Codes

SARAH is a mathematica code that allows any renormalisable theory to be studied with state-of-the-art numerical tools. The user supplies the bare minimum ingredients to specify the theory, and the code computes the Lagrangian, Feynman rules, RGEs, mass matrices, and can write a bespoke spectrum generator for the model that can compute the spectrum at loop level, precision corrections, decays, low energy constraints, and many more. It also produces input files for other codes such as MicrOmegas/CalcHEP, a UFO for MadGraph/Whizard, inputs for Vevacious/Vevacious++, etc. 

I have worked on SARAH since 2012; I contacted the original author Florian Staub after completing my paper on two-loop RGEs in Dirac gaugino models, with the aim of being able to study my favourite class of models in that code. Previously I had to compute everything by hand, and yet this tool would allow me to automatically study models with much more precision practically at the push of a button. Initially after Florian implemented the RGEs from my paper and made other modifications specific to the supersoft operators I dug into the code and tested the Fortran output. One of my first contributions was a prototype of the solution of the tadpole equations via Broyden's method. Apparently few other people were interested in looking inside the code, and so was born a fruitful collaboration. 

In 2014 I started collaborating on implementing a two-loop Higgs mass calculation within SARAH. This was a major drawback compared to the state-of-the-art available for other equivalent spectrum generators for the (N)MSSM. Florian wanted to implement a calculation based on computing the effective potential and taking derivatives numerically; this should be simple to compute (there are only a few classes of diagrams and no external fields) but suffered from many drawbacks such as numerical instability (especially for theories with small vevs); lack of masses for pseudoscalars/other fields; no clear way of solving the Goldstone Boson Catastrophe, etc. The initial release included two versions of this approach, described in this paper. At the same time, I started work on deriving the expressions necessary instead for a diagrammatic computation and entirely writing the module within SARAH myself. This was described in 1503.03098 and is now the default option; I upgraded it subsequently to allow pseudoscalar masses to be computed, and a solution to the Goldstone Boson Catastrophe based on work with my student Johannes Braathen.

Other contributions that I have made to the code are:

• One loop decays 

• Unitarity constraints, including colourful fields 

• Many minor updates and bugfixes

In 2019, Florian Staub left physics, and bequeathed responsibility of the code to me and Werner Porod. I have continued work on development and there are some major new features coming soon

BSMArt is a python code to automate the running of parameter scans using many HEP tools. It is supposed to be lightweight, simple and fast enough to run on a laptop yet powerful enough to use on a cluster. Primarily these are ones in the SARAH family, so I provide scripts for downloading SARAH/SPheno, building the SPheno code and compiling it, and setting up the template files for BSMArt for any model. 

The scans included range from very simple (grids, random); practical (MCMC); external libraries (Diver and MultiNest); and finally my Active Learning algorithm, using machine learning to choose new points that will teach a network the most new information.

Writing your own scan or tool is simple and I welcome collaborations or contributions! I am currently adding new scans and tools, especially for collider observables ...

TLDR is a Two Loop Diagram Repository for storing the (renormalised) scalar self-energies computed by myself and Sebastian Paßehr and described in 1910.02094. 

Other than Mathematica expressions for all of the self energy and tadpole diagrams, both in the full and reduced form, there are also routines for obtaining the finite terms and interfacing with TSIL to compute numerical results. Ultimately the expressions should be incorporated in SARAH; Sebastian and I are working on this!

HackAnalysis is a hackable code for implementing particle level analyses which may have non-standard features, e.g. disappearing tracks. 

I wrote it when working on recasting various LHC searches; initially I was using MadAnalysis, but having problems reproducing certain analysis cutflows where the signal had low efficiencies (so I needed to simulate a lot of events) and I was getting frustrated at how slow it was to create and read Delphes/hepmc files (which I had to repeatedly do). I also  looked at ColliderBit and CheckMATE and saw that they couldn't do what I wanted either: I wanted to be able to break some of the assumptions regarding the event analysis using particle level information, and more honestly, I just wanted to build something so that I could understand what was happening at each step without having to find workarounds or ask other people to add new features. 

When I worked on recasting the CMS disappearing track search this became acute, and I decided to release the analysis along with the framework; it is rather versatile and has multicore operation allowing very fast simulations ideal for prototyping, without needing to write heavy hepmc files (unless you want to); it also includes pileup at particle level, which other codes cannot do (yet). However, I should stress that I don't want to sell another framework, there is no point in having a lot of codes with analyses distributed among them, so I made the syntax similar enough to MadAnalysis (while using a fork of heputils meaning that some of the commands are closer to ColliderBit) so that the analyses can be ported out easily. In fact I have already done this a couple of times ...