US-4.2implemented Run the pipeline with randomised/scrambled data for demos
As a user who wants to share screenshots or GIFs of my financial visualisations without revealing real financial data I want run the pipeline with `--randomize` to scramble account names and amounts so that the plots show realistic structure but no private information.
Acceptance Criteria
- `hledger_plot --randomize` scrambles category names and amounts.
- The Sankey and Treemap plots use scrambled data.
- The scrambled output is deterministic for the same seed (reproducible demos).
No DAG diagram: Post-processing flag — randomisation modifies output in-place; the underlying data path is identical to US-4.1.
Demo video coming soon