EHT Sgr A* Theory Paper Analysis Toolkit

This repository contains tools and Jupyter notebooks for data analysis and visualization for the EHT Sgr A* Theory Paper. All user scripts and notebooks are in the top directory. To use the scripts and notebooks in this package, simply run

pip3 install -r requirements.txt

to obtain the necessary packages.

Dependency

This repository tries to keep its dependency minimal. We only require standard libraries:

• astropy
• click (only used in the top level driver scripts)
• h5py
• jupyterlab (only used to open the notebooks)
• matplotlib
• numpy
• pandas
• parse
• ruamel.yaml (only used in the top level driver scripts)
• tqdm (only used in the top level driver scripts)

Because of this, instead of using custom libraries such as hallmark and mockservation, we put all the dependent functions in the "common/" directory. This makes this repository easy to use and self-contained.

When we are about to finish the Sgr A* papers, it is possible that we will push the code upstream back to the individual packages, and turn this into a meta-package that help us manage all upstream packages.

Design

We design this toolkit so it allows multiple people to work together. To make it a smooth experience, we define three data stage.

• models/: the "input models" of this toolkit include images at different frequencies or full SEDs. These are usually outputs of high performance C or Fortran codes such as ipole and igrmonty that require significant computing resource to run. The number of files is of order of million and total size may reach 10s of TB. The intention is that the people who generate the models are responsible to their own models, who may copy or link the input models in the models/ directory in their own clones of this repository.

• cache/: this toolkit then preprocesses the input models to summary tables and/or compressed movies that are directly relevant to the science. Examples include total flux, image size, spectral filtered movies. The number and total size of these cache files is usually a few orders of mangitude smaller than the files in models/. Because of the more managable size, the intension is that people can share/rsync the full copy of these cache files to their own laptops to perform further analysis.

• output/: stores outputs including LaTeX tables, plots, and movies, that are generated from files in cache/ and possibility models/. The intention is these output files should be identical no matter who generate them. These files should be at publication qualitity that we can place them directly to the Sgr A* Theory Paper.

Other Tools

• SYMBA Pegasus Workflow: https://github.com/bhpire/symba-osg

• ipole Condor Workflow: https://github.com/bhpire/ipole-osg

• igrmonty Condor Workflow: https://github.com/bhpire/igrmonty-osg

• calsz Condor Workflow: https://github.com/bhpire/calsz-osg

Usage

• First, ensure that the files you are analyzing are in the model folder. This is likely efficiently accomplished via symlink.

• Next, you need to make a .yaml file for the library you are analyzing. These are meant to flexibly handle a variety of file formats and naming schemes. Pattern matching one of the existing files is recommended. You are specifying the source file name format, output file name format, image format, and then the order in which you would like the rows in the tables to be output.

• After that, typically do "python cache_summ.py {path to your IPOLE yaml file}" for tables of image observables, or "python cache_sed.py {path to your GRMONTY yaml file}" to do the same for SEDs. Your output will be formatted however you specified in the .yaml files.