Improve State.belief semantics and add explicit topology; mark getSolverParams as deprecated

src/Deprecated.jl

@@ -569,8 +569,10 @@ function deepcopyGraph(
     graphLabel::Symbol = Symbol(getGraphLabel(sourceDFG), "_cp_$(string(uuid4())[1:6])"),
     kwargs...,
 ) where {T <: AbstractDFG}
+    sp = getSolverParams(sourceDFG)
+    sp_kw = sp isa fieldtype(T, :solverParams) ? (; solverParams = sp) : (;)
     destDFG = T(;
-        solverParams = getSolverParams(sourceDFG),
+        sp_kw...,
         graph = sourceDFG.graph,
         agents = deepcopy(sourceDFG.agents),
         graphLabel,
@@ -710,9 +712,15 @@ function findShortestPathDijkstra(
             dfg,
             restrict_labels,
         )
-        return findPath(subdfg, from, to).path
+        result = try
+            findPath(subdfg, from, to)
+        catch ex
+            ex isa DFG.LabelNotFoundError ? nothing : rethrow()
+        end
+        return result === nothing ? Symbol[] : result.path
     else
-        return findPath(dfg, from, to).path
+        result = findPath(dfg, from, to)
+        return result === nothing ? Symbol[] : result.path
     end
 end
 

src/DistributedFactorGraphs.jl

@@ -61,13 +61,13 @@ using StructUtils: @kwarg, @tags
 export StructUtils # export for use in macros
 export AbstractManifold
 export ArrayPartition
-public @format_str   # from FileIO
+export @format_str   # from FileIO
 
 # DFG exports
 const DFG = DistributedFactorGraphs
 export DFG               # module alias for DistributedFactorGraphs
 export getLabel #TODO move
-public @defStateType # macro to define custom variable types
+export @defStateType # macro to define custom variable types
 
 # ------------------------------------------------------------------------------
 #  Types

src/Serialization/StateSerialization.jl

@@ -131,14 +131,18 @@ function unpackOldState(d)
     label = Symbol(d.solveKey)
     !isempty(d.covar) && error("covar field is not supported")
     if label == :parametric
-        belief =
-            StoredBelief(GaussianDensityKind(), statekind; means = vals, covariances = [BW])
+        belief = StoredHomotopyBelief(
+            RootsOnlyTopology(),
+            statekind;
+            means = vals,
+            shapes = [BW],
+        )
     else
-        belief = StoredBelief(
-            NonparametricDensityKind(),
+        belief = StoredHomotopyBelief(
+            LeavesOnlyTopology(),
             statekind;
             points = vals,
-            bandwidth = BW,
+            bandwidths = [BW],
         )
     end
     return State{T, getPointType(T)}(;

src/entities/State.jl

@@ -5,99 +5,160 @@
 abstract type AbstractStateType{N} end
 const StateType = AbstractStateType
 
-##==============================================================================
-## StoredBelief
-##==============================================================================
-abstract type AbstractDensityKind end
+# ==============================================================================
+#  StoredHomotopyBelief
+# ==============================================================================
+"""
+    AbstractHomotopyTopology
+
+Describes the physical layout of the nodes within a `StoredHomotopyBelief`.
+
+Since all beliefs in the Caesar ecosystem are fundamentally Homotopy densities, 
+this trait acts as a lightweight dispatch hint (a "Lens Selector"). It indicates 
+which parts of the tree (Roots vs. Leaves) are currently populated and how they 
+are wired, without requiring downstream packages to inspect the underlying vectors.
+
+**Role of the Topology Trait:**
+- **DFG:** Determines how to serialize and spatial-index the belief in the database.
+- **Visualizers:** Decides how to render the data (e.g., drawing ellipses for roots vs. a point cloud for leaves).
+- **IIF/AMP:** Selects the correct mathematical view to construct (e.g., `MvNormal` vs. `ManifoldKernelDensity` vs. a full `HomotopyDensity`).
+
+!!! note "State, not Strategy"
+    The trait purely describes the *current physical shape* of the data (e.g., "I currently only have Roots populated").
+    It does *not* dictate the solver strategy (e.g., "You must use a parametric solver"). 
+    The math engine is always free to convert or expand the data based on the graph's needs.
+
+**Extending:**
+If the standard tree-based Homotopy model does not fit your specific data layout 
+or solver requirements, you are encouraged to extend this abstract type with your 
+own custom topology struct. Alternatively, if you believe your use case represents 
+a missing core layout, please open an issue to discuss adding it to the 
+foundational ecosystem.
+"""
+abstract type AbstractHomotopyTopology end
+
+# --- 1. The Roots ---
+"L1 structural nodes only. No L2 samples. (Schema: `means`, `weights`, `shapes` populated. `points` empty.)"
+struct RootsOnlyTopology <: AbstractHomotopyTopology end
 
