change the default value of sample in dist_fit()

[jamesmbaazam]

In the current design of dist_fit(), the argument sample is supplied with a default value of NULL, which is checked and replaced with 1000 if it's NULL or < 1000. I think that means it should have a default value of 1000 with a check to see if it's < 1000 when the default is overridden with a supplied value.

In this PR, I change the default value of samples to 1000. If the value of samples is supplied, the function checks if samples is <1000, and if so, sets it to 1000 and throws a warning to inform the user.

[seabbs]

Thanks, James,

See the note about the warning message.

This needs:

• To be rebased on main
• To be redoc'd
• To have a news item
• To have a dev version increment.

Then good to merge.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 9279bf8 is merged into main:

•   :ballot_box_with_check:default: 58.7s -> 58.1s [-14.16%, +12.3%]
•   :ballot_box_with_check:no_delays: 44.4s -> 47s [-6.75%, +18.3%]
•   :ballot_box_with_check:random_walk: 14.8s -> 15s [-7.94%, +11.35%]
•   :ballot_box_with_check:stationary: 37.3s -> 35s [-16.11%, +3.89%]
•   :ballot_box_with_check:uncertain: 54.4s -> 59.9s [-7.82%, +28.05%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[sbfnk]

Last missing task I think is

[ ] To have a dev version increment.

[jamesmbaazam]

I believe this was done here 58db941.

[seabbs]

That isn't in this PR?

[seabbs]

Once things are ready for review FYI I need to be repinged or I won't notice!

[jamesmbaazam]

Last missing task I think is

[ ] To have a dev version increment.

This has been addressed in e91568b.

[seabbs]

Nearly there. Just a small formatting issue.

[seabbs]

LGTM!