replace=True?

[AdamOrmondroyd]

Description

After a discussion with @williamjameshandley, replace=True is likely the more appropriate sampling strategy.

Fixes # (issue)

Type of change

Please delete options that are not relevant.

• Breaking change (fix or feature that would cause existing functionality to not work as expected)

How Has This Been Tested?

Please describe the tests that you ran to verify your changes. Provide instructions so we can reproduce. Please also list any relevant details for your test configuration

• Test A
• Test B

Test Configuration:

• Firmware version:
• Hardware:
• Toolchain:
• SDK:

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
• Any dependent changes have been merged and published in downstream modules

[AdamOrmondroyd]

@williamjameshandley any more thoughts on this?

[williamjameshandley]

This should only make a difference for very small numbers of samples. It may cause confusion when e.g. users put in an unterminated nested sampling run and fgivenx returns a single curve (as there will be effectively only one point with significant weight), but it's better to be statistically correct here.

Reactivating the CI (which has been disabled).

This should bump the minor version number, as it might change some users plots, and we should probably use this as an excuse (in a separate PR) to move to supported python version

[AdamOrmondroyd]

if we're messing with the requirements then the matplotlib requirement needs some attention. I've gotten rid of all but one deprecation warning in #26, at the expense of it no longer working with the latest matplotlib compatible with python 3.8.

[AdamOrmondroyd]

This should only make a difference for very small numbers of samples. It may cause confusion when e.g. users put in an unterminated nested sampling run and fgivenx returns a single curve (as there will be effectively only one point with significant weight), but it's better to be statistically correct here.

If they have fewer than downsamples samples, sampling without replacement would fail anyway