availsim4 2.1.1

Availsim4 is a tool to predict reliability and availability of modern particle accelerators and their related systems.

AvailSim4

Availsim 4 is a software to predict reliability and availability of modern particle accelerators and the related technical systems. It is based on a Monte Carlo Discrete Event Simulation algorithm and allows for advanced availability and reliability statistics.

Requirements

Python >=3.7 with external dependencies; the requirements are listed in setup.py file, with oonly the core ones in the requirements.txt file.

Installation

The default installation process does not feature optional dependencies for visualization and optimized sampling options. To make those available, make sure to follow the instriuctions for installation of optional elements listed below.

AvailSim4 can be installed as whole Python project by clonning this repository or as a Python package through a package manager. Below you can find instructions for both possibilities.

PyPI repository

Use the package manager pip to install AvailSim4.

pip install availsim4

To install all optional dependencies, follow the instructions provided in sections below and then use the following command:

pip install 'availsim4[graphs, quasi-mc, test]'

The optional groups of dependencies are graphs, quasi-mc and test:

• graphs enables the GRAPHS exporter to produce graphs of the components. Additional pre-requirements described below.
• quasi-mc installs the QmcPy library for some optimized sampling features. Additional pre-requirements described below.
• test adds dependencies required to run the test suite (not related to standard use of the framework).

You can also select groups of dependencies, for instance [graphs, quasi-mc] to only use a selected subset of the optional dependencies.

Gitlab

To clone and set up the Python project:

• Clone the project from https://gitlab.cern.ch/availsim4/availsim4core and enter the projects directory.
• Optionally, create a dedicated virtual environment (python -m venv env will create an environment named env).
• Trigger an editable installation with setuptools from the local directory using pip install -e '.[quasi-mc, graphs, test]'. You can modify the list in the brackets to select only desired groups of dependencies.
  • Alternatively, use pip install -r requirements.txt to install required optional dependencies according to the needs. You will need to use python availsim4.py instead of just availsim4 to run the framework.

Visualization tools (optional)

To generate a diagram of the architecture tree, AvailSim4 relies on PyGraphviz package and Graphviz tool. For more information and OS-specific installation guide, please refer to the Graphviz webpage: https://graphviz.org/download/ and PyGraphviz: https://pygraphviz.github.io/documentation/stable/install.html.

General instructions:

• Linux: install Graphviz through your system package manager and then make sure to install PyGraphviz with pip or conda : apt-get update, apt-get install libgraphviz-dev, pip install pygraphviz.
• Windows: Graphviz installation wizard can be downloaded manually from the website. Then proceed to install PyGraphviz with pip or conda. Make sure that dot program's plugins are configured by running dot -c command in your environment. May also require Microsoft C++ Build Tools.
• MacOS: you can install Graphviz using Homebrew package manager by running command brew install graphviz and then pip install pygraphviz.

Pre-requirement for Windows users (optional)

The external library used for Quasi-Monte Carlo requires Windows users to install Microsoft Visual C++ 14.0 or greater. It can be downloaded from "Microsoft C++ Build Tools": https://visualstudio.microsoft.com/visual-cpp-build-tools/, this requirement is needed to use the Quasi-Monte-Carlo algorithm. CERN's Windows users might prefer to use SWAN notebooks which are running on Linux machines.

Usage

To run AvailSim4 from a command line, you can use the following:

availsim4 [-h] 
  --system SYSTEM
  --simulation SIMULATION 
  --output_folder OUTPUT_FOLDER
  [--sensitivity_analysis SENSITIVITYANALYSIS] 
  [--HTCondor]
  [--HTCondor_extra_argument OPTIONAL_EXTRA_ARGUMENT]
  [--nb_processes NB_PROCESS]
  [--children_logic PYTHON_FILE]

Example call starting one of the end-to-end tests manually:

python availsim4.py --system availsim4core/test/E2E/input/convergence/convergence_test_system.xlsx --simulation availsim4core/test/E2E/input/convergence/N1000_simulation.xlsx --output_folder output/E2E_example/

You can see the results after the execution finishes in the output/E2E_example directory.

Alternatively, Availsim4 can be used as a module from within Python scripts or notebooks:

import availsim4core.main


availsim4core.main.start(path_simulation = 'path/simulation/file.xlsx',
            path_system = 'path/system/file.xlsx',
            output_folder = 'path/output/folder/',
            path_sensitivity_analysis="",
            HTCondor=False,
            nb_processes=1)

A notebook, notebook_example_availsim4.ipynb, is provided as an example and can be used as a template (e.g. to use on SWAN for CERN users).

Support features for running AvailSim4 on the HTCondor grid are described in the user guide.

Documentation

A user guide is provided in the user guide. It lists all functionalities of the AvailSim4, along with options and relevant descritions.

For developers wishing to contribute to the code, a dedicated developer guide is being created to provide useful explanations of general implementation concepts and gather all development-specific aspects of the project, such as chosen coding conventions and other decisions - to support coherent approach in the long run.

Contributing

Contributions through pull requests to the project are welcome. For changes, please open an issue detailing the matter to discuss the scope.

This project requires contributors to agree to Developer Certificate of Origin (DCO) by adding a dedicated Signed-off-by: line to all commits. Please use git commit –signoff in order to automate this. All merge requests must include a signature in the form: Signed-off-by: Firstname Lastname <[e-mail]>.

License

Copyright © CERN 2021. Released under the GPL 3.0 only license. All rights not expressly granted are reserved.