Unable to create measurements and gradients for Pose2Pose2

[Affie]

No description provided.

[sam-globus]

Stack trace:

┌ Warning: Unable to create measurements and gradients for Pose2Pose2{MvNormal{Float64,PDMats.PDMat{Float64,Array{Float64,2}},Array{Float64,1}}}(FullNormal(
│ dim: 3
│ μ: [10.0, 0.0, 1.0471975511965976]
│ Σ: [0.010000000000000002 0.0 0.0; 0.0 0.010000000000000002 0.0; 0.0 0.0 0.010000000000000002]
│ )
│ ) during prep of CCW, falling back on no-partial information assumption.  Enable ENV["JULIA_DEBUG"] = "IncrementalInference" for @debug printing to see the error.
└ @ IncrementalInference ~/.julia/packages/IncrementalInference/RyGCk/src/FactorGraph.jl:643
┌ Debug: MethodError(
│  CalcFactor:
│   .factor: Pose2Pose2{MvNormal{Float64,PDMats.PDMat{Float64,Array{Float64,2}},Array{Float64,1}}}
│ , (ProductRepr{Tuple{Array{Float64,1},Array{Float64,2}}}(([9.942635873464976, 0.16903394101000016], [0.0 -0.9629279328387408; 0.9629279328387408 0.0])),), 0x0000000000006dfa)
└ @ IncrementalInference ~/.julia/packages/IncrementalInference/RyGCk/src/FactorGraph.jl:645

[GearsAD]

Hmm, digging further this is a serialization issue. It's failing on rebuildFactorMetadata!(dfg, f).

[dehann]

Oh @GearsAD - don't mind the gradients warning (if that is what you meant by CCW issue), this is new experimental features preparing for JuliaRobotics/IncrementalInference.jl#1010 . At this stage, the code runs fine with or without precomputing the gradients. This is a core side issue that will sort it self out as work towards closing out 1010

[dehann]

I have not dug into the debug CalcFactor method error yet, maybe that is something different

[dehann]

I think Pose2Pose2 gradients work fine when constructing CCW directly in Julia