Feature/crkrenn/multidimensional sampling

[crkrenn]

I've implemented three "NTNF" style sampling interfaces.

Because this is of general interest and can be implemented in maestro alone, I am submitting it here.

Best_candidate sampling is next.

[crkrenn]

This pull request reemphasizes my interest in passing maestro options via the yaml file. Then, we can type maestro run spec instead of maestro run spec --pgen {{one_of_my_pgen_files_I_can_not_remember_which_one}} --autoyes

Pretty please?

AUTOYES: True
PGEN: ./pgen_sample.py

[FrankD412]

I'm wondering if these don't belong in samples -- I'm thinking they might either belong in documentation or in a folder on their own. @jwhite242 is writing some pgen documentation, might it belong there?

[FrankD412]

This pull request reemphasizes my interest in passing maestro options via the yaml file. Then, we can type maestro run spec instead of maestro run spec --pgen {{one_of_my_pgen_files_I_can_not_remember_which_one}} --autoyes

Pretty please?

AUTOYES: True
PGEN: ./pgen_sample.py

So, I've been considering a different but it doesn't feel right to place those settings into the specification itself. They are not required to actually run the study, and are optional. I've been tossing around the idea of using a smaller spec that could list specs with their command line options. A sort of maestro play feature. You could then spin off multiple studies.

[crkrenn]

@FrankD412 & @jsemler, what do you think of a maestro_config.yaml file like the one that Brian D. has been talking about in his work with Jessica?

[crkrenn]

@FrankD412 (cc @jsemler), I want to help you keep your code base clean and maintainable. I'd also like a place to hang my sampling code, and associating it with maestro seems to be the right choice because it only depends on maestro (and some public third-party packages).

Could a third-party directory or LLNL/maestro-examples repo be appropriate?

See the following links for precedent:

https://github.com/gramps-project
https://www.gramps-project.org/wiki/index.php/Addons_development

[FrankD412]

@crkrenn -- I've been meaning to collaborate with users on an example directory. These might be a good start there, I also think they might be good here for advanced documentation. I'm feeling 50/50 on it -- @jwhite242 @jsemler, thoughts?

[crkrenn]

Frank, I don’t think my examples add much to the understanding of writing pgen.py scripts. Your monte-carlo example does a good job there. But, I do think that my examples are something that teams at LLNL would build on, and perhaps elsewhere as well. V/r,

…

-Chris Chris Krenn Lawrence Livermore National Laboratory [email protected]<mailto:[email protected]> 925-423-8636 -------------------------------------------- Q: Why is this email [hopefully] five sentences or less? A: http://five.sentenc.es From: Francesco Di Natale <[email protected]> Reply-To: LLNL/maestrowf <[email protected]> Date: Monday, June 1, 2020 at 6:15 PM To: LLNL/maestrowf <[email protected]> Cc: Ckl <[email protected]>, Mention <[email protected]> Subject: Re: [LLNL/maestrowf] Feature/crkrenn/multidimensional sampling (#271) @crkrenn<https://github.com/crkrenn> -- I've been meaning to collaborate with users on an example directory. These might be a good start there, I also think they might be good here for advanced documentation. I'm feeling 50/50 on it -- @jwhite242<https://github.com/jwhite242> @jsemler<https://github.com/jsemler>, thoughts? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#271 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABOJ5V4ZKIMKTWLT4TWZPILRURHCTANCNFSM4NEQZCXQ>.

[FrankD412]

@ckrenn -- Makes sense. I will look into what's needed to start a new repository of examples and get back to you. In the mean time we can keep this PR open to keep track. Thanks!

[FrankD412]

@crkrenn -- I've started an example repository. Please submit a PR to the collection of examples here.

[FrankD412]

@crkrenn -- do you mind if I close this PR and the other sampling PR since they're now up on the sheetmusic repository?