[andrewbaxter439]

What and WHY

This is a potential solution to #329

Tried a quick solution. This goes through the process of:

• Setting bootstrapAll to false by default
• At the start of the multirun main it overrides this to 'true'
• It then looks in config file to see if this is changed/set
• If then parses command line arguments to see if -bootstrap=false is flagged
• If set to false either in config file or command line it flags up "WARNING: Bootstrapping is disabled. Enabling bootstrapping for multiruns is recommented [sic] for analytical runs."

[Copilot]

🟡 Changes recommended

The integration-test configuration inherits the new bootstrap default without corresponding updates to its existing statistical baselines.

1 open issue

• Preserve integration test baseline after bootstrap default change · New

What changed in this PR

This PR addresses #329 by making regression-coefficient bootstrapping configurable and enabling it by default for multiruns to account for parameter uncertainty.

Changes:

• Makes bootstrapAll mutable while retaining its default of false outside multirun startup.
• Enables bootstrapping for multiruns, with YAML and command-line overrides.
• Adds a warning when multirun bootstrapping is disabled.

File	Description
src/​main/​java/​simpaths/​experiment/​SimPathsMultiRun.java	Sets the multirun default, parses the bootstrap option, and warns when disabled.
src/​main/​java/​simpaths/​data/​Parameters.java	Makes bootstrapAll configurable rather than constant.

_{Review effort: Balanced}

[andrewbaxter439]

This seems a good suggestion! bootstrapping should be disabled in integration tests as a good practice - this should be fixed before merging.

[Copilot]

Added parameter_args.bootstrapAll: false to config/test_run.yml in commit abdee07 to preserve the integration-test baselines.