monad-bayes 1.3.0.4 → 1.3.0.5

raw patch · 6 files changed

+62/−10 lines, 6 filesdep ~basedep ~brickdep ~vtyPVP: major bump suggested

API removals or changes: PVP suggests a major version bump

Dependency ranges changed: base, brick, vty

API changes (from Hackage documentation)

Files

• monad-bayes.cabal +5/−7
• src/Control/Monad/Bayes/Class.hs +36/−0
• src/Control/Monad/Bayes/Inference/SMC2.hs +4/−0
• src/Control/Monad/Bayes/Population.hs +5/−1
• src/Control/Monad/Bayes/Sampler/Strict.hs +10/−1
• src/Control/Monad/Bayes/Weighted.hs +2/−1

@@ -1,15 +1,13 @@ cabal-version:   2.2 name:            monad-bayes-version:         1.3.0.4+version:         1.3.0.5 license:         MIT license-file:    LICENSE.md copyright:       2015-2020 Adam Scibior maintainer:      dominic.steinitz@tweag.io author:          Adam Scibior <adscib@gmail.com> stability:       experimental-tested-with:-  GHC ==9.0.2 || ==9.2.7 || ==9.4.5 || ==9.6.4 || ==9.8.2 || ==9.10.1-+tested-with:     GHC ==9.0 || ==9.2 || ==9.4 || ==9.6 || ==9.8 || ==9.10 homepage:        http://github.com/tweag/monad-bayes#readme bug-reports:     https://github.com/tweag/monad-bayes/issues synopsis:        A library for probabilistic programming.@@ -40,8 +38,8 @@  common deps   build-depends:-    , base             >=4.15     && <4.21-    , brick            >=2.3.1    && <2.6+    , base             >=4.15     && <4.22+    , brick            ^>=2.10     , containers       >=0.6      && <0.8     , foldl            ^>=1.4     , free             ^>=5.2@@ -67,7 +65,7 @@     , text             >=1.2      && <2.2     , transformers     >=0.5.6    && <0.7     , vector           >=0.12.0   && <0.14-    , vty              >=6.1      && <6.3+    , vty              >=6.4      && <6.6     , vty-unix         ^>=0.2.0.0  common test-deps

@@ -71,6 +71,8 @@     Measure,     Kernel,     Log (ln, Exp),+    MonadUniformRange,+    uniformR,   ) where @@ -105,6 +107,7 @@ import Statistics.Distribution.Normal (normalDistr) import Statistics.Distribution.Poisson qualified as Poisson import Statistics.Distribution.Uniform (uniformDistr)+import System.Random (UniformRange)  -- | Monads that can draw random variables. class (Monad m) => MonadDistribution m where@@ -114,6 +117,8 @@     m Double    -- | Draw from a uniform distribution.+  --+  -- See also 'MonadUniformRange.uniformR' for types other than 'Double'.   uniform ::     -- | lower bound a     Double ->@@ -339,6 +344,19 @@ histogramToList :: Histogram -> [(Double, Double)] histogramToList = H.asList +-- | Monads that can efficiently draw uniform variables of+-- 'UniformRange' types.+class (Monad m) => MonadUniformRange m where+  -- | Draw from a uniform distribution for some 'UniformRange' type.+  uniformR ::+    (UniformRange a) =>+    -- | lower bound l (inclusive for discrete types)+    a ->+    -- | upper bound u (inclusive for discrete types)+    a ->+    -- | \(\sim \mathcal{U}(l, u)\).+    m a+ ---------------------------------------------------------------------------- -- Instances that lift probabilistic effects to standard tranformers. @@ -351,6 +369,9 @@  instance (MonadMeasure m) => MonadMeasure (IdentityT m) +instance (MonadUniformRange m) => MonadUniformRange (IdentityT m) where+  uniformR l u = lift $ uniformR l u+ instance (MonadDistribution m) => MonadDistribution (ExceptT e m) where   random = lift random   uniformD = lift . uniformD@@ -364,11 +385,17 @@   random = lift random   bernoulli = lift . bernoulli +instance (MonadUniformRange m) => MonadUniformRange (ExceptT e m) where+  uniformR l u = lift $ uniformR l u+ instance (MonadFactor m) => MonadFactor (ReaderT r m) where   score = lift . score  instance (MonadMeasure m) => MonadMeasure (ReaderT r m) +instance (MonadUniformRange m) => MonadUniformRange (ReaderT r m) where+  uniformR l u = lift $ uniformR l u+ instance (Monoid w, MonadDistribution m) => MonadDistribution (WriterT w m) where   random = lift random   bernoulli = lift . bernoulli@@ -379,6 +406,9 @@  instance (Monoid w, MonadMeasure m) => MonadMeasure (WriterT w m) +instance (Monoid w, MonadUniformRange m) => MonadUniformRange (WriterT w m) where+  uniformR l u = lift $ uniformR l u+ instance (MonadDistribution m) => MonadDistribution (StateT s m) where   random = lift random   bernoulli = lift . bernoulli@@ -390,6 +420,9 @@  instance (MonadMeasure m) => MonadMeasure (StateT s m) +instance (MonadUniformRange m) => MonadUniformRange (StateT s m) where+  uniformR l u = lift $ uniformR l u+ instance (MonadDistribution m) => MonadDistribution (ContT r m) where   random = lift random @@ -397,3 +430,6 @@   score = lift . score  instance (MonadMeasure m) => MonadMeasure (ContT r m)++instance (MonadUniformRange m) => MonadUniformRange (ContT r m) where+  uniformR l u = lift $ uniformR l u

