @@

11{-# LANGUAGE GADTs #-}

22module TutorialD.Interpreter.RODatabaseContextOperator where

33import ProjectM36.Base

4+import ProjectM36.Attribute (attributeForName)

5+import ProjectM36.Relation (attributes)

46import ProjectM36.Error

57import ProjectM36.InclusionDependency

68import qualified ProjectM36.Client as C

@@ -26,6 +28,7 @@ data RODatabaseContextOperator where

2628 ShowRelationVariables :: RODatabaseContextOperator

2729 ShowAtomFunctions :: RODatabaseContextOperator

2830 ShowDatabaseContextFunctions :: RODatabaseContextOperator

31+ ShowSortedDataFrame :: RelationalExpr -> AttributeName -> RODatabaseContextOperator

2932 Quit :: RODatabaseContextOperator

3033 deriving (Show)

3134

@@ -81,6 +84,7 @@ roDatabaseContextOperatorP = typeP

8184 <|> showTypesP

8285 <|> showAtomFunctionsP

8386 <|> showDatabaseContextFunctionsP

87+ <|> showSortedDataFrameP

8488 <|> quitP

8589

8690--logically, these read-only operations could happen purely, but not if a remote call is required

@@ -148,6 +152,18 @@ evalRODatabaseContextOp sessionId conn ShowDatabaseContextFunctions = do

148152 Left err -> pure $ DisplayErrorResult (T.pack (show err))

149153 Right rel -> evalRODatabaseContextOp sessionId conn (ShowRelation (ExistingRelation rel))

150154

155+evalRODatabaseContextOp sessionId conn (ShowSortedDataFrame expr attrName) = do

156+ res <- C.executeRelationalExpr sessionId conn expr

157+ case res of

158+ Left err -> pure $ DisplayErrorResult $ T.pack (show err)

159+ Right rel -> do

160+ let attrs = attributes rel

161+ case attributeForName attrName attrs of

162+ Left err -> pure $ DisplayErrorResult $ T.pack (show err)

163+ Right attr -> do

164+ pure $ DisplayDataFrameResult $ sortDataFrameBy attr . toDataFrame $ rel

165+

166+

151167evalRODatabaseContextOp _ _ Quit = pure QuitResult

152168

153169interpretRODatabaseContextOp :: C.SessionId -> C.Connection -> T.Text -> IO TutorialDOperatorResult

@@ -155,4 +171,10 @@ interpretRODatabaseContextOp sessionId conn tutdstring = case parse roDatabaseCo

155171 Left err -> pure $ DisplayErrorResult (T.pack (show err))

156172 Right parsed -> evalRODatabaseContextOp sessionId conn parsed

157173

158−

174+showSortedDataFrameP :: Parser RODatabaseContextOperator

175+showSortedDataFrameP = do

176+ reservedOp ":showsorteddataframe"

177+ ShowSortedDataFrame <$> relExprP <*> identifier

178+

179+

180+

src/lib/ProjectM36/Base.hs+43−6

@@ -37,7 +37,7 @@ data Atom = IntegerAtom Integer |

3737 BoolAtom Bool |

3838 RelationAtom Relation |

3939 ConstructedAtom DataConstructorName AtomType [Atom]

40− deriving (Eq, Show, Binary, Typeable, NFData, Generic)

40+ deriving (Eq, Ord, Show, Binary, Typeable, NFData, Generic)

4141

4242instance Hashable Atom where

4343 hashWithSalt salt (ConstructedAtom dConsName _ atoms) = salt `hashWithSalt` atoms

@@ -77,7 +77,9 @@ data AtomType = IntAtomType |

7777 TypeVariableType TypeVarName

7878 --wildcard used in Atom Functions and tuples for data constructors which don't provide all arguments to the type constructor

7979 deriving (Eq, NFData, Generic, Binary, Show)

