dataframe-0.3.3.8: A fast, safe, and intuitive DataFrame library.

Safe Haskell	None
Language	Haskell2010

DataFrame.Functions

Synopsis

name :: Show a => Expr a -> Text
col :: Columnable a => Text -> Expr a
as :: Columnable a => Expr a -> Text -> NamedExpr
(.=) :: Columnable a => Text -> Expr a -> NamedExpr
ifThenElse :: Columnable a => Expr Bool -> Expr a -> Expr a -> Expr a
lit :: Columnable a => a -> Expr a
lift :: (Columnable a, Columnable b) => (a -> b) -> Expr a -> Expr b
lift2 :: (Columnable c, Columnable b, Columnable a) => (c -> b -> a) -> Expr c -> Expr b -> Expr a
toDouble :: (Columnable a, Real a) => Expr a -> Expr Double
div :: (Integral a, Columnable a) => Expr a -> Expr a -> Expr a
mod :: (Integral a, Columnable a) => Expr a -> Expr a -> Expr a
(.==) :: (Columnable a, Eq a) => Expr a -> Expr a -> Expr Bool
eq :: (Columnable a, Eq a) => Expr a -> Expr a -> Expr Bool
(.<) :: (Columnable a, Ord a) => Expr a -> Expr a -> Expr Bool
lt :: (Columnable a, Ord a) => Expr a -> Expr a -> Expr Bool
(.>) :: (Columnable a, Ord a) => Expr a -> Expr a -> Expr Bool
gt :: (Columnable a, Ord a) => Expr a -> Expr a -> Expr Bool
(.<=) :: (Columnable a, Ord a, Eq a) => Expr a -> Expr a -> Expr Bool
leq :: (Columnable a, Ord a, Eq a) => Expr a -> Expr a -> Expr Bool
(.>=) :: (Columnable a, Ord a, Eq a) => Expr a -> Expr a -> Expr Bool
geq :: (Columnable a, Ord a, Eq a) => Expr a -> Expr a -> Expr Bool
and :: Expr Bool -> Expr Bool -> Expr Bool
(.&&) :: Expr Bool -> Expr Bool -> Expr Bool
or :: Expr Bool -> Expr Bool -> Expr Bool
(.||) :: Expr Bool -> Expr Bool -> Expr Bool
not :: Expr Bool -> Expr Bool
count :: Columnable a => Expr a -> Expr Int
collect :: Columnable a => Expr a -> Expr [a]
mode :: (Columnable a, Eq a) => Expr a -> Expr a
minimum :: (Columnable a, Ord a) => Expr a -> Expr a
maximum :: (Columnable a, Ord a) => Expr a -> Expr a
sum :: (Columnable a, Num a, Unbox a) => Expr a -> Expr a
mean :: (Columnable a, Real a, Unbox a) => Expr a -> Expr Double
variance :: (Columnable a, Real a, Unbox a) => Expr a -> Expr Double
median :: (Columnable a, Real a, Unbox a) => Expr a -> Expr Double
percentile :: Int -> Expr Double -> Expr Double
stddev :: (Columnable a, Real a, Unbox a) => Expr a -> Expr Double
zScore :: Expr Double -> Expr Double
pow :: (Columnable a, Num a) => Int -> Expr a -> Expr a
relu :: (Columnable a, Num a) => Expr a -> Expr a
min :: (Columnable a, Ord a) => Expr a -> Expr a -> Expr a
max :: (Columnable a, Ord a) => Expr a -> Expr a -> Expr a
reduce :: (Columnable a, Columnable b) => Expr b -> a -> (a -> b -> a) -> Expr a
whenPresent :: (Columnable a, Columnable b) => (a -> b) -> Expr (Maybe a) -> Expr (Maybe b)
whenBothPresent :: (Columnable a, Columnable b, Columnable c) => (a -> b -> c) -> Expr (Maybe a) -> Expr (Maybe b) -> Expr (Maybe c)
recode :: (Columnable a, Columnable b) => [(a, b)] -> Expr a -> Expr (Maybe b)
recodeWithDefault :: (Columnable a, Columnable b) => b -> [(a, b)] -> Expr a -> Expr b
firstOrNothing :: Columnable a => Expr [a] -> Expr (Maybe a)
lastOrNothing :: Columnable a => Expr [a] -> Expr (Maybe a)
splitOn :: Text -> Expr Text -> Expr [Text]
match :: Text -> Expr Text -> Expr (Maybe Text)
matchAll :: Text -> Expr Text -> Expr [Text]
parseDate :: Text -> Expr Text -> Expr (Maybe Day)
daysBetween :: Expr Day -> Expr Day -> Expr Int
bind :: forall a b m. (Columnable a, Columnable (m a), Monad m, Columnable b, Columnable (m b)) => (a -> m b) -> Expr (m a) -> Expr (m b)
generateConditions :: TypedColumn Double -> [Expr Bool] -> [Expr Double] -> DataFrame -> [Expr Bool]
generatePrograms :: Bool -> [Expr Bool] -> [Expr Double] -> [Expr Double] -> [Expr Double] -> [Expr Double]
isLiteral :: Expr a -> Bool
deduplicate :: Columnable a => DataFrame -> [Expr a] -> [(Expr a, TypedColumn a)]
equivalent :: DataFrame -> Expr Double -> Expr Double -> Bool
synthesizeFeatureExpr :: Text -> BeamConfig -> DataFrame -> Either String (Expr Double)
f1FromBinary :: Vector Double -> Vector Double -> Maybe Double
f1FromCounts :: Int -> Int -> Int -> Maybe Double
fitClassifier :: Text -> Int -> Int -> DataFrame -> Either String (Expr Int)
percentiles :: DataFrame -> [Expr Double]
roundToSigDigits :: Int -> Double -> Double
roundTo2SigDigits :: Double -> Double
fitRegression :: Text -> Int -> Int -> DataFrame -> Either String (Expr Double)
data LossFunction
- = PearsonCorrelation
- | MutualInformation
- | MeanSquaredError
- | F1
getLossFunction :: LossFunction -> Vector Double -> Vector Double -> Maybe Double
data BeamConfig = BeamConfig {
- searchDepth :: Int
- beamLength :: Int
- lossFunction :: LossFunction
- includeConditionals :: Bool
}
defaultBeamConfig :: BeamConfig
beamSearch :: DataFrame -> BeamConfig -> TypedColumn Double -> [Expr Double] -> [Expr Bool] -> [Expr Double] -> Maybe (Expr Double)
pickTopN :: DataFrame -> TypedColumn Double -> BeamConfig -> [(Expr Double, TypedColumn a)] -> [Expr Double]
pickTopNBool :: DataFrame -> TypedColumn Double -> [(Expr Bool, TypedColumn Bool)] -> [Expr Bool]
satisfiesExamples :: DataFrame -> TypedColumn Double -> Expr Double -> Bool
isReservedId :: Text -> Bool
isVarId :: Text -> Bool
isHaskellIdentifier :: Text -> Bool
sanitize :: Text -> Text
typeFromString :: [String] -> Q Type
declareColumns :: DataFrame -> DecsQ

