v0.20.3

IncrementalInference v0.20.3

Diff since v0.20.2

Closed issues:

• Segfault on copyGraph! / deepcopy_internal (jl1.4.2) (#1113)
• Does not yet run Julia 1.6-beta1, typename error (#1120)
• "deconvolution tools via differential factors" test fail on julia1.6-beta (#1123)

Merged pull requests:

• fast forward dfg integration tests (#1119) (@dehann)
• towards #1010, consolidate numerics (#1121) (@dehann)
• _findSubgraphsFactorType use nameof (#1122) (@Affie)
• hotfix to restore tests (#1125) (@dehann)
• consolidate to one _solveCCWNumeric! (#1126) (@dehann)
• bug fix (#1127) (@dehann)
• fast-forward dfg integration tests (#1128) (@dehann)
• v0.20.3-rc1 (#1130) (@dehann)

v0.20.2

IncrementalInference v0.20.2

Diff since v0.20.1

Merged pull requests:

• v0.20.2-rc1 (#1117) (@dehann)

v0.20.1

IncrementalInference v0.20.1

Diff since v0.20.0

Merged pull requests:

• update DERelative on new CalcFactor API (#1107) (@dehann)
• Switch to Github CI (#1108) (@dehann)
• update multihypo for CalcFactor (#1109) (@dehann)
• remaining CalcFactor deprecations (#1110) (@dehann)
• rm deprecated code for v0.20 (#1111) (@dehann)
• v0.20.1-rc1 (#1112) (@dehann)
• add testFactorResidualBinary (#1114) (@dehann)
• v0.20.1-rc2 (#1115) (@dehann)

v0.20.0

IncrementalInference v0.20.0

Diff since v0.19.1

Closed issues:

• standardize special sampler using new FMd / CalcResidual (#927)
• tag v0.19.1 (#1092)

Merged pull requests:

• _getCCW clean up and improvements (#1095) (@dehann)
• deconv cleanup, more testing before CalcResidual (#1097) (@dehann)
• basic deconv cleanup before CalcFactor switch (#1098) (@dehann)
• first test pass on CalcFactor (LinearRelative) (#1100) (@dehann)
• towards CalcFactor and cleanup (#1101) (@dehann)
• introduce getSample(::CalcFactor, ::Int) (#1102) (@dehann)

v0.19.1

IncrementalInference v0.19.1

Diff since v0.19.0

Closed issues:

• parseusermultihypo assertion error (#1086)

Merged pull requests:

• fast forward dfg integration testing (#1084) (@dehann)
• fix multihypo tolerance, needs backport (#1087) (@dehann)
• fix multihypo tolerance, needs backport (#1088) (@dehann)
• v0.19.1-rc1 (#1090) (@dehann)

v0.18.2

IncrementalInference v0.18.2

Diff since v0.18.1

Closed issues:

• Remove Union types from FactorMetadata (#383)
• Clique label field and label attribute is a bit confusing with getter/setter. (#551)
• test repeatCSMStep! (#742)
• Consolidate buildTreeFromOrdering! functions (#983)
• Bug, FMD not doing multihypo correctly (#1025)
• xdot error in drawTree (#1043)
• rn resetBuildTree! -> buildTreeReset! (#1045)
• animateStateMachineHistoryIntervalCompound not found from csmAnimate (#1046)
• rn internal tags,eg :__LIKELIHOODMESSAGE__ (#1048)
• rn tree.variableOrder -> tree.eliminationOrder (#1049)
• Missing deprecation for fetchMsgUpThis (#1053)
• Double adding DOWN COMMON MsgPriors (#1068)
• Deprecate AbstractRelativeRoots shuffleX, preserve partial (#1072)
• tag v0.19.0 (#1085)

Merged pull requests:

