Rethink2 to pymc v5

[yahrMason]

Rethinking_2 Conversion to PyMC v5+

I am working on updating the Rethinking_2 notebooks to the current version of PyMC + Pytensor. The changes I am making are:

• updating the environment yml and Readme
• changing the causalgraphicalmodels dependency to this updated repo as the original is no longer maintained and is not compatible with current python versions: https://github.com/BirkhoffG/causalgraphicalmodels/tree/master
• replace all theano/aesara references with pytensor
• adopt modern arviz naming conventions (ie idata as opposed to trace)
• use coords to manage model dimensions
• make sure all notebooks run with current versions of PyMC/Pytensor/Arviz

I made an initial pass through the notebooks and changed all aesara/theano refs to pytensor, and changed the names of all inference data objects to the standard idata convention. Now I need to go through the notebooks and execute them and make sure that they run properly. The following list tracks my progress. Anyone feel free to clone my branch and make edits/updates.

To Do:

• [x] Update Readme and yml
• [x] Add missing data sets
• [x] Add notes about coords and idata_kwargs={"log_likelihood": True}
• [x] Chapter 2 Notebook
• [x] Chapter 3 Notebook
• [x] Chapter 4 Notebook
• [x] Chapter 5 Notebook
• [x] Chapter 6 Notebook
• [x] Chapter 7 Notebook
• [x] Chapter 8 Notebook
• [x] Chapter 9 Notebook
• [x] Chapter 10 Notebook
• [x] Chapter 11 Notebook
• [x] Chapter 12 Notebook
• [x] Chapter 13 Notebook
• [x] Chapter 14 Notebook
• [x] Chapter 15 Notebook
• [x] Chapter 16 Notebook

[review-notebook-app[bot]]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[yahrMason]

pre-commit is passing locally. Not sure what's causing the issue.

[twiecki]

Amazing effort!

[twiecki]

The pre-commit looks to be a version problem maybe.

[twiecki]

@yahrMason Can you rebase? should fix pre-commit issues.

[yahrMason]

@twiecki ah yep, that solved it. Thanks for that.

[twiecki]

I think @cluhmann had started on a conversion as well, so maybe he can take a look.

[yahrMason]

This will also close #192 and #225