BOA Paper Tracking Issue

[madeline-scyphers]

This is a tracking issue to keep track of GH issues that are important for the BOA release paper in some way

100% necessary

• Better documentation for other language support/ScriptWrapper interface #70
• Pin sqlalchemy<2.0  #64
• Add documentation for acquisition function/generation step options and scheduler options from Ax #74
• add doctest #71
• switch 'get_best_trials' in example to 'best_fitted_trials' in examples for clarity #67
• Error restarting trials when using R script wrapper  #66
• Fix typo in "--config-path" in instructions for launching boa  #65
• Add MVP data dump #77
• Release to conda #51
• Publish to PyPi #50
• Move ax-platform to be installed by conda rather than pip  #28
• Setup BOA DOIs for each time a new version releases #78
• Fix windows CI  #18
• create pass through metric #88
• Make docs more user-friendly #125
• Add documentation/examples for how to restart the scheduler #127
• metric names can't be 0 (int type) #126

Stretch Level 1

• Add asynchronous opt interface #73
• change metric config options from "metric", "synthetic_metric", "sklearn_metric", etc. to just metric, with an optional type key #72
• create a best_raw_trial that doesn't use_model_prediction=True at all #68
• Add general plotting function to summarize optimization output #58
• Make getting output and plotting notebook entirely in boa #35 (58 is part of 35)
• Add global stopping criteria to json options  #54
• Document Axs generation step options #52
• Add more examples for the docs, including examples in other languages #79
• Add better checks for configuration file #83
• Handle configuration with dataclasses instead of dictionary #84
• Add jinja2 support for config file(and maybe for user scripts) #128

Stretch Level 2

• Need testing/documentation for using R outside of conda #62