Enhance/refactor workflow-generated boilerplates

[tsalo]

Summary

The command-line interfaces in the workflows module generate methods descriptions that can be pasted directly into manuscripts. Currently, those boilerplates are just multiline strings. I'd like to improve them a bit.

• BibTeX references instead of just strings
• Updated descriptions. I think they're a little out-of-date at the moment.
• Write out the text to files as well. I think they are currently just used in the logger.

Also, are there any tools out there that would help make better boilerplates?

[tsalo]

As part of this, the conperm workflow boilerplate is out of date.

[tsalo]

What if we implemented methods within individual Estimators that built descriptions from the Estimator parameters? Alternately, we could generate the boilerplates and add them as an attribute to MetaResult objects, if we're only interested in Estimators and Correctors.

For object-based descriptions:

meta = ALE()
corrector = FWECorrector(method="montecarlo")

result = meta.fit(dataset)
boilerplate = meta.generate_description()
corr_result = corrector.transform(result)
boilerplate2 = corrector.generate_description()

For result-based descriptions:

meta = ALE()
corrector = FWECorrector(method="montecarlo")

result = meta.fit(dataset)
boilerplate = result.description_
corr_result = corrector.transform(result)
boilerplate2 = corr_result.description_