Documentation

name :: Show a => Expr a -> Text Source #

col :: Columnable a => Text -> Expr a Source #

as :: Columnable a => Expr a -> Text -> NamedExpr Source #

(.=) :: Columnable a => Text -> Expr a -> NamedExpr infixr 0 Source #

ifThenElse :: Columnable a => Expr Bool -> Expr a -> Expr a -> Expr a Source #

lit :: Columnable a => a -> Expr a Source #

lift :: (Columnable a, Columnable b) => (a -> b) -> Expr a -> Expr b Source #

lift2 :: (Columnable c, Columnable b, Columnable a) => (c -> b -> a) -> Expr c -> Expr b -> Expr a Source #

toDouble :: (Columnable a, Real a) => Expr a -> Expr Double Source #

div :: (Integral a, Columnable a) => Expr a -> Expr a -> Expr a Source #

mod :: (Integral a, Columnable a) => Expr a -> Expr a -> Expr a Source #

(.==) :: (Columnable a, Eq a) => Expr a -> Expr a -> Expr Bool Source #

eq :: (Columnable a, Eq a) => Expr a -> Expr a -> Expr Bool Source #

(.<) :: (Columnable a, Ord a) => Expr a -> Expr a -> Expr Bool Source #

lt :: (Columnable a, Ord a) => Expr a -> Expr a -> Expr Bool Source #

(.>) :: (Columnable a, Ord a) => Expr a -> Expr a -> Expr Bool Source #

gt :: (Columnable a, Ord a) => Expr a -> Expr a -> Expr Bool Source #

(.<=) :: (Columnable a, Ord a, Eq a) => Expr a -> Expr a -> Expr Bool Source #

leq :: (Columnable a, Ord a, Eq a) => Expr a -> Expr a -> Expr Bool Source #

(.>=) :: (Columnable a, Ord a, Eq a) => Expr a -> Expr a -> Expr Bool Source #

geq :: (Columnable a, Ord a, Eq a) => Expr a -> Expr a -> Expr Bool Source #

and :: Expr Bool -> Expr Bool -> Expr Bool Source #

(.&&) :: Expr Bool -> Expr Bool -> Expr Bool Source #

or :: Expr Bool -> Expr Bool -> Expr Bool Source #

(.||) :: Expr Bool -> Expr Bool -> Expr Bool Source #

not :: Expr Bool -> Expr Bool Source #

count :: Columnable a => Expr a -> Expr Int Source #

collect :: Columnable a => Expr a -> Expr [a] Source #

mode :: (Columnable a, Eq a) => Expr a -> Expr a Source #

minimum :: (Columnable a, Ord a) => Expr a -> Expr a Source #

maximum :: (Columnable a, Ord a) => Expr a -> Expr a Source #

sum :: (Columnable a, Num a, Unbox a) => Expr a -> Expr a Source #

mean :: (Columnable a, Real a, Unbox a) => Expr a -> Expr Double Source #

variance :: (Columnable a, Real a, Unbox a) => Expr a -> Expr Double Source #

median :: (Columnable a, Real a, Unbox a) => Expr a -> Expr Double Source #

percentile :: Int -> Expr Double -> Expr Double Source #

stddev :: (Columnable a, Real a, Unbox a) => Expr a -> Expr Double Source #

zScore :: Expr Double -> Expr Double Source #

pow :: (Columnable a, Num a) => Int -> Expr a -> Expr a Source #

relu :: (Columnable a, Num a) => Expr a -> Expr a Source #

min :: (Columnable a, Ord a) => Expr a -> Expr a -> Expr a Source #

max :: (Columnable a, Ord a) => Expr a -> Expr a -> Expr a Source #

reduce :: (Columnable a, Columnable b) => Expr b -> a -> (a -> b -> a) -> Expr a Source #

