Managing Python™ packages can be a daunting task. Often, numerous incompatibilities exist between Python™ packages commonly used in Earth sciences (see example about GDAL and Xarray in Jupyter notebook). Commonly required tasks vary depending on the user's need. Here are some capabilities I need for my workflow that are not covered by the standard Anaconda package installation:
- geodetic 2-D and 3-D coordinate transformations
- geodetic computations
- handling commonly used data formats such as GeoData frames
- image processing using sophisticated tools available in dedicated packages such as OpenCV (Open Source Computer Vision Library)
- work specific tools, such as AI-based tools e.g., Segment Anything Model (SAM)
- collaborative code developing and open sharing of code in commonly used repositories such as GitHub
This repository contains a Jupyter notebook with instructions for installing essential Python™ packages for Earth scientists and testing the functionality of the installed packages in various computational and operating system environments. The Jupyter notebook has the advantage that individual cells can be run depending on which packages are installed. The Jupyter notebook and corresponding Python™ script have been tested under the following environments:
- Windows 11 Pro (23H2) Python™ 3.11.7
- Windows 10 Enterprise (21H2) Python™ 3.9.13
- Linux POSIX Release: 5.15.146.1-microsoft-standard-WSL2 Python™ 3.11.7
- Linux POSIX Release: 5.10.198-187.748.amzn2.x86_64 Python™ 3.11.8 (selected packages tested on the CryoCloud JupyterHub)
After publishing this repository several excellent suggestions were made in the discussions on how to avoid the issues I've experienced with package management. Make sure to check them out.