-"""Single Gaussian (mean + covariance)."""
-struct GaussianDensityKind <: AbstractDensityKind end
+# --- 2. The Leaves ---
+"L2 raw samples only. No L1 structure. (Schema: `points`, `bandwidths` populated. `means` empty.)"
+struct LeavesOnlyTopology <: AbstractHomotopyTopology end
 
-"""Kernel density / particle-based (points + shared bandwidth)."""
-struct NonparametricDensityKind <: AbstractDensityKind end
+# --- 3. The Full Trees ---
+"Tree packed in arrays using 2i, 2i+1 math.(Schema: L1 and L2 populated. Parent arrays empty.)"
+struct ImplicitTreeTopology <: AbstractHomotopyTopology end
 
-"""Homotopy between particles and Gaussian."""
-struct HomotopyDensityKind <: AbstractDensityKind end
+"Full tree using adjacency lists. (Schema: L1, L2, and Parent arrays fully populated.)"
+struct ExplicitTreeTopology <: AbstractHomotopyTopology end
 
-function StructUtils.lower(::StructUtils.StructStyle, p::AbstractDensityKind)
+function StructUtils.lower(::StructUtils.StructStyle, p::AbstractHomotopyTopology)
     return StructUtils.lower(Packed(p))
 end
-@choosetype AbstractDensityKind resolvePackedType
-
-# TODO naming? Density, DensityRepresentation, StoredBelief, BeliefState, etc?
-# TODO flatten in State? likeley not for easier serialization of points.
-@kwdef struct StoredBelief{T <: StateType, P}
-    statekind::T = T()# NOTE duplication for serialization, TODO maybe only in State and therefore belief cannot deserialize separately.
-    """Discriminator for which representation is active."""
-    densitykind::AbstractDensityKind = NonparametricDensityKind()
-
-    #--- Parametric fields (Gaussian / GMM / Homotopy leading modes) ---
-    """On-manifold component means.
-    Gaussian: length 1. Homotopy: leading (tree_kernel) means."""
-    means::Vector{P} = P[] # previously `val[1]` for Gaussian
-    """Component covariances, matching `means`."""
-    covariances::Vector{Matrix{Float64}} = Matrix{Float64}[] # previously `covar` existed but was stored in `bw` (hacky)
-    "Component weights, matching `means`."
+@choosetype AbstractHomotopyTopology resolvePackedType
+
+"""
+    StoredHomotopyBelief{T <: StateType, P}
+
+A multi-resolution "Grove of Trees" representing a manifold belief.
+Each tree can be as deep (ExplicitTreeTopology) or as shallow (RootsOnlyTopology) 
+as the evidence requires, but they all speak the same language of Nodes and Parents.
+
+These are the internal raw beliefs and need to be viewed through a lens such as 
+provided by AMP for features like pdf evaluation. Organized into structural 
+Tree/Branch layers (L1) and empirical Leaf layers (L2).
+
+!!! warning "Raw Data Container"
+    `StoredHomotopyBelief` is the raw data schema used for database storage and serialization.
+    Mutating this structure in-place is discouraged. Rather, construct a new `State` object 
+    and call `addState!` or `mergeState!`.
+"""
+@kwdef struct StoredHomotopyBelief{T <: StateType, P}
+    statekind::T = T()# NOTE duplication for serialization and self description.
+    """A hint for downstream solvers on how to interpret this data (The 'How')"""
+    topologykind::AbstractHomotopyTopology = LeavesOnlyTopology()
+
+    # L1 Nodes
+    """
+    [Order 0] The relative importance or probability of each node in L1.
+    """
     weights::Vector{Float64} = Float64[]
+    """
+    [Order 1] The location/center of each node, stored directly on the manifold.
+    """
+    means::Vector{P} = P[] # previously `val[1]` for Gaussian
+    """
+    [Order 2] The spread/curvature of each node (e.g., Covariance or Precision matrix). 
+    """
+    shapes::Vector{Matrix{Float64}} = Matrix{Float64}[] # previously `covar` existed but was stored in `bw` (hacky)
 
