replace #1291 , some code cleanup

src/ApproxConv.jl

@@ -252,7 +252,7 @@ function calcVariableDistanceExpectedFractional(ccwl::CommonConvWrapper,
 end
 
 # Add entrypy on a point in `points` on manifold M, only on dimIdx if in p 
-function addEntropyOnManifoldHack!( M::ManifoldsBase.AbstractManifold,
+function addEntropyOnManifold!( M::ManifoldsBase.AbstractManifold,
                                     points::Union{<:AbstractVector{<:Real},SubArray}, 
                                     dimIdx::AbstractVector, 
                                     spreadDist::Real,
@@ -281,16 +281,6 @@ function addEntropyOnManifoldHack!( M::ManifoldsBase.AbstractManifold,
 
   end
   #
-  # manis = convert(Tuple, M) # LEGACY, TODO REMOVE
-  # # TODO deprecate
-  # maniAddOps, _, _, _ = buildHybridManifoldCallbacks(manis)
-  # # add 1σ "noise" level to max distance as control
-  # # 1:size(addEntr, 1)
-  # for dim in dimIdx, idx in 1:length(addEntr)
-  #   if (p === :) || dim in p
-  #     addEntr[idx][dim] = maniAddOps[dim](addEntr[idx][dim], spreadDist*(rand()-0.5))
-  #   end
-  # end
   nothing
 end
 
@@ -333,7 +323,7 @@ function computeAcrossHypothesis!(ccwl::Union{<:CommonConvWrapper{F},
       # consider duplicate convolution approximations for inflation off-zero
       # ultimately set by dfg.params.inflateCycles
       for iflc in 1:inflateCycles
-        addEntropyOnManifoldHack!(mani, addEntr, 1:getDimension(mani), spreadDist, cpt_.p)
+        addEntropyOnManifold!(mani, addEntr, 1:getDimension(mani), spreadDist, cpt_.p)
         # no calculate new proposal belief on kernels `allelements[count]`
         skipSolve ? @warn("skipping numerical solve operation") : approxConvOnElements!(ccwl, allelements[count])
       end
@@ -348,7 +338,7 @@ function computeAcrossHypothesis!(ccwl::Union{<:CommonConvWrapper{F},
         addEntr = view(ccwl.params[sfidx], allelements[count])
         # dynamic estimate with user requested speadNH of how much noise to inject (inflation or nullhypo)
         spreadDist = calcVariableDistanceExpectedFractional(ccwl, sfidx, certainidx, kappa=spreadNH)
-        addEntropyOnManifoldHack!(mani, addEntr, 1:getDimension(mani), spreadDist)
+        addEntropyOnManifold!(mani, addEntr, 1:getDimension(mani), spreadDist)
 
     elseif hypoidx == 0
       # basically do nothing since the factor is not active for these allelements[count]
@@ -358,7 +348,7 @@ function computeAcrossHypothesis!(ccwl::Union{<:CommonConvWrapper{F},
       # dynamic estimate with user requested speadNH of how much noise to inject (inflation or nullhypo)
       spreadDist = calcVariableDistanceExpectedFractional(ccwl, sfidx, certainidx, kappa=spreadNH)
       # # make spread (1σ) equal to mean distance of other fractionals
-      addEntropyOnManifoldHack!(mani, addEntr, 1:getDimension(mani), spreadDist)
+      addEntropyOnManifold!(mani, addEntr, 1:getDimension(mani), spreadDist)
     else
       error("computeAcrossHypothesis -- not dealing with multi-hypothesis case correctly")
     end
@@ -518,7 +508,7 @@ function evalPotentialSpecific( Xi::AbstractVector{<:DFGVariable},
         addEntr[m] .= ccwl.measurement[1][m]
       end
       # ongoing part of RoME.jl #244
-      addEntropyOnManifoldHack!(mani, addEntrNH, 1:getDimension(mani), spreadDist)
+      addEntropyOnManifold!(mani, addEntrNH, 1:getDimension(mani), spreadDist)
   else
     i = 0
     for dimnum in fnc.partial
@@ -529,7 +519,7 @@ function evalPotentialSpecific( Xi::AbstractVector{<:DFGVariable},
       # @show size(addEntr), dimnum, nhmask
       addEntrNHp = view(view(addEntr, (1:length(ahmask))[ahmask]), dimnum)
       # ongoing part of RoME.jl #244
-      addEntropyOnManifoldHack!(mani, addEntrNHp, dimnum:dimnum, spreadDist)
+      addEntropyOnManifold!(mani, addEntrNHp, dimnum:dimnum, spreadDist)
     end
   end
   return addEntr

src/FGOSUtils.jl

@@ -140,19 +140,7 @@ function manikde!(pts::AbstractVector{P},
   return AMP.manikde!(M, pts)
 end
 
-# manikde!( pts::AbstractVector{<:Real}, 
-#           vartype::Type{<:ContinuousScalar}) = manikde!(reshape(pts,1,:), vartype)
-#
 
-# TEMPORARY legacy wrapper
-# function manikde!(ptsArr::Vector{Vector{Float64}}, 
-#                   bw::Vector{Float64}, 
-#                   varType::Union{InstanceType{<:InferenceVariable}, InstanceType{<:AbstractFactor}} )
-#   #
-#   arr = Matrix{Float64}(undef, length(ptsArr[1]), length(ptsArr))
-#   @cast arr[i,j] = ptsArr[j][i]
-#   manikde!( arr, bw, varType )  
-# end
 
 """
     $SIGNATURES

src/FactorGraph.jl

@@ -594,20 +594,21 @@ Notes
 - Will not work in all situations, but good enough so far.
   - # TODO standardize via domain or manifold definition...??
 """
-function calcZDim(cf::CalcFactor{T}) where {T <: FunctorInferenceType}
+function calcZDim(cf::CalcFactor{T}) where {T <: AbstractFactor}
   #
-  # zdim = T != GenericMarginal ? size(getSample(usrfnc, 2)[1],1) : 0
-  zdim = if T != GenericMarginal
-    # vnds = Xi # (x->getSolverData(x)).(Xi)
-    # NOTE try to make sure we get matrix back (not a vector)
-    smpls = sampleFactor(cf, 2)[1]
-    length(smpls[1])
-  else
-    0
-  end
-  return zdim
+
+  # NOTE try to make sure we get matrix back (not a vector)
+  smpls = sampleFactor(cf, 2)[1]
+  return length(smpls[1])
 end
 
+# FIXME THIS IS NEW REPLACEMENT FUNCTION
+# function calcZDim(cf::CalcFactor{T}) where T <: AbstractFactor
+#   return manifold_dimension(getManifold(cf.factor))
+# end
+
+calcZDim(cf::CalcFactor{<:GenericMarginal}) = 0
+
 calcZDim(cf::CalcFactor{<:ManifoldPrior}) = manifold_dimension(cf.factor.M)
 
 

src/IncrementalInference.jl

@@ -422,6 +422,7 @@ include("Factors/GenericFunctions.jl")
 # Refactoring in progress
 include("Factors/MsgLikelihoods.jl")
 
+include("entities/GraphConstraintTypes.jl")
 include("CliqueTypes.jl")
 
 include("JunctionTreeTypes.jl")
@@ -444,7 +445,6 @@ include("TreeBasedInitialization.jl")
 include("HeatmapSampler.jl")
 
 # special variables and factors, see RoME.jl for more examples
-include("GraphConstraintTypes.jl")
 include("Factors/Mixture.jl")
 include("Factors/DefaultPrior.jl")
 include("Factors/LinearRelative.jl")