Setting up the environment

During this week I’ve set the project repository. I followed these posts ^{1 2} to start the project using current best practices. This includes reusing audreyr’s cookiecutter and pypackage template.

The first impression is that the template contains a bit too many text and config files (running tests, generating documentation, deploying to PyPI), but I believe that getting familliar with them will pay off in the next stages of the project.

The template includes:

• Makefile with commands for building, cleaning, testing, deploying
• Travis-CI automatic testing configuration file
• Makefile for building Sphinx-powered documentation

Which is basically all that’s required to automate project workflow.

Other than that

Meanwhile I am also learning about cross-validation and the Scikit-learn package. I will try to use the existing cross-validation from Scikit-learn to find the regularization parameter for smoothing the data in kCSD. This is useful for determining and reducing the noise from the potential measurements.

My idea is to create hacky and experimental versions as scripts in IPython notebook and then gradually refactor and move new features to the project on the INCF repo.