v3.2.1

AbstractMCMC v3.2.1

Diff since v3.2.0

Closed issues:

• Standard InferenceData-like API (#74)

Merged pull requests:

• Build stable documentation (#76) (@devmotion)
• Forward discard_initial and thinning to bundle_samples (#78) (@devmotion)

v3.2.0

AbstractMCMC v3.2.0

Diff since v3.1.0

Merged pull requests:

• add NestedSamplers to integration tests (#69) (@mileslucas)
• Add Documenter (#70) (@devmotion)
• Add MCMCSerial (#71) (@cpfiffer)

v3.1.0

AbstractMCMC v3.1.0

Diff since v3.0.2

Merged pull requests:

• Add Metadata struct (#64) (@cpfiffer)

v3.0.2

AbstractMCMC v3.0.2

Diff since v3.0.1

Closed issues:

• Update convergence sampling signature to use state (#67)

Merged pull requests:

• add state to isdone signature (#68) (@mileslucas)

v3.0.1

AbstractMCMC v3.0.1

Diff since v3.0.0

Merged pull requests:

• fix callback signature in convergence mcmcsample (#66) (@mileslucas)

v3.0.0

AbstractMCMC v3.0.0

Diff since v2.5.0

Breaking
callback functions passed to sample now also take the kwargs passed to samples as well as the current state, i.e. one must follow the following signature:

callback(rng, model, sampler, sample, state, iteration; kwargs...)

Merged pull requests:

• Include state and kwargs... to callback (#56) (@theogf)
• Add integration tests (#62) (@devmotion)

v2.5.0

AbstractMCMC v2.5.0

Diff since v2.4.0

Merged pull requests:

• Use TerminalLoggers also on Windows for Julia >= 1.5.3 (#61) (@devmotion)

v2.4.0

AbstractMCMC v2.4.0

Diff since v2.1.0

Closed issues:

• Option to thin (#49)
• TagBot trigger issue (#52)
• Documentation on role of AbstractModel instances for MCMC (#54)

Merged pull requests:

• Implemented thinning option in mcmcsample (#50) (@luiarthur)
• Add Github CI (#51) (@devmotion)
• Fixed missing semi-colon for kwargs in steps (#53) (@torfjelde)
• Allow to enable and disable progress logging globally (#59) (@devmotion)
• Limit update frequency of progress bars (#60) (@devmotion)

v2.1.0

AbstractMCMC v2.1.0

Diff since v2.0.0

Merged pull requests:

• Add documentation to README (#43) (@devmotion)
• Add chainsstack (#45) (@devmotion)
• Provide generic implementation of bundle_samples for Vector{T} (#46) (@devmotion)
• Add discard_initial keyword argument (#47) (@devmotion)
• Update CompatHelper and TagBot (#48) (@devmotion)

v2.0.0

AbstractMCMC v2.0.0

Diff since v1.0.1

Closed issues:

• [RFC] Sampler states (#31)
• Unnecessary copies in multi-threaded parallel sampling (#37)
• Bring back transitions_init and transitions_save! (#41)

Merged pull requests:

• Use BangBang to widen containers if needed (#35) (@devmotion)
• Check if the number of chains is greater than the number of samples per chain (#36) (@devmotion)
• Use nchains instead of Threads.nthreads() (#38) (@cpfiffer)
• Add a Sample transducer (#39) (@devmotion)
• Fix undefined reference error when type widening (#40) (@mohamed82008)
• Switch design to immutable samplers (#42) (@devmotion)