GitHub

Compiling UlamMethod.jl Into a Standalone App

Creating The App

This experimental feature allows you to compile UlamMethod.jl into a standalone executable that can be used independently of Julia. Currently, only 2D data are supported. Note that creating the app requires a julia installation.

  1. Create a folder where you want the app to be stored and cd into it. For this example, we'll use UlamApp.

  2. Clone https://github.com/70Gage70/UlamMethod.jl into UlamApp. If you use GitHub CLI this is accomplished quickly by: gh repo clone 70Gage70/UlamMethod.jl.

  3. Run the following command from the terminal. This will compile the app, taking 5-10 minutes. It will also add the package PackageCompiler to your main environment. Run julia -e 'import Pkg; Pkg.rm("PackageCompiler") afterwards if you don't want it any more.

julia -e 'import Pkg; Pkg.add("PackageCompiler"); Pkg.develop(path = "UlamMethod.jl"); using PackageCompiler, UlamMethod; create_app("UlamMethod.jl/", "UlamMethodCompiled")'

The path to the executable is now .../UlamApp/UlamMethodCompiled/bin/UlamMethod. This can e.g. be added to your PATH for convenient access.

Using The App

The expected syntax is ./UlamMethod ARG1 ARG2 ... ARG9, where each ARG is defined as follows

  • ARG1: The path to the file containing the input data. This expects a .mat file that contains the variables "x0", "xT", "y0" and "yT".
  • ARG2: The file to write the output. This should also be a .mat file.
  • ARG3: The type of binner. Options are [rec, tri, hex, vor].
  • ARG4: The number of bins.
  • ARG5 - ARG8: These args control the fraction of data taken to be in nirvana on each side of the rectangular boundary in the order left, right, bottom, top.
  • ARG9: The reinjection algorithm. Options are [data, stat, unif].

Example

The following terminal command will generate some test data in the file traj-test.mat:

julia -e 'import Pkg; Pkg.activate(temp = true); Pkg.add(["UlamMethod", "MAT"]); using UlamMethod, MAT; traj = Trajectories(2, 1000); matwrite("traj-test.mat", Dict("x0" => traj.x0[1,:], "y0" => traj.x0[2,:], "xT" => traj.xT[1,:],"yT" => traj.xT[2,:]))'

We'll apply Ulam's method with 100 rectangular bins and "data" reinjection (the default). We'll let 2% of the data be in nirvana on the left side of the boundary. The output file will be ulam-out.mat. Due to precompilation issues, the very first time you run the app, it will be slow.

./UlamMethod traj-test.mat ulam-out.mat rec 100 0.02 0.0 0.0 0.0 data

The output .mat file contains the transition probability matrix "P" as well as the vertices defining the associated bins "bins_verts". Refer to the "info" field for more information regarding the format of the output.