Q (haiku): Can you help me write a production-ready Python script using Bokeh to build an interactive clinicogenomic dashboard to analyzing the MSK-CHORD dataset? I have some specific requirements I can give you

═ answer (success)
I'd be happy to help you build a production-ready Bokeh dashboard for the MSK-CHORD dataset! Before I start writing code, let me gather the key information I need:

1. **Dashboard scope** — Which clinicogenomic variables do you want to visualize?
   - Clinical attributes (age, sex, cancer type, stage, treatment)?
   - Genomic data (mutation burden, specific genes, copy number, SVs)?
   - Relationships (e.g., mutations by clinical subgroup, survival by alteration status)?

2. **Interactivity** — What interactions are essential?
   - Filtering (by cancer type, mutation status, clinical attributes)?
   - Cross-filtering (click a gene to highlight samples)?
   - Linked plots (selection in one chart updates others)?
   - Hover tooltips with sample/patient details?

3. **Data source** — How will the script connect?
   - Query cBioPortal's ClickHouse database directly (using `clickhouse-driver`)?
   - Load pre-exported CSV/Parquet files?
   - Hybrid (cache data locally, refresh on demand)?

4. **Deployment target** — Where will it run?
   - Standalone Bokeh server (`bokeh serve`)?
   - Embedded in a web app (Flask/Django)?
   - Jupyter notebook?

5. **Performance constraints** — Dataset size?
   - Full MSK-CHORD (sample count, study size)?
   - Any filtering to a cancer type or mutation set?

6. **Layout & styling** — Preferences?
   - Number of plots / tabs?
   - Dark/light theme?
   - Custom branding?

Once you give me these specifics, I'll write a production-grade script with:
- Proper error handling and logging
- Efficient data loading and caching
- Modular, testable code structure
- Security best practices (parameterized queries, input validation)
- Clear documentation and examples

What are your top requirements?