Simple Xsuite study

Introduction

This minimal repository aims at turning the initial Dynamic Aperture template, designed to create large, parallelized, parameter scans, into a simpler set of scripts for making simulations of a given accelerator model with Xsuite, under a fixed configuration.

Installation

The script generate.sh should take care of:

1. Cloning the repository for the LHC version of interest (HL 1.6, HL 1.3, run III, run III ions). Note that the runIII configurations are only available on CERN AFS for now.
2. Installing the required packages, along with Python if needed
3. Restructuring the repository to keep only the necessary files
4. Mutate the configuration files to match the desired study

Simply run the following command:

bash generate.sh

And follow the instructions. Don't forget to re-activate the virtual environment using source miniforge/bin/activate every time you open a new terminal and intend to run the Python scripts.

Keep in mind that it remains quite basic, and most likely not very robust. In the event of an issue, please follow the manual installation guide.

Usage

As in the initial template, the simulation is done in two steps, with two separate configuration files:

1. The Python script 1_build_distr_and_collider/1_build_distr_and_collider.py builds a simple collider as a json file from the optics file and madx model specified in the configuration file 1_build_distr_and_collider/config.yaml.
2. The Python script 2_track_particles/2_track_particles.py sets the initial distribution and configuration of the collider, and tracks the particles for a given number of turns, as specified in the configuration file 2_track_particles/config.yaml.

A typical pipeline would therefore be:

1. Modify the configuration files 1_build_distr_and_collider/config.yaml and 2_track_particles/config.yaml to match the desired study.

2. Run the collider building script:

   cd 1_build_distr_and_collider
   python 1_build_distr_and_collider.py

3. When done, run the particle tracking script:

   cd ../2_track_particles
   python 2_track_particles.py

4. Postprocess the results, which are available as a datadrame in the 2_track_particles/output_particles.parquet folder, which you can read with, for instance, pandas, using the following command:

import pandas as pd

df = pd.read_parquet('output_particles.parquet')

License

This repository is licensed under the MIT License.