Let’s assume that you have installed conda or miniconda or mamba. We are going to use that to create an environment for this package. We will add jupyter and fitsio using conda because that seems to work better, and add this package using pip, because that works better.
conda create -n nb_to_pack -c conda-forge -y python=3.10.5 jupyter fitsio
conda activate nb_to_pack
git clone https://github.com/KIPAC/NBToPackage.git
cd NBToPackage
pip install -e .
You can start up a jupyter notebook session (pointed at the directory with the notebooks) like so:
jupyter-notebok nb &
From there can launch either notebook. Note that we saved a version of the old notebook that was run, while we saved a ‘clean’ version of the new-notebook, to make it clearer what the new code looks like.
Ok, let’s try make the plots from the command line tool we build to replace the notebook. First let’s see how to run the tool.
bin/piff-syst-plots -h
usage: piff_syst_plots [-h] -d FILE [-o DIR] [-c FILE] [-f] [-a] [-l]
[NAMES ...]
positional arguments:
NAMES Names of plots to draw, leave blank for all plots
options:
-h, --help show this help message and exit
-d FILE, --data FILE Path to data
-o DIR, --output DIR Output directory
-c FILE, --config FILE Yaml configuration file
-p , --plot_type Plot file type: ['pdf', 'png', 'eps', 'jpg']
-a, --annotate Store provenance metdata with pdf
-l, --list_plotters list plotters
Ok, let’s make all the plots using the data/piff_sample_cat.fits file provided.
bin/piff_syst_plots -d data/piff_sample_cat.fits -o plots -p pdf