Data.Vector.Compact.IntVec

Signed integer version of dynamic word vectors.

See Data.Vector.Compact.WordVec for more details.

Note: for unsigned integers, you really should use WordVec instead, because that is significantly faster, and has much more specialized functions implemented.

This module should be imported qualified (to avoid name clashes with Prelude).

The dynamic Word vector type

newtype IntVec

A dynamic int vector is a (small) vector of small signed integers stored compactly

Constructors

IntVec WordVec

Instances

Eq IntVec

Defined in Data.Vector.Compact.IntVec

Methods

(==) :: IntVec -> IntVec -> Bool

(/=) :: IntVec -> IntVec -> Bool

Ord IntVec

Defined in Data.Vector.Compact.IntVec

Methods

compare :: IntVec -> IntVec -> Ordering

(<) :: IntVec -> IntVec -> Bool

(<=) :: IntVec -> IntVec -> Bool

(>) :: IntVec -> IntVec -> Bool

(>=) :: IntVec -> IntVec -> Bool

max :: IntVec -> IntVec -> IntVec

min :: IntVec -> IntVec -> IntVec

Show IntVec

Defined in Data.Vector.Compact.IntVec

Methods

showsPrec :: Int -> IntVec -> ShowS

show :: IntVec -> String

showList :: [IntVec] -> ShowS

data Shape

The "shape" of a dynamic word vector

Constructors

Shape

Fields

• shapeLen :: !Int

  length of the vector

• shapeBits :: !Int

  bits per element (quantized to multiples of 4)

Instances

Eq Shape

Defined in Data.Vector.Compact.WordVec

Methods

(==) :: Shape -> Shape -> Bool

(/=) :: Shape -> Shape -> Bool

Show Shape

Defined in Data.Vector.Compact.WordVec

Methods

showsPrec :: Int -> Shape -> ShowS

show :: Shape -> String

showList :: [Shape] -> ShowS

vecShape :: IntVec -> Shape

vecLen :: IntVec -> Int

The length of the vector

vecBits :: IntVec -> Int

The number of bits per element used to encode the vector

Show instance

showIntVec :: IntVec -> String

showsPrecIntVec :: Int -> IntVec -> ShowS

Empty vector, singleton

null :: IntVec -> Bool

empty :: IntVec

singleton :: Int -> IntVec

isSingleton :: IntVec -> Maybe Int

Conversion to/from lists

fromList :: [Int] -> IntVec

Note: fromList xs = fromList' (lenMinMax xs)

fromList' :: (Int, (Int, Int)) -> [Int] -> IntVec

usage: fromList' (len,(min,max)) xs where min and max are the minimum and maximum (or just a lower and upper bound) appearing in the list.

fromList'' :: Shape -> [Int] -> IntVec

Don't use this unless you really know what you are doing!

lenMinMax :: [Int] -> (Int, (Int, Int))

Computes the length, minimum and maximum of a list, traversing it only once (instead of 3 times).

toList :: IntVec -> [Int]

toRevList :: IntVec -> [Int]

toRevList == reverse . toList

Indexing

unsafeIndex :: Int -> IntVec -> Int

Indexing starts from 0. No bound checks are done.

safeIndex :: Int -> IntVec -> Maybe Int

Head, tail, etc

head :: IntVec -> Int

tail :: IntVec -> IntVec

cons :: Int -> IntVec -> IntVec

Prepends an element

uncons :: IntVec -> Maybe (Int, IntVec)

last :: IntVec -> Int

snoc :: IntVec -> Int -> IntVec

Appends an element

concat :: IntVec -> IntVec -> IntVec

Generic operations

fold :: (a -> Int -> a) -> a -> IntVec -> a

Left fold

naiveMap :: (Int -> Int) -> IntVec -> IntVec

boundedMap :: (Int, Int) -> (Int -> Int) -> IntVec -> IntVec

If you have (nearly sharp) lower and upper bounds for the result of your of function on your vector, mapping can be more efficient

naiveZipWith :: (Int -> Int -> Int) -> IntVec -> IntVec -> IntVec

boundedZipWith :: (Int, Int) -> (Int -> Int -> Int) -> IntVec -> IntVec -> IntVec

If you have (nearly sharp) lower and upper bounds for the result of your of function on your vector, zipping can be more efficient

listZipWith :: (Int -> Int -> a) -> IntVec -> IntVec -> [a]

Number of bits needed

bitsNeededForMinMax :: (Int, Int) -> Int

usage: bitsNeededForMinMax (min,max)

bitsNeededFor :: Int -> Int

Note: this automatically rounds up to multiples of 4

roundBits :: Int -> Int

We only allow multiples of 4.