Clean up TQDM usage

[WardBrian]

Submission Checklist

• Run unit tests
• Declare copyright holder and open-source license: see below

Summary

This fixes a longstanding issue that was preventing us from using the default format from tqdm. @bob-carpenter recently pointed out how useful this output is, so this restores it.

We can't really test this output, so I was just running the cmdstanpy_tutorial.py file with an artificially high number of iterations.

Copyright and Licensing

Please list the copyright holder for the work you are submitting (this will be you or your assignee, such as a university or company):

Simons Foundation

By submitting this pull request, the copyright holder is agreeing to license the submitted work under the following licenses:

• Code: BSD 3-clause (https://opensource.org/licenses/BSD-3-Clause)

[mitzimorris]

I'm re-running my case-study notebook now. I agree that the time to completion is useful. The iterations/second is misleading in that it's raw iterations, not EFF/sec, and the iterations / sec is about 10x's the EFF/sec.

can we get timing without iterations/sec?

[mitzimorris]

EFF/sec

tqdm for sum_to_zero_vector:

tqdm for soft sum-to-zero constraint:

[WardBrian]

The iterations/second is misleading in that it's raw iterations, not EFF/sec

I don't see how it's misleading at all. The sample interface asks you to specify a number of iterations, not a number of effective draws, and the progress is in terms of the same thing.

the iterations / sec is about 10x's the EFF/sec

This will be entirely model dependent. If your ESS is 10x smaller than your number of samples, then yes, this is expected

[mitzimorris]

point taken re iterations vs EFF.

[mitzimorris]

LGTM!