-    #--- Non-parametric / Homotopy leaves ---
-    """On-manifold sample points. For KDE/HomotopyDensity, these are the leaf kernel means."""
+    # L2 Nodes
+    """
+    The raw empirical samples on the manifold. Used for KDE and particle representations.
+    """
     points::Vector{P} = P[] # previously `val`
-    """Shared kernel bandwidth matrix used with ManifoldKernelDensity, see field `covar` for the parametric covariance"""
-    bandwidth::Union{Nothing, Matrix{Float64}} = zeros(getDimension(T), getDimension(T)) #previously `bw` ---
-    # bandwidth::Matrix{Float64} = zeros(getDimension(T), getDimension(T))
-    # TODO is bandwidth[s] matrix or vector or ::Vector{Matrix{Float64} or ::Vector{Vector{Float64}?
-    # JSON.parse(JSON.json(zeros(0, 0)), Matrix{Float64}) errors, so trying with nothing union
-end
+    """
+    The second-order bandwidths for the non-parametric points, supports variable bandwidth kernels.
+    """
+    bandwidths::Vector{Matrix{Float64}} = Matrix{Float64}[] #previously `bw` ---
 
-#FIXME remove old name before v0.29
-const BeliefRepresentation = StoredBelief
+    # --- Topology (The Hierarchy) ---
+    """
+    L1 Internal Topology: mean_parents[i] = j means means[i] is a child of means[j]. A value of 0 indicates a Root node.
+    """
+    mean_parents::Vector{Int} = Int[]
+    """
+    L2-to-L1 Bridge: point_parents[i] = j means points[i] is governed by means[j]. Points are leaves.
+    """
+    point_parents::Vector{Int} = Int[]
+end
 
-JSON.omit_empty(::Type{<:StoredBelief}) = true
+JSON.omit_empty(::Type{<:StoredHomotopyBelief}) = true
 
-function StoredBelief(T::AbstractStateType)
-    return StoredBelief{typeof(T), getPointType(T)}(; statekind = T)
+function StoredHomotopyBelief(T::AbstractStateType)
+    return StoredHomotopyBelief{typeof(T), getPointType(T)}(; statekind = T)
 end
 
-function StoredBelief(::NonparametricDensityKind, T::AbstractStateType; kwargs...)
-    return StoredBelief{typeof(T), getPointType(T)}(;
+function StoredHomotopyBelief(::LeavesOnlyTopology, T::AbstractStateType; kwargs...)
+    return StoredHomotopyBelief{typeof(T), getPointType(T)}(;
         statekind = T,
-        densitykind = NonparametricDensityKind(),
-        bandwidth = zeros(getDimension(T), getDimension(T)),
+        topologykind = LeavesOnlyTopology(),
+        bandwidths = [zeros(getDimension(T), getDimension(T))],
         kwargs...,
     )
 end
 
-function StoredBelief(::GaussianDensityKind, T::AbstractStateType; kwargs...)
-    return StoredBelief{typeof(T), getPointType(T)}(;
+function StoredHomotopyBelief(::RootsOnlyTopology, T::AbstractStateType; kwargs...)
+    return StoredHomotopyBelief{typeof(T), getPointType(T)}(;
         statekind = T,
-        densitykind = GaussianDensityKind(),
-        bandwidth = nothing,
+        topologykind = RootsOnlyTopology(),
         kwargs...,
     )
 end
 
 function StructUtils.fielddefaults(
     ::StructUtils.StructStyle,
-    ::Type{StoredBelief{T, P}},
+    ::Type{StoredHomotopyBelief{T, P}},
 ) where {T, P}
     return (
         statekind = T(),
-        densitykind = NonparametricDensityKind(),
+        topologykind = LeavesOnlyTopology(),
         means = P[],
-        covariances = Matrix{Float64}[],
+        shapes = Matrix{Float64}[],
         weights = Float64[],
         points = P[],
-        bandwidth = nothing,
+        bandwidths = Matrix{Float64}[],
+        mean_parents = Int[],
+        point_parents = Int[],
     )
 end
 
 function resolveStoredBeliefType(lazyobj)
     statekind = liftStateKind(lazyobj.statekind[])
-    return StoredBelief{typeof(statekind), getPointType(statekind)}
+    return StoredHomotopyBelief{typeof(statekind), getPointType(statekind)}
 end
 
-@choosetype StoredBelief resolveStoredBeliefType
+@choosetype StoredHomotopyBelief resolveStoredBeliefType
 
