My awesome project using TeNPy

Initial setup to get started

This is a small template repository to simplify setting up projects that use TeNPy. To get started:

• Visit https://github.com/tenpy/project_template/ and click on the big green "Use this template" button to create a custom fork of the repository. You can keep that fork private, if you want to!
• git clone your new fork repository onto your computer/on the cluster. After the git clone, call git submodule init and git submodule update to download TeNPy.
• Setup your Python environment, e.g. module load python/3.8 on a cluster or conda activate custom_env to activate a conda environment. You might find it usefull to set up a custom module based on the provided module_example. Adjust the files in cluster_templates/*.sh to set correct paths/load the modules etc, as needed.
• Once you have the correct Python environment loaded, cd TeNPy/ and bash ./compile.sh to compile TeNPy.

Implement custom models and other classes

There's a model_custom.py which serves as an example how to implement a new model class; you can add further classes (structured in other python modules, if desired), as much as you like!

Submitting a job / running simulations

• Activate your python environment.
• Create a subfolder where you want to have the output files and cd there.
• Create a yaml file with the simulation parameters, for example the example_submit.yml. (You can also have multiple yaml files, which then get "merged" together; this allows to have a "default_parameters.yml" file which is always the same, and a "changing_parameters.yml" which contains what you changed compared to previous runs...)
• Call cluster_jobs.py submit example_submit.yml. This will create a job script (e.g. MyJob.slurm.sh) and job config (MyJob.config.yml) containing all the parameters which you just submitted. Finally, it will submit that job to the cluster (if you use the SlurmJob or SGEJob for job_config.class in the yaml file) or directly run everything (if you use JobConfig instead). The examples are set up such that you only see warnings/errors in the terminal output (or MyJob.job*.out on a cluster), but you get the lengthy log files with the same name as the output files (also containing error messages/warnings).
• You can immediately update the example_submit.yml after job submission and submit slightly different parameters as a separate job, if desired.

Profit!!!

Do some awesome science and write a great paper about it :)