Various packages no longer available via conda

[ktyle]

Hi there! As inspired by @fperez , as part of a Geoscience Data Analysis & Visualization class that I am teaching (in its first incarnation this semester) at UAlbany, I thought it would be cool to have my students (and I) once again attempt to reproduce Laken and Stordal 2016, as was done in @fperez 's class @ UC-Berkeley in 2017! One issue that I'm immediately running into is that a few of the pinned packages in environment.yml are no longer available from conda or conda-forge:

seaborn-0.6
matplotlib-1.5.0
python-3.4.4

It looks like the HBGWL_analysis notebook can mostly be reproduced using seaborn-0.8, mpl-2.0.2, and Python-3.5.4. But an interesting question arises in terms of achieving long-term reproducibility when the original environment can no longer be readily reproduced ...

--Kevin

[benlaken]

Hi Kevin, thank’s very much for your interest in reproducing this work. I am happy to know that people are still finding use out of it, and I am very much in favor of using it as a learning exercise for students. I must admit that I haven’t looked at this notebook in a while or attempted to run it, but I will make some time to update it and ensure compatibility with currently supported libs so that getting it running is not an impossible task. I will try and get the Master branch running when I have some free time in the next week or so.

[ktyle]

Hey @benlaken that's super! I can say that all the figures / tables in the paper can be reproduced using Python 3.5.4 , MPL 2.0.2, Seaborn 0.8, with the exception of Table 1 (no analog in the notebook) and Figure 3 (the notebook-generated figure has frequency as the x-axis and is not broken up by season). In my present-day environment (output from watermark below), errors pretty quickly crop up, the first being due to deprecated Pandas Dataframe.resample arguments.

---

CPython 3.7.8
IPython 7.18.1

numpy 1.19.1
scipy 1.5.2
pandas 1.1.2
statsmodels 0.12.0
patsy 0.5.1
matplotlib 3.3.2
seaborn 0.11.0
tabulate 0.8.7