ENH: let a convergence study be reproduced from one seed

[thc1006]

simulate_convergence had no way to take a seed, so each batch derived a root of its own and the study could not be repeated. Addresses #1077.

Pull request type

• Code changes (bugfix, features)

Checklist

• Tests for the changes have been added (if needed)
• Docs have been reviewed and added / updated
• Lint (ruff check / ruff format --check, pylint) has passed locally
• All tests have passed locally
• CHANGELOG.md: one line, since the workflow that would have written it has not run since CI: run the changelog job for pull requests from forks #1112 (BUG: the changelog workflow stopped running, and CHANGELOG.md is 37 merged pull requests behind #1173).

Current behavior

simulate_convergence runs batches until the confidence interval settles, calling simulate(append=True) for each one. It takes no seed and passes none on, so every batch draws fresh entropy. Two consequences: the same study cannot be run twice, and the same max_simulations reaches a different set of samples depending on the batch_size that divided it up, which is the opposite of what a batch size should mean.

New behavior

It takes a keyword-only random_seed and hands it to each batch. The seed fixes the study rather than the batch: every batch continues the root the rows already record, so a given max_simulations reaches the same samples whichever batch_size was used to get there. Leaving it out keeps today's behavior, which draws fresh entropy and reproduces nothing.

The eight lines that do this are a pass-through. What makes them mean something is #1187, which merged last week: an append now continues the root its rows were drawn with rather than starting a second one, and a simulation takes its seed from its own index rather than from the worker or the batch that happened to run it. Before that, passing the same seed to every batch would have restarted the sequence each time.

Breaking change

• No

random_seed defaults to None, which is what every existing call already does.

Additional information

Three tests. One runs the same study twice and compares what was drawn; one runs the same max_simulations under two different batch sizes and compares those; the third is the control, that a study with no seed still runs. Accepting the seed and not passing it on turns the first two red and leaves the control green.

This began as a branch from August that carried thirteen commits. Eleven of them have since merged, as #1181 and #1187, so only the two here are new. I took those onto current develop rather than rebasing the branch, since the diff against develop was dominated by what the old base was missing rather than by anything this adds.

Verified on a50d07d5:

pytest tests/                      2647 passed, 10 failed
  the same on develop a50d07d5     2644 passed, 10 failed
ruff check . / ruff format --check clean
pylint rocketpy/ tests/ docs/      10.00/10, exit 0

The ten are the same ten on both, all of them optional dependencies this machine does not have.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 91.58%. Comparing base (a50d07d) to head (41344c7).

Additional details and impacted files

@@           Coverage Diff            @@
##           develop    #1198   +/-   ##
========================================
  Coverage    91.58%   91.58%           
========================================
  Files          132      132           
  Lines        18404    18404           
========================================
+ Hits         16855    16856    +1     
+ Misses        1549     1548    -1     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.