Data.TDigest.Vector.Internal

Documentation

data TDigest (compression :: Nat)

TDigest is a vector of centroids plus not yet merged elements.

The size of structure is dictated by compression, *𝛿*. And is *O(𝛿)*.

Constructors

TDigest

Fields

• tdigestTotalWeight :: !Size

  sum of vector and buffer size

• tdigestData :: !(Vector Centroid)

  actual data. *Invariants:* sorted by mean; length <= 2 𝛿 (soft)

• tdigestBufferSize :: !Size
• tdigestBuffer :: [Double]

  addition buffer, elements with weight 1. *Invariants:* length 2 <= 𝛿

• tdigestDirection :: !Bool

  direction is a hack, so we merge from left and right. *TODO* remove?

Instances

KnownNat comp => Reducer Double (TDigest comp)

Both cons and snoc are insert

Defined in Data.TDigest.Vector.Internal

Methods

unit :: Double -> TDigest comp

snoc :: TDigest comp -> Double -> TDigest comp

cons :: Double -> TDigest comp -> TDigest comp

Show (TDigest compression)

Defined in Data.TDigest.Vector.Internal

Methods

showsPrec :: Int -> TDigest compression -> ShowS

show :: TDigest compression -> String

showList :: [TDigest compression] -> ShowS

KnownNat comp => Semigroup (TDigest comp)

Defined in Data.TDigest.Vector.Internal

Methods

(<>) :: TDigest comp -> TDigest comp -> TDigest comp

sconcat :: NonEmpty (TDigest comp) -> TDigest comp

stimes :: Integral b => b -> TDigest comp -> TDigest comp

KnownNat comp => Monoid (TDigest comp)

Defined in Data.TDigest.Vector.Internal

Methods

mempty :: TDigest comp

mappend :: TDigest comp -> TDigest comp -> TDigest comp

mconcat :: [TDigest comp] -> TDigest comp

NFData (TDigest comp)

Defined in Data.TDigest.Vector.Internal

Methods

rnf :: TDigest comp -> ()

KnownNat comp => HasHistogram (TDigest comp) Maybe

Defined in Data.TDigest.Vector.Internal

Methods

histogram :: TDigest comp -> Maybe (NonEmpty HistBin)

totalWeight :: TDigest comp -> Weight

size :: TDigest comp -> Int

Size of structure

totalWeight :: TDigest comp -> Weight

minimumValue :: KnownNat comp => TDigest comp -> Mean

Center of left-most centroid. Note: may be different than min element inserted.

>>> minimumValue (tdigest [1..100] :: TDigest 3)
1.0

maximumValue :: KnownNat comp => TDigest comp -> Mean

Center of right-most centroid. Note: may be different than max element inserted.

>>> maximumValue (tdigest [1..100] :: TDigest 3)
100.0

ksize

Arguments

:: Double

compression parameter, 𝛿

-> Double

quantile, q

-> Double

notional index, k

Mapping from quantile *q* to notional index *k* with compression parameter *𝛿*.

>>> ksize 42 0
0.0

>>> ksize 42 1
42.0

• q@ is clamped.:

>>> ksize 42 2
42.0

clamp :: Double -> Double

ksizeInv

Arguments

:: Double

compression parameter, 𝛿

-> Double

notional index, k

-> Double

quantile, q

Inverse of ksize.

>>> ksizeInv 42 0
0.0

>>> ksizeInv 42 42
1.0

>>> ksizeInv 42 (ksize 42 0.3)
0.3

merge :: Int -> Double -> [(Mean, Weight)] -> [(Mean, Weight)]

emptyTDigest :: TDigest comp

combineTDigest :: forall comp. KnownNat comp => TDigest comp -> TDigest comp -> TDigest comp

finalize :: forall comp. KnownNat comp => TDigest comp -> TDigest comp

Flush insertion buffer

forceCompress :: forall comp. KnownNat comp => TDigest comp -> TDigest comp

compress :: forall comp. KnownNat comp => TDigest comp -> TDigest comp

sizeCoefficient :: Num a => a

valid :: TDigest comp -> Bool

isRight . validate

validate :: TDigest comp -> Either String (TDigest comp)

Check various invariants in the TDigest structure.

insert

Arguments

:: KnownNat comp

=> Double

element

-> TDigest comp

-> TDigest comp

Insert single value into TDigest.

insert'

Arguments

:: KnownNat comp

=> Double

element

-> TDigest comp

-> TDigest comp

Insert single value, don't compress TDigest even if needed.

This may violate the insertion buffer size invariant.

For sensibly bounded input, it makes sense to let TDigest grow (it might grow linearly in size), and after that compress it once.

singleton :: Double -> TDigest comp

Make a TDigest of a single data point.

tdigest :: (Foldable f, KnownNat comp) => f Double -> TDigest comp

Strict foldl' over Foldable structure.

>>> :set -XDataKinds