minimal-generator

Quick start without Atom

1. Download Julia from https://julialang.org/downloads/ and double click to install.

2. Download Gurobi and obtain a free academic license at https://www.gurobi.com/academia/academic-program-and-licenses/.

3. Open a Julia REPL, and enter

   julia> ENV["GUROBI_HOME"] = "/Library/gurobi912/mac64"
   

   This tells Julia where to find the Gurobi installation.

4. Import the minimal-generator library files by running the following commands in the julia REPL:

   using Pkg
   cd("path/to/source/folder")
   include("installRequirements.jl")
   

5. Follow the example in exampleRun.jl.  For example, start by importing a point cloud, computing homology, and plotting a barcode:

   # import a data file:
   pc = readdlm("data/Synthetic-data/Gamma/Gamma-pointcloud/2x100-Gamma-4.csv")
   # compute homology of pc in dimension 1
   C = computeHomology(pc, false, 1)
   # plot the bar code for a pointcloud
   plotBarCode(C)
   

Quick start with Atom

Requirements:

1. Download Julia from https://julialang.org/downloads/ and double click to install.
2. Download Atom from https://atom.io/.
3. Open Atom, go to preference -> packages, install package: uber-juno. (configure the right )
4. Download Gurobi and obtain a free academic license at https://www.gurobi.com/academia/academic-program-and-licenses/. Open a Julia cones ENV["GUROBI_HOME"] = "/Library/gurobi912/mac64"
5. Open the folder containing all the scripts in Atom.
6. Open the file named installRequirements.jl, run this file to install required packages and load all the other scripts into the environment. I had to run Pkg.build("HDF5") to compile Eirene.

To compute homology and optimize generators for a given pointcloud, look at an example pipeline in the file exampleRun.jl.

Some output functions:

• point cloud --> C, where C is some object with "all the information you need" about the computation, we used a customized object called "homologyObject" to represent C.
• (C,m) --> the euclidean coordinates of the mth point in the point cloud (assuming the input is a point cloud)
• (C,d) --> dth boundary matrix
• (C,d) --> dth barcode
• (C,d,n) --> birth time of nth cell in dimension d
• (C,d,n) --> vertices of nth cell in dimension d
• (C,d,k) --> row index of the kth pivot element for the boundary matrix in dim d
• (C,d,k) --> col index of the kth pivot element for the boundary matrix in dim d
• (C,d,l) --> generator for lth persistent homology class (expressed as a linear combination of simplices; concretely, this would probably take the form of a 1-column sparse matrix)