Exercise 06: add PRM and RRT* planning tasks

[angelostefanini0]

Changes

• Removed the SAT tasks and related exercise material from collision_primitives.py, data.py, and ex06.py, restructuring Exercise 06 into seven tasks.
• Updated docs/06-collisionchecking.md to document the new exercise structure, PRM and RRT*, and the expected evaluation behavior.
• Added the student-facing PRM and RRT* interfaces in sampling_planners.py.
• Added deterministic public planning scenarios in sampling_data.py, covering direct-path and obstacle cases.
• Added visualizations in visualization.py for planning scenarios, PRM samples, obstacles, queries, and returned paths.
• Fixed Tasks 4 and 5 by extending Task 4 sensing to cover the complete robot motion and restoring the random seed after CVXPY validation so generated cases remain reproducible and independent.
• Added public evaluation for Tasks 6 and 7 in ex06.py. PRM paths are checked for endpoints, sample membership, bounds, collision-free edges, connection radius, and shortest-path cost; RRT* paths are checked for endpoints, bounds, collision-free edges, deterministic behavior on validation scenarios, and cost relative to the reference solution.
• Deep-copied planner inputs before validation and execution, and validated implementations on both trivial and obstacle scenarios.
• Aligned the pre-commit Black environment with the project's Python 3.11 requirement.

Validation

• All public reference cases for Tasks 1–7 pass with accuracy 1.0.
• Formatting, compilation, imports, scenario determinism, and visualization smoke tests pass.