SpaceML Heliophysics Notebooks: Solar Dynamics Observatory ML Dataset

Heliophysics notebooks corresponding to the SDO ML Dataset

Notebooks:

• 01: SDO ML Dataset (2018)
  • A demonstration of interacting with a sample of the SDO ML dataset. Galvez, R. et al 2019 ApJS.
• 02: Temperature Maps (2018)
  • A deep-learning approach to differential emission measure inversion using a 1x1 Convolutional Neural Network. Wright, P. et al 2019 Zenodo.
• 04: AIA Autocalibration (2019)
  • Demonstration inference of the SDO AIA Autocalibration method: Dos Santos et al 2021 A&A.

The following notebooks are currently under development:

• 03: EVE Virtual Instrument (2018)
  • Under development, based on the work of Szenicer, A. et al 2019 Science Advances.

Interacting with each notebook:

Each notebook is contained within its own folder:

.
└── notebooks
    └── ##_<project>_<year> # Each project has its own folder named sequentially, with the project name, and year of the project
        ├── README.md
        ├── <project>_colab.ipynb # A Jupyter notebook designed to be executed on Google Colab.
        ├── <project>.ipynb # The corresponding local development version of the colab notebook.
        ├── environment.yml # Conda environment file
        └── requirements.txt # Requirements file

For local development, the necessary environment can be created as follows (under the assumption that an anaconda installation exists).

cd notebooks/<project>

conda env create -f environment.yml
conda activate <environment>

# start the jupyter notebook app
jupyter notebook

Contributions

Contributions are welcome as pull requests to the main branch, and should mirror the stucture of existing projects.

• A requirements file can be produced with pip freeze > requirements.txt, however, to minimize the nunmber of redundant packages in that list, first create a virtual environment, and pip install packages there (Anaconda is popular among scientists).

  conda create --name <name>
  conda activate <name>
  conda list #this should be empty
  

• Formatting with Black (https://black.readthedocs.io/en/stable/) is preferred; see https://github.com/drillan/jupyter-black for the Jupyter notebook integration:

  pip install black
  jupyter nbextension install https://github.com/drillan/jupyter-black/archive/master.zip --user
  jupyter nbextension enable jupyter-black-master/jupyter-black