`MethodError` when attempting to construct a `Float32` LOBSTER

[ali-ramadhan]

Using the main branch this MWE

using Oceananigans
using OceanBioME

grid = RectilinearGrid(Float32,
	topology = (Bounded, Bounded, Bounded),
    size = (10, 10, 3),
    x = (0, 100),
    y = (0, 100),
    z = (-10, 0)
)

LOBSTER(; grid, surface_photosynthetically_active_radiation = (x, y, t) -> 100.0f0)

fails with

ERROR: MethodError: no method matching var"#TwoBandPhotosyntheticallyActiveRadiation#1"(::RectilinearGrid{…}, ::Float64, ::Float64, ::Float64, ::Float64, ::Float64, ::Float64, ::Float64, ::Float64, ::var"#9#10", ::Type{…})

Closest candidates are:
  var"#TwoBandPhotosyntheticallyActiveRadiation#1"(::Oceananigans.Grids.AbstractGrid{FT}, ::FT, ::FT, ::FT, ::FT, ::FT, ::FT, ::FT, ::FT, ::SPAR, ::Type{TwoBandPhotosyntheticallyActiveRadiation}) where {FT, SPAR}
   @ OceanBioME ~/.julia/packages/OceanBioME/7c6HZ/src/Light/2band.jl:97

Stacktrace:
 [1] top-level scope
   @ REPL[11]:1
Some type information was truncated. Use `show(err)` to see complete types.

---

Environment:

julia> versioninfo()
Julia Version 1.10.9
Commit 5595d20a287 (2025-03-10 12:51 UTC)
Build Info:
  Official https://julialang.org/ release
Platform Info:
  OS: Linux (x86_64-linux-gnu)
  CPU: 48 × AMD Ryzen Threadripper 7960X 24-Cores
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-15.0.7 (ORCJIT, znver3)
Threads: 1 default, 0 interactive, 1 GC (on 48 virtual cores)
Environment:
  LD_PRELOAD = /usr/NX/lib/libnxegl.so

[ali-ramadhan]

Hmmm, I don't think this pattern

OceanBioME.jl/src/Light/2band.jl

Lines 97 to 106 in 5dafe8f

	function TwoBandPhotosyntheticallyActiveRadiation(; grid::AbstractGrid{FT},
	water_red_attenuation::FT = 0.225, # 1/m
	water_blue_attenuation::FT = 0.0232, # 1/m
	chlorophyll_red_attenuation::FT = 0.037, # 1/(m * (mgChl/m³) ^ eʳ)
	chlorophyll_blue_attenuation::FT = 0.074, # 1/(m * (mgChl/m³) ^ eᵇ)
	chlorophyll_red_exponent::FT = 0.629,
	chlorophyll_blue_exponent::FT = 0.674,
	pigment_ratio::FT = 0.7,
	phytoplankton_chlorophyll_ratio::FT = 1.31,
	surface_PAR::SPAR = default_surface_PAR) where {FT, SPAR} # mgChl/mol N

maintains type stability based on this MWE:

using Oceananigans
using Oceananigans.Grids: AbstractGrid

f(grid::AbstractGrid{FT}, x::FT = 7.4) where FT = @show x, typeof(x)

then

julia> grid = RectilinearGrid(Float64,
           topology = (Bounded, Bounded, Bounded),
           size = (10, 10, 3),
           x = (0, 100),
           y = (0, 100),
           z = (-10, 0)
       );

julia> f(grid);
(x, typeof(x)) = (7.4, Float64)

but

julia> grid = RectilinearGrid(Float32,
           topology = (Bounded, Bounded, Bounded),
           size = (10, 10, 3),
           x = (0, 100),
           y = (0, 100),
           z = (-10, 0)
       );

julia> f(grid);
ERROR: MethodError: no method matching f(::RectilinearGrid{Float32, Bounded, Bounded, Bounded, Oceananigans.Grids.StaticVerticalDiscretization{…}, Float32, Float32, OffsetArrays.OffsetVector{…}, OffsetArrays.OffsetVector{…}, CPU}, ::Float64)

Closest candidates are:
  f(::AbstractGrid{FT}, ::FT) where FT
   @ Main REPL[25]:1
  f(::AbstractGrid{FT}) where FT
   @ Main REPL[25]:1

Stacktrace:
 [1] f(grid::RectilinearGrid{Float32, Bounded, Bounded, Bounded, Oceananigans.Grids.StaticVerticalDiscretization{…}, Float32, Float32, OffsetArrays.OffsetVector{…}, OffsetArrays.OffsetVector{…}, CPU})
   @ Main ./REPL[25]:1
 [2] top-level scope
   @ REPL[34]:1
Some type information was truncated. Use `show(err)` to see complete types.

[jagoosw]

Thank you for finding this! I hadn't realised that this didn't work I'll have a look at how we can better construct, maybe the new oceaniangas default types will be helpful

[johnryantaylor]

We have been discussing type stability here. Maybe this is something that @AdelekeBankole can look at if he has time

[AdelekeBankole]

Thank you @ali-ramadhan and @johnryantaylor, yes I will also have a look at this issue.

[jagoosw]

This will be resolved in #276

[ali-ramadhan]

This is still an issue but I'll open a PR with a fix.

[jagoosw]

Sorry! Do you know where its coming from?

[ali-ramadhan]

No worries at all! I approved #276 but should have remembered this issue and mentioned that we should add tests for this in that PR.

It didn't take long to fix and I just opened #324 which includes tests. Hoping the tests pass there!