Skip to content
Tutorials for the IRIS-HEP tutorial at APS DPF, July 29, 2019.
Jupyter Notebook
Branch: master
Clone or download

Latest commit

Fetching latest commit…
Cannot retrieve the latest commit at this time.

Files

Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data
img
.gitignore
01-scientific-python-ecosystem.ipynb
02-histograms.ipynb
03-columnar-data-analysis.ipynb
04-uproot.ipynb
05-accelerating-python.ipynb
LICENSE
README.md
environment.yml
misc-fractal.ipynb

README.md

2019-07-29-dpf-python

Tutorials for the IRIS-HEP tutorial at APS DPF, July 29, 2019, 2‒6pm in Shillman 425, Northeastern University.

How to participate

The preferred way to participate is to run the JupyterLab code with us, altering examples and asking us "what if" questions as we go along, as well as using the notebooks as starting points for the five-minute exercises.

You can run all of these notebooks on a public cloud service called Binder:

Launch Binder

Navigate in the JupyterLab file view (left sidebar) to the desired lesson. Note that Binder cannot save data permanently (reloading your web browser will take you to a new instance), and it may take a minute or two to start up.

Running everything on your own computer

Alternatively, you might want to follow along using your own computer so that you can save data and have copies of the software after the tutorial. If you want a copy of all software packages (a few GB) and you have a non-Windows computer, install conda and git and clone this repository:

git clone https://github.com/jpivarski/2019-07-29-dpf-python.git
cd 2019-07-29-dpf-python

Then use

conda env create -f environment.yml

to install nearly everything in an isolated environment named dpf-python. You can delete this without affecting anything else with the command conda remove --name dpf-python --all.

To use the software, you have to enter the isolated environment. Type

conda activate dpf-python

to enter the environment (and exit it by closing the terminal or conda deactivate). One additional package, JupyterLab, must be installed manually in the isolated environment:

conda install jupyterlab

Start JupyterLab by typing

jupyter lab

A web browser tab should open with the same environment as the Binder site above, or a URL will be provided that you can copy-paste into a web browser. However, the software will be running on your computer.

Cherry-picking software packages or manually installing

Chances are, you won't want to install everything defined by environment.yml, only the packages most relevant to your work. They can be installed individually with pip or conda. There are, however, a few constraints:

  • ROOT can be installed with conda on Macs and UNIX-like computers (e.g. Linux), but not with conda on Windows and not with pip. You can install ROOT manually, but be sure to install the latest version, release 6.18/00 to get the RDataFrame.AsNumpy feature. Also, interaction between ROOT and Python might not work if the compile-time and runtime versions of Python differ. Note that conda installs a custom Python version and pip doesn't.
  • We do very little with root-numpy, but if you install this package, it should be with pip and it should be after ROOT and Python are already installed because it compiles against both of them.
  • The tutorial examples use Python f-strings for formatting; these require at least Python 3.6. None of the exercises have been tested with Python 2, though most of the software described in this tutorial will support Python 2.7 until it reaches its end of life in December 2019.
You can’t perform that action at this time.