Refactor dry data tracking into a DryData class with centralized printing

[jyejare]

• Replace the shared module-level dry_data dict with a DryData class, instantiated per provider run and passed through providerCleanup and every resource entity instead of being re-imported everywhere.
• Fix DryData.reset_resource() to clear all sub-lists (delete/stop/skip) so stale entries can't leak between regions/zones.
• Centralize the repeated dry-run print/append logic into utils.print_dry_data(), used by all providers (aws, azure, gce, vmware, podman).
• Fix resourcewise_data() to operate on plain dicts, matching how it's actually invoked, and propagate dry_data through entities/providers.py.
• Add tests/test_dry_data.py covering DryData behavior and verifying two regions/zones produce disjoint, non-repeating dry data with no resources missing from the aggregated output.
• Add Docs/DRY_DATA_REFACTOR.md documenting the refactor.