SRC Brokering

SRC brokering for JAE Intro Scholarship

How to run a function with Docker

Note: This is a general ReadMe.md. You should use one of the folders. Name a folder . In each folder, there is an example with that folder name. This is a generalitation.

Some functions

I have used Docker Desktop aplication. The steps for a succesful execution are:

docker build -t image1-fn <folder_version>

In this step, we create an image called image1-fn with a created dockerfile, and we specify the directory in wich we have the programme and the dockerfile.

docker run --name contenedor2-fn -v $(pwd)/<folder_version>/images:/<folder_version>/images -e FSOURCE='images' image1-fn

We create a conainer called contenedor2-fn, and path for mounting the images directory is specified. Consequently, result will be in that directory. We write the image we want in te container too.

Finally, we have executed the programme, which gets the image perro_husky.jpg in images, gets the face borders and save the result in images.

This is a simple version of the program. If you want more complexity, please use one of the next links:.

example-fn-logging: With this version, you could see hot the program runs with 'docker logs '.

example-fn-optimised: In this version, Dockerfile use a low weight image

Work

In this branch, you will find documentation and code about a system with only a node. If you want to see documentation and code about a general algorithm based in MOEA, please, go to branch Multinode-System. Documentation about WSClean and a bit of DNNs are here too.

Now it will be explained every folder in this branch.

• /docs: Here you will find three folders: OSCAR, broker and wsclean.

• /docs/OSCAR: There is a Readme.md where you can read all about OSCAR instalation: what is OSCAR, ways to install it,...

• /docs/wsclean: There is a Readme.md where it is summarized WSClean library from the point of view of a mathematical and/or computer scientist. All about arguments that could be interesting, what is WSClean,...

• /docs/broker: Broker is a set of algorithms that leads all about system control. You will find two folders: DNN/Transformer and MOEA.

• /docs/broker/DNN/Transformer: Information about transformers for an interpretation problem. It is not very developed.

• /docs/broker/MOEA: All developed about MOEA algorithms for broker is here, and a script that interpolates functions because it was thought to be necessary. You will find two folders. In Platypus folder there are only examples made for training. In Workflow ideas folder you will find ideas about algorithm.

• /docs/broker/MOEA/Workflow ideas: Ideas explained in detail about theorical behaviour. Platypus folder has the environment and script that simulates the algorithm.

• /docs/broker/MOEA/Workflow ideas/Platypus: Environment with docs you need to use Platypus library. Within file moea_problem_broker.py you can find the simulation of the algorithm and how model has been implemented.