Poincare's Section for positive momentum about taylorintegration.jl HOT 17 CLOSED

Which version (of Julia and the package) are you using?

from taylorintegration.jl.

I am using the 1.0.3 version of Julia and the 0.4.1 version of TaylorIntegration.

from taylorintegration.jl.

What about TaylorSeries.jl? I am trying to have the same set up as you have to see if I can reproduce the problem.

from taylorintegration.jl.

Note that the documentation was created with Julia v1.0, which I guess corresponds to Julia v1.0.3, and there was no problem...

from taylorintegration.jl.

So, I've checked it again and I can't reproduce the problem. Everything runs fine for me using Julia 1.0.3, TaylorIntegration v0.4.1, and TaylorSeries v0.8.0. I haven't checked it, but everything should also run without any problem for more recent versions of TaylorSeries.

from taylorintegration.jl.

I'm sorry, I realized I wasn't completely clear about my issue. I got the error from this part of the code:

using TaylorIntegration

for i in 1:nconds
    rand1 = rand()
    rand2 = rand()
    x_ini .= x0 .+ 0.005 .* [0.0, sqrt(rand1)*cos(2pi*rand2), 0.0, sqrt(rand1)*sin(2pi*rand2)]
    px!(x_ini, E0)

    tv_i, xv_i, tvS_i, xvS_i, gvS_i = taylorinteg(henonheiles!, g, x_ini, 0.0, 135.0,
        25, 1e-25, maxsteps=30000);
    tvSv[i] = vcat(0.0, tvS_i)
    xvSv[i] = vcat(transpose(x_ini), xvS_i)
    gvSv[i] = vcat(0.0, gvS_i)
end

However, it seems that the real problem is the definition of the function:

function g(t, x, dx)
    px = constant_term(x[3])
    
    if px > zero(px)
        # ...return x
        return x[1]
    else
        #otherwise, discard the crossing
        return nothing
    end
end

If I change g(t,x,dx)for

g(t, x, dx)=x[1]

The program works

from taylorintegration.jl.

I actually checked if I could reproduce the documentation (step by step, copy-pasting what it is written), and I had no problem. That's the reason I gave all the details of the packages that I used. Yet, I know that the definition of g is kind of subtle.

@PerezHz Do you have an idea what is going on here?

from taylorintegration.jl.

Sorry for the late reply... I also was unable to reproduce the problem; it would seem as if the version is lower than 0.4.0, since it can't find a method of surfacecrossing which was introduced in 0.4.0 (specifically, those methods were added in #62), are you completely sure your installed version of TaylorIntegration is 0.4.1? Could you please post the output in the REPL of using Pkg; Pkg.status()?

from taylorintegration.jl.

Yes, of course!

Status `~/.julia/environments/v1.0/Project.toml`
  [6e4b80f9] BenchmarkTools v0.4.1
  [7073ff75] IJulia v1.15.2
  [b964fa9f] LaTeXStrings v1.0.3
  [429524aa] Optim v0.18.1
  [58dd65bb] Plotly v0.2.0
  [91a5bcdd] Plots v0.22.5
  [438e738f] PyCall v1.18.5
  [d330b81b] PyPlot v2.7.0
  [92b13dbe] TaylorIntegration v0.4.1+ [`~/.julia/dev/TaylorIntegration`]
  [6aa5eb33] TaylorSeries v0.9.2

from taylorintegration.jl.

I can't recognize the commit you are in in TaylorIntegration, [92b13dbe]. Can you confirm that commit or where did it diverge from master?

from taylorintegration.jl.

I think [92b13dbe] refers to TaylorIntegration's UUID; but still, yes, it looks as if the local used version of TaylorIntegration is not exactly v0.4.1... Since the output of status looks like you have dev'd TaylorIntegration, have you tried doing git checkout v0.4.1 in the dir ~/.julia/dev/TaylorIntegration and then checking if the problem disappears?

from taylorintegration.jl.

You are right, @PerezHz, it does refers to the uuid.

from taylorintegration.jl.

Oh, thank you very much.

When I tried git checkout v0.4.1, I got error: pathspec 'v0.4.1' did not match any file(s) known to git..
So, I pulled from ~/.julia/dev/TaylorIntegration and then I tried doing git checkout v0.4.1 and it worked.

However, I don´t know if the pull was the best option.

from taylorintegration.jl.

So, I pulled from ~/.julia/dev/TaylorIntegration and then I tried doing git checkout v0.4.1 and it worked.

So, just to confirm: has the original issue been solved?

However, I don´t know if the pull was the best option.

If you keep TaylorIntegration dev'd then probably you will have to update the versions manually as you just did; but if you don't expect to do many fine-tweakings to TaylorIntegration, then perhaps the best option is to remove TaylorIntegration from your Julia environment and reinstall via using Pkg; Pkg.add("TaylorIntegration") so that you use registered versions and then updating to new releases doing e.g. using Pkg; Pkg.update("TaylorIntegration")

from taylorintegration.jl.

Another way is woth Pkg.free:

(v1.1) pkg> ?free
  free pkg[=uuid] ...

  Free a pinned package pkg, which allows it to be upgraded or downgraded
  again. If the package is checked out (see help develop) then this command
  makes the package no longer being checked out.

from taylorintegration.jl.

So, just to confirm: has the original issue been solved?

Yes, It has been solved.

@lbenet and @PerezHz, thank you for your help.

from taylorintegration.jl.

Then I close this issue.

from taylorintegration.jl.