 ##==============================================================================
 ## State
@@ -121,7 +182,7 @@ $(TYPEDFIELDS)
     """
     Generic stored belief for this state.
     """
-    belief::StoredBelief{T, P} = StoredBelief{T, P}()#; statekind = T())
+    belief::StoredHomotopyBelief{T, P} = StoredHomotopyBelief{T, P}()#; statekind = T())
     """List of symbols for separator variables for this state, used in variable elimination and inference computations."""
     separator::Vector{Symbol} = Symbol[]
     """False if initial numerical values are not yet available or stored values are not ready for further processing yet."""
@@ -132,16 +193,7 @@ $(TYPEDFIELDS)
     marginalized::Bool = false #TODO renamed from ismargin v0.29
     """How many times has a solver updated this state estimate."""
     solves::Int = 0 # TODO renamed from solvedCount v0.29
-
-    #TODO belief container that can be used for active solver beliefs such as a HomotopyDensity
-    # The type is defined by a trait saved in the StoredBelief and 
-    # verbs such as `hydrate!(state)` `persist!(state)` can be used at data at checkpoints.
-    # Forcing an explicit `persist!` acts as a state checkpoint, 
-    #ensuring the graph only ever stores fully committed solver results rather than half-computed intermediate math.
-    # abstract type AbstractActiveBelief end
-    # active_belief::Base.RefValue{<:AbstractActiveBelief} = Ref{AbstractActiveBelief}() & (ignore = true,)
 end
-
 # OLD deprecated fields, removed in v0.29, kept here for reference during transition
 # val::Vector{P} = Vector{P}()
 # bw::Matrix{Float64} = zeros(0, 0)
@@ -153,6 +205,22 @@ end
 # events::Dict{Symbol, Threads.Condition} = Dict{Symbol, Threads.Condition}()    
 # dontmargin::Bool = false
 
+# ==============================================================================
+#  FUTURE VIEW WRAPPER (Internal DFG Placeholder)
+# ==============================================================================
+# NOTE: The `StoredHomotopyBelief` is currently expressive and fast enough that 
+# DFG does not need to store a resolved view next to it in memory. 
+#
+# If future profiling requires it, DFG will introduce a verbose View wrapper 
+# to hold the raw data alongside the instantiated read-only math object.
+#
+# abstract type AbstractHomotopyBeliefView end
+# 
+# struct HomotopyBeliefView{T, P, M} <: AbstractHomotopyBeliefView
+#     stored::StoredHomotopyBelief{T, P}
+#     math_engine::M # Read-only instantiated solver object (e.g., AMP.HomotopyDensity)
+# end
+
 ##------------------------------------------------------------------------------
 ## Constructors
 function State{T}(; kwargs...) where {T <: StateType}
@@ -178,7 +246,7 @@ function StructUtils.fielddefaults(
     ::Type{State{T, P}},
 ) where {T, P}
     return (
-        belief = StoredBelief{T, P}(; statekind = T()),
+        belief = StoredHomotopyBelief{T, P}(; statekind = T()),
         separator = Symbol[],
         initialized = false,
         observability = Float64[],
@@ -189,12 +257,12 @@ function StructUtils.fielddefaults(
 end
 
 refMeans(state::State) = state.belief.means
-refCovariances(state::State) = state.belief.covariances
+refCovariances(state::State) = state.belief.shapes
 refWeights(state::State) = state.belief.weights
 refPoints(state::State) = state.belief.points
-refBandwidth(state::State) = state.belief.bandwidth
-
-getDensityKind(state::State) = state.belief.densitykind
+refBandwidth(state::State) = state.belief.bandwidths[1]
+refBandwidths(state::State) = state.belief.bandwidths
+getTopologyKind(state::State) = state.belief.topologykind
 
 # we can also do somthing like this:
 function getComponent(state::State, i)

src/entities/equality.jl

@@ -20,7 +20,7 @@ implement compare if needed.
 const GeneratedCompareUnion = Union{
     Agent,
     Graphroot,
-    StoredBelief,
+    StoredHomotopyBelief,
     State,
     Blobentry,
     Bloblet,

src/services/AbstractDFG.jl

@@ -26,8 +26,19 @@ getGraphLabel(dfg::AbstractDFG) = getLabel(getGraph(dfg))
 
 """
     $(SIGNATURES)
+
+!!! warning "Deprecated"
+    `getSolverParams(dfg)` is deprecated in DFG v0.29 Pass `SolverParams` directly
+    to `solveTree!()` as a keyword argument instead.
 """
-getSolverParams(dfg::AbstractDFG) = dfg.solverParams
+function getSolverParams(dfg::AbstractDFG)
+    Base.depwarn(
+        "getSolverParams(dfg) is deprecated. SolverParams will be removed from the DFG object. " *
+        "Pass SolverParams directly to solveTree!() as a keyword argument instead.",
+        :getSolverParams,
+    )
+    return dfg.solverParams
+end
 
 """
     $(SIGNATURES)

src/services/compare.jl

@@ -155,7 +155,7 @@ end
 #Compare State
 function compare(a::State, b::State)
     refPoints(a) != refPoints(b) && @debug("val is not equal") === nothing && return false
-    refBandwidth(a) != refBandwidth(b) &&
+    refBandwidths(a) != refBandwidths(b) &&
         @debug("bw is not equal") === nothing &&
         return false
     # a.BayesNetOutVertIDs != b.BayesNetOutVertIDs &&