whenPresent :: (Columnable a, Columnable b) => (a -> b) -> Expr (Maybe a) -> Expr (Maybe b) Source #

whenBothPresent :: (Columnable a, Columnable b, Columnable c) => (a -> b -> c) -> Expr (Maybe a) -> Expr (Maybe b) -> Expr (Maybe c) Source #

recode :: (Columnable a, Columnable b) => [(a, b)] -> Expr a -> Expr (Maybe b) Source #

recodeWithDefault :: (Columnable a, Columnable b) => b -> [(a, b)] -> Expr a -> Expr b Source #

firstOrNothing :: Columnable a => Expr [a] -> Expr (Maybe a) Source #

lastOrNothing :: Columnable a => Expr [a] -> Expr (Maybe a) Source #

splitOn :: Text -> Expr Text -> Expr [Text] Source #

match :: Text -> Expr Text -> Expr (Maybe Text) Source #

matchAll :: Text -> Expr Text -> Expr [Text] Source #

parseDate :: Text -> Expr Text -> Expr (Maybe Day) Source #

daysBetween :: Expr Day -> Expr Day -> Expr Int Source #

bind :: forall a b m. (Columnable a, Columnable (m a), Monad m, Columnable b, Columnable (m b)) => (a -> m b) -> Expr (m a) -> Expr (m b) Source #

generateConditions :: TypedColumn Double -> [Expr Bool] -> [Expr Double] -> DataFrame -> [Expr Bool] Source #

generatePrograms :: Bool -> [Expr Bool] -> [Expr Double] -> [Expr Double] -> [Expr Double] -> [Expr Double] Source #

isLiteral :: Expr a -> Bool Source #

deduplicate :: Columnable a => DataFrame -> [Expr a] -> [(Expr a, TypedColumn a)] Source #

equivalent :: DataFrame -> Expr Double -> Expr Double -> Bool Source #

Checks if two programs generate the same outputs given all the same inputs.

synthesizeFeatureExpr Source #

Arguments

:: Text	Target expression
-> BeamConfig
-> DataFrame
-> Either String (Expr Double)

f1FromBinary :: Vector Double -> Vector Double -> Maybe Double Source #

f1FromCounts :: Int -> Int -> Int -> Maybe Double Source #

fitClassifier Source #

Arguments

:: Text	Target expression
-> Int	Depth of search (Roughly, how many terms in the final expression)
-> Int	Beam size - the number of candidate expressions to consider at a time.
-> DataFrame
-> Either String (Expr Int)

percentiles :: DataFrame -> [Expr Double] Source #

roundToSigDigits :: Int -> Double -> Double Source #

roundTo2SigDigits :: Double -> Double Source #

fitRegression Source #

Arguments

:: Text	Target expression
-> Int	Depth of search (Roughly, how many terms in the final expression)
-> Int	Beam size - the number of candidate expressions to consider at a time.
-> DataFrame
-> Either String (Expr Double)

data LossFunction Source #

Constructors

PearsonCorrelation
MutualInformation
MeanSquaredError
F1

getLossFunction :: LossFunction -> Vector Double -> Vector Double -> Maybe Double Source #

data BeamConfig Source #

Constructors

BeamConfig
Fields searchDepth :: Int beamLength :: Int lossFunction :: LossFunction includeConditionals :: Bool

defaultBeamConfig :: BeamConfig Source #

beamSearch Source #

Arguments

:: DataFrame
-> BeamConfig	Parameters of the beam search.
-> TypedColumn Double	Examples
-> [Expr Double]	Constants
-> [Expr Bool]	Conditions
-> [Expr Double]	Programs
-> Maybe (Expr Double)

pickTopN :: DataFrame -> TypedColumn Double -> BeamConfig -> [(Expr Double, TypedColumn a)] -> [Expr Double] Source #

pickTopNBool :: DataFrame -> TypedColumn Double -> [(Expr Bool, TypedColumn Bool)] -> [Expr Bool] Source #

satisfiesExamples :: DataFrame -> TypedColumn Double -> Expr Double -> Bool Source #

isReservedId :: Text -> Bool Source #

isVarId :: Text -> Bool Source #

isHaskellIdentifier :: Text -> Bool Source #

sanitize :: Text -> Text Source #

typeFromString :: [String] -> Q Type Source #

declareColumns :: DataFrame -> DecsQ Source #