• standardize eliminationOrder (#1050) (@dehann)
• Updating LikelihoodMessage and _MsgJointRelPr (#1054) (@dehann)
• fast-forward DFG testing (#1064) (@dehann)
• updates and maintenance to csm debugging tools (#1066) (@dehann)
• refactoring fmd part 1, (#1070) (@dehann)
• some cleanup and moving code around (#1071) (@dehann)
• remove partial Roots factor (#1073) (@dehann)
• simplify numericCCW Minimize (#1074) (@dehann)
• drop shuffleX, depr (::ccw)(x...) (#1075) (@dehann)
• drop ccw.cpt[].Y (#1076) (@dehann)
• rm fmd fields, fix fmd multihypo (#1077) (@dehann)
• update AbstractRelative Factor and FactorMinimize (#1078) (@dehann)
• add depr for fetchMsgUpThis (#1079) (@dehann)
• fix draw/show tree during solve, (#1080) (@dehann)
• rn internal tags :LIKELIHOOD... (#1081) (@dehann)
• depr _defaultFactorMetadata, resetBuildTree! (#1082) (@dehann)
• fix multihypo tolerance, needs backport (#1088) (@dehann)

v0.19.0

IncrementalInference v0.19.0

Diff since v0.18.1

Closed issues:

• Remove Union types from FactorMetadata (#383)
• Clique label field and label attribute is a bit confusing with getter/setter. (#551)
• test repeatCSMStep! (#742)
• Consolidate buildTreeFromOrdering! functions (#983)
• Bug, FMD not doing multihypo correctly (#1025)
• xdot error in drawTree (#1043)
• rn resetBuildTree! -> buildTreeReset! (#1045)
• animateStateMachineHistoryIntervalCompound not found from csmAnimate (#1046)
• rn internal tags,eg :__LIKELIHOODMESSAGE__ (#1048)
• rn tree.variableOrder -> tree.eliminationOrder (#1049)
• Missing deprecation for fetchMsgUpThis (#1053)
• Double adding DOWN COMMON MsgPriors (#1068)
• Deprecate AbstractRelativeRoots shuffleX, preserve partial (#1072)

Merged pull requests:

• standardize eliminationOrder (#1050) (@dehann)
• Updating LikelihoodMessage and _MsgJointRelPr (#1054) (@dehann)
• fast-forward DFG testing (#1064) (@dehann)
• updates and maintenance to csm debugging tools (#1066) (@dehann)
• refactoring fmd part 1, (#1070) (@dehann)
• some cleanup and moving code around (#1071) (@dehann)
• remove partial Roots factor (#1073) (@dehann)
• simplify numericCCW Minimize (#1074) (@dehann)
• drop shuffleX, depr (::ccw)(x...) (#1075) (@dehann)
• drop ccw.cpt[].Y (#1076) (@dehann)
• rm fmd fields, fix fmd multihypo (#1077) (@dehann)
• update AbstractRelative Factor and FactorMinimize (#1078) (@dehann)
• add depr for fetchMsgUpThis (#1079) (@dehann)
• fix draw/show tree during solve, (#1080) (@dehann)
• rn internal tags :LIKELIHOOD... (#1081) (@dehann)
• depr _defaultFactorMetadata, resetBuildTree! (#1082) (@dehann)

v0.18.1

IncrementalInference v0.18.1

Diff since v0.18.0

Closed issues:

• TreeClique index field and underlining graphs structure indexes (#540)
• [DEV] Overview of issues related to parametric and consolation (#558)
• BayesTree alternative to Graphs.jl (#609)
• Manifest VariableTypes (Singleton model) and new FMD (#783)
• Parent CSM should resolve sibling init order (if partials or similar) (#954)
• tag v0.18.0 (#1041)
• Debug tools are broken on new CSM (#1055)

Merged pull requests:

• WIP to towards better correlation of differential samples (#1026) (@dehann)
• Allow dispatch on cliq id vs graph index by using CliqueId (#1047) (@Affie)
• add factor EuclidDistance (#1052) (@dehann)
• CliqueId{Int64}.value on csm debug, and test (#1056) (@dehann)
• add repeatCSMStep (#1057) (@dehann)
• add csmc.enableLogging, docs, and depr (#1058) (@dehann)
• add seq to printHistoryLine, and maint (#1060) (@dehann)
• v0.18.1-rc1 (#1062) (@dehann)

v0.18.0

IncrementalInference v0.18.0

Diff since v0.17.0

Closed issues:

