Bump version for simlogger nt patch (#8)

Project.toml

@@ -1,7 +1,7 @@
 name = "FSimBase"
 uuid = "731e945d-f3ed-49a1-bee2-3998705ef155"
 authors = ["JinraeKim <kjl950403@gmail.com> and contributors"]
-version = "0.2.0"
+version = "0.2.1"
 
 [deps]
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
@@ -14,11 +14,11 @@ Transducers = "28d57a85-8fef-5791-bfe6-a80928e7c999"
 
 [compat]
 DataFrames = "1.2"
-DiffEqBase = "6.73"
-DiffEqCallbacks = "2.17"
+DiffEqBase = "6.77"
+DiffEqCallbacks = "2.18"
 NamedTupleTools = "0.13"
 Reexport = "1.2"
-SimulationLogger = "0.2"
+SimulationLogger = "0.3"
 Transducers = "0.4"
 julia = "1"
 

src/APIs/simulation.jl

@@ -15,28 +15,6 @@ function _initialise(__Problem, state0, __dyn, p, t0, tf)
  prob
 end
 
-"""
-Extended to deal with empty Dict.
-"""
-function _namedtuple(x::Dict)
- if x == Dict()
- return NamedTuple() # empty NamedTuple
- else
- return namedtuple(x)
- end
-end
-
-function recursive_namedtuple(x::Any)
- x
-end
-
-function recursive_namedtuple(d::Dict)
- _namedtuple(
- Dict(k => recursive_namedtuple(v) for (k, v) in d)
- )
-end
-
-
 """
 Simulator struct.
 # NOTICE
@@ -116,9 +94,9 @@ end
 function _push!(df::DataFrame, integrator::DEIntegrator, log_func)
  x = integrator.u
  t = integrator.t
- __log_dict__ = log_func(x, t, integrator)
- __log_nt__ = recursive_namedtuple(__log_dict__)
- push!(df, (; time=t, sol=__log_nt__))
+ __logger__ = log_func(x, t, integrator)
+ # __log_nt__ = recursive_namedtuple(__logger__)
+ push!(df, (; time=t, sol=__logger__))
 end
 
 function Base.push!(simulator::Simulator, df::DataFrame, flag=true)
@@ -156,7 +134,7 @@ function DiffEqBase.solve(simulator::Simulator;
  elseif saveat != nothing && savestep != nothing
  error("Assign values of either `saveat` or `savestep`")
  end
- saved_values = SavedValues(Float64, Dict)
+ saved_values = SavedValues(Float64, NamedTuple)
  if log_func != nothing
  cb_save = SavingCallback(log_func, saved_values;
  saveat=saveat, tdir=Int(sign(tf-t0)))
@@ -167,74 +145,12 @@ function DiffEqBase.solve(simulator::Simulator;
  kwargs...)
  df = DataFrame(
  time = saved_values.t,
- sol = saved_values.saveval |> Map(recursive_namedtuple) |> collect,
+ sol = saved_values.saveval,
  )
  df
 end
 
 
-# function sim(state0, dyn, p=nothing;
-# solver=nothing, # DifferentialEquations.jl will find a default solver
-# t0=0.0, tf=1.0,
-# callback::DECallback=CallbackSet(),
-# log_off=false,
-# saveat=nothing,
-# savestep=nothing,
-# kwargs...,
-# )
-# if saveat == nothing && savestep == nothing
-# saveat = [] # default
-# elseif saveat != nothing && savestep == nothing
-# # nothing
-# elseif saveat == nothing && savestep != nothing
-# saveat = t0:savestep:tf
-# elseif saveat != nothing && savestep != nothing
-# error("Assign values of either `saveat` or `savestep`")
-# end
-# tspan = (t0, tf)
-# iip = true
-# __dyn = (dx, x, p, t) -> dyn(dx, x, p, t)
-# prob = ODEProblem{iip}(__dyn, state0, tspan, p) # true: isinplace
-# saved_values = SavedValues(Float64, Dict)
-# cb_save = nothing
-# if log_off == false
-# # logging function
-# if isinplace(prob)
-# __log_indicator__ = __LOG_INDICATOR__() # just an indicator for logging
-# if hasmethod(dyn, Tuple{Any, Any, Any, Any, __LOG_INDICATOR__})
-# log_func = function (x, t, integrator::DEIntegrator; kwargs...)
-# x = copy(x) # `x` merely denotes a "view"
-# dyn(zero.(x), x, integrator.p, t, __log_indicator__; kwargs...)
-# end
-# cb_save = SavingCallback(log_func, saved_values;
-# saveat=saveat, tdir=Int(sign(tspan[2]-tspan[1])))
-# end
-# else
-# error("Not tested")
-# end
-# callback = CallbackSet(callback, cb_save) # save values "after all other callbacks"
-# end
-# sol = solve(prob, solver;
-# callback=callback,
-# kwargs...)
-# prob, sol
-# if log_off == true
-# return prob, sol
-# else
-# # recursive NamedTuple conversion from Dict; https://discourse.julialang.org/t/how-to-make-a-named-tuple-from-a-dictionary/10899/34?u=ihany
-# recursive_namedtuple(x::Any) = x
-# recursive_namedtuple(d::Dict) = _namedtuple(
-# Dict(k => recursive_namedtuple(v) for (k, v) in d)
-# )
-# df = DataFrame(
-# time = saved_values.t,
-# sol = saved_values.saveval |> Map(recursive_namedtuple) |> collect,
-# )
-# return prob, df
-# end
-# end
-
-
 """
 # Notes
 - The basic concept is borrowed from [an MRAC example](https://jonniedie.github.io/ComponentArrays.jl/stable/examples/adaptive_control/).

test/minimal.jl

@@ -18,7 +18,7 @@ function main()
  end
  simulator = Simulator(
  state0, dynamics!, p;
- Problem=ODEProblem,
+ Problem=:ODE,
  solver=Tsit5(),
  tf=tf,
  )