search

lanl-ansi / inverse_ising

Julia implementation of RISE, logRISE and RPLE algorithms for the inverse Ising problem
Other
8 stars 3 forks source link

Trouble with input errors #9

Open rtrauscht opened 3 years ago

rtrauscht commented 3 years ago

I have been unable to get Inverse_Ising.jl to work- I consistently get errors like this:

ERROR: LoadError: MethodError: no method matching strip(::CSV.Column{String,PooledString}) Closest candidates are: strip(::AbstractString) at strings/util.jl:217 strip(::AbstractString, ::Any) at strings/util.jl:218

On the stacktrace it shows the failure is at the beginning of the script at argument parsing on the line that reads: method = strip(args[1])

I have tried to faithfully reproduce the use environment: Have noted that the last update was for Julia 1, and am using Julia V 1.1.1 which looks like it was the current version at the time that the update was published. I can't get things to work with the most recent version of Julia either. Have made sure I activated the environment specified by the toml file in the package using CSV v0.5.5 DataFrames v0.18.3 Ipopt v0.5.4 JuMP v0.19.2 StatsBase v0.30.0 LinearAlgebra

Is this type of error familiar at all? Is there another dependency/Julia version that I am missing that could cause such an error?

I am new to Julia , but it seems in my experimentation that CSV processing is somewhat finicky. Perhaps it is an issue of properly formatting the input file? As per recommendation in the docs I have placed the five terms (method, regularization, symmetrization, input file, output filename) in the arguments.csv file (using as input the synthetic example provided in the repository) but the script still always crashes at the argument parsing stage.
I note that the original form of the arguments file provided with the code is not in that same format- could you perhaps add to the repository a copy of an arguments.csv file in that form that is verified to work on your system?

Many thanks in advance for any attention you can give this matter. It's a a very nice paper.

ccoffrin commented 3 years ago

@rtrauscht, this code is not actively maintained and the issue you were having was related to Julia's packages changing their API over the last several years. I have just pushed a fix for the latest version of Julia and the dependent packages that should resolve this issue.

I can recommend you have a look at the package GraphicalModelLearning.jl for a more well maintained software with similar functionality to this one.