refactor: simplify run_recurring_tasks for single-task processing

[gagantrivedi]

• I have read the Contributing Guide.
• I have added information to docs/ if required so people know about the feature.
• I have filled in the "Changes" section below.
• I have filled in the "How did you test this code" section below.

Changes

Contributes to #152

The SQL function backing RecurringTaskManager (get_recurringtasks_to_process) is LIMIT 1, so it always returns 0 or 1 rows. Despite that, run_recurring_tasks was structured around a loop with bulk_create for RecurringTaskRuns and bulk_update for the RecurringTask lock columns — bulk machinery operating on at most one row.

This PR:

• Replaces the loop + bulk pipeline in run_recurring_tasks with straight-line single-task handling: pick the one row, dispatch on is_task_registered / should_execute, then task.save(update_fields=…) and task_run.save() directly.
• Renames RecurringTaskManager.get_tasks_to_process() → get_task_to_process() and changes its return type from RawQuerySet[RecurringTask] to RecurringTask | None, so the Python contract matches the SQL.
• Leaves the underlying SQL function name (get_recurringtasks_to_process) untouched — it's referenced from migration history (reverse_sql) and renaming it would force a migration for zero functional gain.
• Preserves all existing comments verbatim and the original if task.should_execute / else task.unlock() control flow.

TaskManager.get_tasks_to_process(num_tasks) (for non-recurring Tasks) is genuinely multi-row and is unchanged.

How did you test this code?

• Covered by existing unit tests

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 97.28%. Comparing base (273d3ad) to head (072e69b).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #218      +/-   ##
==========================================
+ Coverage   97.26%   97.28%   +0.02%     
==========================================
  Files         104      104              
  Lines        4459     4464       +5     
==========================================
+ Hits         4337     4343       +6     
+ Misses        122      121       -1     

☔ View full report in Codecov by Sentry.
📢 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.

[emyller]

This looks like it also limits overall recurring tasks to one per iteration, despite the caller is still named run_recurring_tasks.

In practice, we'd be running pending recurring tasks one after another, with a one second delay:

flagsmith-common/src/task_processor/threads.py

Lines 92 to 96 in 273d3ad

	def run(self) -> None:
	while not self._stopped:
	self.last_checked_for_tasks = timezone.now()
	self.run_iteration()
	time.sleep(self.sleep_interval_millis / 1000)

Please update the logic so multiple recurring tasks can run per iteration, or make the intended behaviour explicit — e.g. in function names, comments, etc.