@@ -23,6 +23,7 @@   ( MonadDistribution (random),     MonadFactor (..),     MonadMeasure,+    MonadUniformRange (uniformR),   ) import Control.Monad.Bayes.Inference.MCMC import Control.Monad.Bayes.Inference.RMSMC (rmsmc)@@ -45,6 +46,9 @@  instance (MonadDistribution m) => MonadDistribution (SMC2 m) where   random = lift random++instance (MonadUniformRange m) => MonadUniformRange (SMC2 m) where+  uniformR l u = lift $ uniformR l u  instance (Monad m) => MonadFactor (SMC2 m) where   score = SMC2 . score

@@ -45,6 +45,7 @@   ( MonadDistribution (..),     MonadFactor (..),     MonadMeasure,+    MonadUniformRange (..),     factor,   ) import Control.Monad.Bayes.Weighted@@ -95,6 +96,9 @@   bernoulli = lift . bernoulli   categorical = lift . categorical +instance (MonadUniformRange m) => MonadUniformRange (ListT m) where+  uniformR l u = lift $ uniformR l u+ instance (MonadFactor m) => MonadFactor (ListT m) where   score = lift . score @@ -102,7 +106,7 @@  -- | A collection of weighted samples, or particles. newtype PopulationT m a = PopulationT {getPopulationT :: WeightedT (ListT m) a}-  deriving newtype (Functor, Applicative, Monad, MonadIO, MonadDistribution, MonadFactor, MonadMeasure)+  deriving newtype (Functor, Applicative, Monad, MonadIO, MonadDistribution, MonadFactor, MonadMeasure, MonadUniformRange)  instance MonadTrans PopulationT where   lift = PopulationT . lift . lift

@@ -2,6 +2,8 @@ {-# LANGUAGE DerivingStrategies #-} {-# LANGUAGE GeneralizedNewtypeDeriving #-} {-# LANGUAGE ImportQualifiedPost #-}+{-# LANGUAGE UnboxedTuples #-}+{-# LANGUAGE UndecidableInstances #-}  -- | -- Module      : Control.Monad.Bayes.Sampler@@ -39,7 +41,11 @@         random,         uniform       ),+    MonadUniformRange+      ( uniformR+      ),   )+import Control.Monad.Primitive (PrimMonad) import Control.Monad.Reader (MonadIO, ReaderT (..)) import Control.Monad.ST (ST) import Control.Monad.Trans (MonadTrans)@@ -49,7 +55,7 @@  -- | The sampling interpretation of a probabilistic program -- Here m is typically IO or ST-newtype SamplerT g m a = SamplerT {runSamplerT :: ReaderT g m a} deriving (Functor, Applicative, Monad, MonadIO, MonadTrans)+newtype SamplerT g m a = SamplerT {runSamplerT :: ReaderT g m a} deriving (Functor, Applicative, Monad, MonadIO, MonadTrans, PrimMonad)  -- | convenient type synonym to show specializations of SamplerT -- to particular pairs of monad and RNG@@ -70,6 +76,9 @@   bernoulli p = SamplerT (ReaderT $ MWC.bernoulli p)   categorical ps = SamplerT (ReaderT $ MWC.categorical ps)   geometric p = SamplerT (ReaderT $ MWC.geometric0 p)++instance (StatefulGen g m) => MonadUniformRange (SamplerT g m) where+  uniformR l u = SamplerT (ReaderT $ uniformRM (l, u))  -- | Sample with a random number generator of your choice e.g. the one -- from `System.Random`.

@@ -28,6 +28,7 @@   ( MonadDistribution,     MonadFactor (..),     MonadMeasure,+    MonadUniformRange,     factor,   ) import Control.Monad.State (MonadIO, MonadTrans, StateT (..), lift, mapStateT, modify)@@ -36,7 +37,7 @@ -- | Execute the program using the prior distribution, while accumulating likelihood. newtype WeightedT m a = WeightedT (StateT (Log Double) m a)   -- StateT is more efficient than WriterT-  deriving newtype (Functor, Applicative, Monad, MonadIO, MonadTrans, MonadDistribution)+  deriving newtype (Functor, Applicative, Monad, MonadIO, MonadTrans, MonadDistribution, MonadUniformRange)  instance (Monad m) => MonadFactor (WeightedT m) where   score w = WeightedT (modify (* w))