/PolarizedStacker

Polarized Snapshot Image Hierarchical Stacking

Primary LanguageJulia

PolarizedStacker

Polarized Snapshot Image Hierarchical Stacking

This package will read in a number of eht-imaging snapshot files and stack the posteriors together using an approximate hierarchical modeling scheme. Please see EHTC IV SgrA* for more information.

Installation

To install this package you need to first have a local version of Julia. To install Julia I recommend the juliaup package https://github.com/JuliaLang/juliaup and installing the Julia 1.8 series juliaup add 1.8.5. Once Julia is installed clone the repo and you should be good to go.

Running the script

To run the stacker you need to run

juila -p NCORES main.jl directorylist priors_list.txt

where

  • NCORES is the number of cores you wish to use. Note we parallelize on the list of directories in directorylist so NCORES should be fewer than the number of directories
  • directorylist is a file where each line is a path to the directory of the folder of the snapshot results. For instance if fitting two separate models directorylist would be 
        /path/to/model1/directory
    /path/to/model2/directory
    an example can be found in listdir. We recommend that all runs or data are place in the Data/ folder.
  • priors_list.txt a text file with the list of priors and parameter names used. Two example priors are included (priors_example_mring_m_stokesi_2_m_lp_3_add_floor=False.txt and priors_example_mring_m_stokesi_2_m_lp_3_m_cp_1_add_floor=False.txt). Note that any "distance" variables e.g. diameter and width have to be in the same units as the chain files. For instance usually the chain files are in $\mu$as and so the priors also have to be in $\mu$as.

The output of the stacking will be in directorylist/StackedResults and will contain

  1. stacker_chain.h5 which is a HDF5 file that groups all the snapshot results together
  2. stacker_chain_ha_trunc.csv which is a CSV file containing the chain of the results. Note this is the entire chain I would typically just recommend taking the last quarter of it for parameter inference.
  3. stacker_chain_ckpt.jld2 is the checkpoint file for sampling that can be used to restart the script.

Post-processing

To post-process the images we have included a postprocess script that will convert a stacker result to a set of images. To run this do

julia postprocess.jl "path/to/stacker_chain_ha_trunc.csv" "path/to/output/directory" "path/to/prior_list.txt"

if the results are for a linear polarized images and

julia postprocess.jl "path/to/stacker_chain_ha_trunc.csv" "path/to/output/directory" "path/to/prior_list" -c

for a circularly polarized image. By default this will produce 500 fits images in the output directory, but this can be changed with the -n option.