CLN, CI: check notebooks follow black formatting

[MarcoGorelli]

Suggested by @AlexAndorra here, opening this as a placeholder so I don't forget.

Any preferences on black formatting, or are the defaults OK? Will you be looking to applying it to the rest of the codebase as well in the future?

to-do list (updated 2020-10-12)

• docs/source/notebooks/GLM-negative-binomial-regression.ipynb
• docs/source/notebooks/SMC2_gaussians.ipynb
• docs/source/notebooks/bayes_param_survival_pymc3.ipynb
• docs/source/notebooks/bayesian_neural_network_advi.ipynb
• docs/source/notebooks/blackbox_external_likelihood.ipynb
• docs/source/notebooks/data_container.ipynb
• docs/source/notebooks/empirical-approx-overview.ipynb
• docs/source/notebooks/api_quickstart.ipynb
• docs/source/notebooks/convolutional_vae_keras_advi.ipynb
• docs/source/notebooks/dependent_density_regression.ipynb
• docs/source/notebooks/gaussian_mixture_model.ipynb
• docs/source/notebooks/ODE_with_manual_gradients.ipynb
• docs/source/notebooks/factor_analysis.ipynb
• docs/source/notebooks/conditional-autoregressive-model.ipynb
• docs/source/notebooks/hierarchical_partial_pooling.ipynb
• docs/source/notebooks/MLDA_gravity_surveying.ipynb
• docs/source/notebooks/lasso_block_update.ipynb
• docs/source/notebooks/GP-MeansAndCovs.ipynb
• docs/source/notebooks/gaussian_process.ipynb
• docs/source/notebooks/howto_debugging.ipynb
• docs/source/notebooks/marginalized_gaussian_mixture_model.ipynb
• docs/source/notebooks/model_comparison.ipynb
• docs/source/notebooks/model_averaging.ipynb
• docs/source/notebooks/dp_mix.ipynb
• docs/source/notebooks/gaussian-mixture-model-advi.ipynb
• docs/source/notebooks/GP-MaunaLoa2.ipynb
• docs/source/notebooks/log-gaussian-cox-process.ipynb
• docs/source/notebooks/sampler-stats.ipynb
• docs/source/notebooks/getting_started.ipynb
• docs/source/notebooks/sampling_callback.ipynb
• docs/source/notebooks/sampling_compound_step.ipynb
• docs/source/notebooks/lda-advi-aevb.ipynb
• docs/source/notebooks/posterior_predictive.ipynb
• docs/source/notebooks/stochastic_volatility.ipynb
• docs/source/notebooks/weibull_aft.ipynb
• docs/source/notebooks/updating_priors.ipynb
• docs/source/notebooks/normalizing_flows_overview.ipynb
• docs/source/notebooks/rugby_analytics.ipynb
• docs/source/notebooks/variational_api_quickstart.ipynb
• docs/source/notebooks/probabilistic_matrix_factorization.ipynb
• docs/source/notebooks/survival_analysis.ipynb
• docs/source/notebooks/putting_workflow.ipynb
• docs/source/notebooks/multilevel_modeling.ipynb
• docs/source/notebooks/cox_model.ipynb

---

This is a begginer-friendly issue, all you need to do is:

pip install -U nbqa
nbqa black notebook1.ipynb notebook2.ipynb

replacing notebook1.ipynb notebook2.ipynb with some notebooks from above. You should then check over the diff to make sure nothing unexpected has been changed - please report any errors/issues with nbqa to https://github.com/nbQA-dev/nbQA . Feel free to ping me if you want/need help with this

NOTE: sometimes, you may need to disable black for a block of code - see the style guide for how to do that

[AlexAndorra]

Thanks for the initiative! No real preference per se, we just set the line length to 100 instead of the default 88 in pyproject.toml, as we did on ArviZ.
I don't know about applying it to the rest of the codebase, as reviewing these changes is quite hard. We tend to do it gradually, on a file by file basis but I'm not aware of a concerted effort to do it on all the codebase.

[MarcoGorelli]

@AlexAndorra do you want to add the beginner-friendly tag here?

[AlexAndorra]

Done @MarcoGorelli 😉

[satrio-hw]

Hi, I'd like to help here.. So, just to make sure.. I need to reformat given notebook using nbqa black, isn't it?

[MarcoGorelli]

Hi, I'd like to help here

Thanks!

.. So, just to make sure.. I need to reformat given notebook using nbqa black, isn't it?

Yup - but because these notebooks are user-facing, you may sometimes need/want to use "# fmt: off and "# fmt: on - see docs/source/notebooks/BEST.ipynb from https://github.com/pymc-devs/pymc3/pull/4148/files for an example

[satrio-hw]

noted, I'll try to working on this files first:

• docs/source/notebooks/AR.ipynb
• docs/source/notebooks/Bayes_factor.ipynb
• docs/source/notebooks/BEST.ipynb
• docs/source/notebooks/GLM-hierarchical-advi-minibatch.ipynb
• docs/source/notebooks/GLM-linear.ipynb
• docs/source/notebooks/Euler-Maruyama_and_SDEs.ipynb
• docs/source/notebooks/GLM-hierarchical.ipynb
• docs/source/notebooks/GLM.ipynb
• docs/source/notebooks/DEMetropolisZ_EfficiencyComparison.ipynb
• docs/source/notebooks/DEMetropolisZ_tune_drop_fraction.ipynb

[MarcoGorelli]

@satrio-hw sorry, I didn't get round to checking these off yet, I've taken on many of them in #4148

I'll check them off now, feel free to take any of the rest

[satrio-hw]

understood, than I'll take :

• docs/source/notebooks/GLM.ipynb
• docs/source/notebooks/GLM-robust.ipynb
• docs/source/notebooks/GLM-rolling-regression.ipynb
• docs/source/notebooks/GP-Marginal.ipynb
• docs/source/notebooks/GP-Kron.ipynb
• docs/source/notebooks/GP-smoothing.ipynb
• docs/source/notebooks/GP-TProcess.ipynb
• docs/source/notebooks/GP-SparseApprox.ipynb
• docs/source/notebooks/MLDA_simple_linear_regression.ipynb
• docs/source/notebooks/GP-Latent.ipynb