• draw fg and bayes tree with xdot, separate pdf (#691)
• Update and document CSM colors (#1007)
• Julia error and exit on 1000 particles (#1013)
• prepBatchTree! not drawing bayes net (#1015)
• drawTree not working with filepath and imgs keyword arguments (#1016)
• rename ODERelative -> DERelative (#1031)
• some graphs cannot be initialized, need warning (#1035)

Merged pull requests:

• default to params.N (#1014) (@dehann)
• User path in drawTree with spyCliqMat (#1017) (@Affie)
• add docs to Mixture (#1018) (@dehann)
• more fixes, and docs (#1019) (@dehann)
• initial ODERelative (#1020) (@dehann)
• new approxConv for chain of elements (#1021) (@dehann)
• Removing bayes net draw while deprecating older Graphs.jl (#1022) (@dehann)
• deprecate setCliqueDrawColor! (#1023) (@dehann)
• new test for fmd bug on multihypo (#1024) (@dehann)
• Towards removing Graphs.jl for tree (#1027) (@Affie)
• Rename softtype -> variableType, getSofttype -> getVariableType (#1028) (@Affie)
• Update BeliefTypes.jl (#1029) (@dehann)
• Update FactorGraphTypes.jl (#1030) (@dehann)
• setMarginalized! function (#1032) (@dehann)
• Set MetaBayesTree as default tree (#1033) (@Affie)
• add deleteClique! function (#1034) (@Affie)
• Limit treeinit iters with limittreeinit_iters and @error cleanly exit (#1036) (@Affie)
• fix warn of name collision Dates.Second (#1037) (@dehann)
• rm Graphs.jl dependency from IIF (#1038) (@dehann)
• Rename ODERelative to DERelative (#1039) (@dehann)
• cleanup before release (#1040) (@dehann)

v0.17.0

IncrementalInference v0.17.0

Diff since v0.16.2

Closed issues:

• Finalize state for CSM loggers (#321)
• csmc.condition should all be made thread safe (#597)
• ensure csm up msg updated in recycle after repeat down (#599)
• Tree init regression in buildup to DFG v0.6 IIF v0.9 (#602)
• Decide if msg channels in BTND or edges only (#675)
• Undo CliqStatus back to an enum (#744)
• Consolidate btnd.solvableDims into LikelihoodMessage (#910)
• Consolidate CSM down msg functions (#938)
• Should isCliqMarginalizedFromVars be deprecated (#962)
• getSuggestedPPE --> getPPESuggested (#973)
• BoundsError on getCliqVarInitOrderUp, solving Hexagonal (#982)
• Take! Clique State Machine Resolution (#989)
• solveCliqDownFrontalProducts! causes BoundsError and DimensionMismatch errors (#993)
• switch to defVariable macro (#995)
• Unable to open document “file:///tmp/caesar/random/bt_4684e640.pdf”. (#996)
• Bug - Mixture Prior is not an AbstractPrior (#998)
• Intermittent error in addLikelihoodsDifferential! (solveFactorMeasurements) (#1009)

Merged pull requests:

• restore 458 461 test (#501) (@dehann)
• Move fetch functions to attic and reorganize (#987) (@Affie)
• Maintenance on take and Restore CliqueStatus enum (#988) (@Affie)
• CSM updates (#991) (@Affie)
• Split up CSM and clique up recycling (#992) (@Affie)
• fix #993 (#994) (@dehann)
• fix #996, FourDoorExample.jl (#997) (@dehann)
• Mixture isPrior (#1000) (@Affie)
• Use @defVariable for Sphere1D and ContinuousScalar (#1001) (@Affie)
• Basic skip upsolve state (#1002) (@Affie)
• useMsgLikelihoods=true (#1003) (@dehann)
• Split down CSM (#1004) (@Affie)
• drop deprecations at end v0.16.x (#1005) (@dehann)
• Updates and cleanup of CSM (#1006) (@Affie)
• hists as NamedTuple CSMHistoryTuple (#1008) (@Affie)
• fix solveFactorMeasurements (#1011) (@Affie)