80−

80+

81+instance Ord AtomType where

82+ compare = undefined

8183type TypeVarMap = M.Map TypeVarName AtomType

8284

8385instance Hashable TypeVarMap where

@@ -122,6 +124,16 @@ instance Eq RelationTupleSet where

122124

123125instance NFData RelationTupleSet where rnf = genericRnf

124126

127+sortDataFrameBy :: Attribute -> DataFrame -> DataFrame

128+sortDataFrameBy attr (DataFrame attrs tupSet) =

129+ case attr `elem` attrs of

130+ True -> DataFrame attrs $ sortTupleSetBy (compareRelationTupleBy attr) tupSet

131+ False -> error $ "No such attribute: " ++ show attr

132+

133+sortTupleSetBy :: (RelationTuple -> RelationTuple -> Ordering) -> RelationTupleSet -> RelationTupleSet

134+sortTupleSetBy compare' (RelationTupleSet tuples) =

135+ RelationTupleSet $ L.sortBy compare' tuples

136+

125137--the same hash must be generated for equal tuples so that the hashset equality works

126138instance Hashable RelationTuple where

127139 --sanity check the tuple for attribute and tuple counts

@@ -145,15 +157,32 @@ instance Binary RelationTuple

145157instance Eq RelationTuple where

146158 (==) tuple1@(RelationTuple attrs1 _) tuple2@(RelationTuple attrs2 _) = attributesEqual attrs1 attrs2 && atomsEqual

147159 where

148− atomForAttribute attr (RelationTuple attrs tupVec) = case V.findIndex (== attr) attrs of

149− Nothing -> Nothing

150− Just index -> tupVec V.!? index

151160 atomsEqual = V.all (== True) $ V.map (\attr -> atomForAttribute attr tuple1 == atomForAttribute attr tuple2) attrs1

152161

162+atomForAttribute attr (RelationTuple attrs tupVec) =

163+ case V.findIndex (== attr) attrs of

164+ Nothing -> Nothing

165+ Just index -> tupVec V.!? index

166+

153167instance NFData RelationTuple where rnf = genericRnf

154168

169+data DataFrameTuple = DataFrameTuple Attributes (V.Vector Atom) deriving (Eq, Show, Generic)

170+

171+compareRelationTupleBy :: Attribute -> RelationTuple -> RelationTuple -> Ordering

172+compareRelationTupleBy attr tuple1@(RelationTuple attrs1 tupVec1) tuple2@(RelationTuple attrs2 tupVec2) =

173+ case attr `elem` attrs1 && attr `elem` attrs2 of

174+ True -> compare (atomForAttribute attr tuple1) (atomForAttribute attr tuple2)

175+ False -> error "Comparing two RelationTuples by an non-existent common attribute."

176+

155177data Relation = Relation Attributes RelationTupleSet deriving (Show, Generic,Typeable)

156178

179+instance Ord Relation where

180+ compare (Relation attrs1 tupSet1) (Relation attrs2 tupSet2) = undefined

181+ (<) = undefined

182+ (<=) = undefined

183+ (>=) = undefined

184+ (>) = undefined

185+

157186instance Eq Relation where

158187 Relation attrs1 tupSet1 == Relation attrs2 tupSet2 = attributesEqual attrs1 attrs2 && tupSet1 == tupSet2

159188

@@ -167,7 +196,15 @@ instance Hashable Relation where

167196 sortedAttrs = map snd (sortedAttributesIndices attrs)

168197

169198instance Binary Relation

170−

199+

200+data DataFrame = DataFrame Attributes RelationTupleSet deriving (Show, Generic, Typeable)

201+

202+toDataFrame :: Relation -> DataFrame

203+toDataFrame (Relation attr tupSet) = DataFrame attr tupSet

204+

205+fromDataFrame :: DataFrame -> Relation

206+fromDataFrame (DataFrame attr tupSet) = Relation attr tupSet

207+

171208-- | Used to represent the number of tuples in a relation.

172209data RelationCardinality = Countable | Finite Int deriving (Eq, Show, Generic, Ord)

173210