Fixed incorrect calls to to_linked_internal_transform

[torfjelde]

Came across these isuse when trying the following with JET.jl 👀

julia> using DynamicPPL, Distributions, JET

julia> @model function demo()
           x = Vector{Float64}(undef, 2)
           x[1] ~ Normal()
           x[2] ~ Normal()
           return x
       end
demo (generic function with 4 methods)

julia> model = demo();

julia> vi = DynamicPPL.typed_varinfo(model);

julia> for vn in keys(vi)
           DynamicPPL.set_flag!(vi, vn, "del")
       end^C

julia> f, argtypes = DynamicPPL.DebugUtils.gen_evaluator_call_with_types(model, DynamicPPL.typed_varinfo(model), SamplingContext());

julia> report_call(f, argtypes)
═════ 1 possible error found ═════
┌ demo(__model__::Model{…}, __varinfo__::TypedVarInfo{…}, __context__::SamplingContext{…}) @ Main ./REPL[124]:4
│┌ tilde_assume!!(context::SamplingContext{…}, right::Normal{…}, vn::VarName{…}, vi::TypedVarInfo{…}) @ DynamicPPL /Users/tfjelde/.julia/packages/DynamicPPL/Awq82/src/context_implementations.jl:144
││┌ tilde_assume(context::SamplingContext{…}, right::Normal{…}, vn::VarName{…}, vi::TypedVarInfo{…}) @ DynamicPPL /Users/tfjelde/.julia/packages/DynamicPPL/Awq82/src/context_implementations.jl:52
│││┌ tilde_assume(::Random.TaskLocalRNG, ::DefaultContext, ::SampleFromPrior, ::Normal{…}, ::VarName{…}, ::TypedVarInfo{…}) @ DynamicPPL /Users/tfjelde/.julia/packages/DynamicPPL/Awq82/src/context_implementations.jl:67
││││┌ tilde_assume(::DynamicPPL.IsLeaf, rng::Random.TaskLocalRNG, context::DefaultContext, sampler::SampleFromPrior, right::Normal{…}, vn::VarName{…}, vi::TypedVarInfo{…}) @ DynamicPPL /Users/tfjelde/.julia/packages/DynamicPPL/Awq82/src/context_implementations.jl:72
│││││┌ assume(rng::Random.TaskLocalRNG, sampler::SampleFromPrior, dist::Normal{…}, vn::VarName{…}, vi::TypedVarInfo{…}) @ DynamicPPL /Users/tfjelde/.julia/packages/DynamicPPL/Awq82/src/context_implementations.jl:260
││││││ no matching method found `to_linked_internal_transform(::TypedVarInfo{@NamedTuple{x::DynamicPPL.Metadata{Dict{VarName{:x, Accessors.IndexLens{Tuple{Int64}}}, Int64}, Vector{Normal{Float64}}, Vector{VarName{:x, Accessors.IndexLens{Tuple{Int64}}}}, Vector{Float64}, Vector{Set{DynamicPPL.Selector}}}}, Float64}, ::Normal{Float64})`: f = DynamicPPL.to_linked_internal_transform(vi::TypedVarInfo{@NamedTuple{x::DynamicPPL.Metadata{Dict{…}, Vector{…}, Vector{…}, Vector{…}, Vector{…}}}, Float64}, dist::Normal{Float64})
│││││└────────────────────

[torfjelde]

This raises the quesitons as to whether it would be worth making use of test_call and test_opt from JET.jl for certain cases, e.g. the demo models 👀

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 2 lines in your changes missing coverage. Please review.

Project coverage is 84.35%. Comparing base (48921d3) to head (361c45e).

Files with missing lines	Patch %	Lines
src/context_implementations.jl	0.00%	2 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #726   +/-   ##
=======================================
  Coverage   84.35%   84.35%           
=======================================
  Files          35       35           
  Lines        4212     4212           
=======================================
  Hits         3553     3553           
  Misses        659      659           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[coveralls]

Pull Request Test Coverage Report for Build 12058780410

Details

• 0 of 2 (0.0%) changed or added relevant lines in 1 file are covered.
• 25 unchanged lines in 4 files lost coverage.
• Overall coverage remained the same at 84.354%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/context_implementations.jl	0	2	0.0%

Files with Coverage Reduction	New Missed Lines	%
src/model.jl	1	93.68%
src/varinfo.jl	6	86.3%
src/simple_varinfo.jl	6	86.6%
src/threadsafe.jl	12	57.76%

Totals
Change from base Build 12056044639:	0.0%
Covered Lines:	3553
Relevant Lines:	4212

---

💛 - Coveralls