JuliaDiff / JuliaDiff/ChainRules.jl
generic frule for constructors?
Nobody has claimed this yet.
- Dominant language
- Julia
- Stars
- 475
- Forks
- 98
- PR merge metrics
- No merged PRs in 30d
Description
Like for https://github.com/JuliaDiff/ChainRules.jl/issues/153
there is a generic frule for default constructors.
It relies on the arguments being the same as the fields so would have to detect that.
Its basically:
function frule(::Type{P}, args..., _, dargs...) where P
y = P(args...)
dy = Composite{P}(; zip(fieldnames(P, dargs)...)
end
the more serious implementation might be:
function frule(::Type{P}, all_args...) where P
nargs = fieldcount(P)
length(all_args) == 2nargs + 1 || return nothing
args = @inbounds all_args[1:nargs]
all(typeof.(args) .== fieldtypes(P)) || return nothing
dargs = @inbounds all_args[end-nargs : end] # skip dself as constructors never functors
y = P(args...)
dy = Composite{P}(; zip(fieldnames(P, dargs)...)
end
But this may well need to be written as a generated function if it doesn't constant fold good.
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reviewing ChainRules.jl issue #153 and the existing frule handling for default constructors. The issue names no implementation file or test; determine whether matching constructor arguments to field types and producing the requested primal and tangent values needs a generated function, then establish tests for the supported and rejected signatures.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- julia
- Domain
- backend
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 30/100