refactor: delete stale files and consolidate .mindmodel structure

Deleted stale root-level Python files:
- main.py (unused 'Hello world' script)
- verify.py (unused table info script)
- scraper.py (unused MotionScraper class)
- scheduler.py (unused DataUpdateScheduler class)

Deleted duplicate .mindmodel root YAML files (subdirectory versions are more comprehensive):
- anti-patterns.yaml, architecture.yaml, conventions.yaml
- dependencies.yaml, domain.yaml, domain-glossary.yaml
- stack.yaml, tech-stack.yaml, workflows.yaml

Added comprehensive .mindmodel subdirectories:
- constraints/ (naming, db-schema, error-handling, types, etc.)
- patterns/ (api, architecture, database, python, streamlit, etc.)
- examples/ (code examples for each pattern)
- anti-patterns/, architecture/, conventions/, dependencies/, domain/, stack/

Updated ARCHITECTURE.md to reflect current codebase:
- Removed references to non-existent files
- Added missing files (explorer.py, explorer_helpers.py, pipeline/)
- Added directory structure documentation
- Updated tech stack to include scipy, sklearn, umap

Updated .gitignore:
- Added patterns for generated analysis files
- Added .worktrees/ pattern (was already in gitignore but dir was deleted)

Removed empty .worktrees/ directory

.mindmodel/stack/stack.yaml

@@ -0,0 +1,41 @@
+# Tech Stack
+
+## Runtime & Language
+- **Python ≥3.13** (type: runtime)
+- Streamlit (type: web framework) - multi-page app: Home, Stemwijzer, Explorer (4 tabs)
+
+## Data Layer
+- **DuckDB** (type: database) - 9 tables: motions, mp_votes, svd_vectors, mp_party_history, etc.
+- **ibis** (type: ORM) - DuckDB backend for Pythonic SQL
+- Query mode: duckdb:// path or :memory: (see database.py:50-51)
+
+## ML / Analytics
+- **scikit-learn** (type: ML) - clustering, Procrustes alignment
+- **UMAP** (type: dimensionality reduction) - 2D political compass projection
+- **scipy** (type: scientific computing) - spatial/alignment algorithms
+- **numpy** (type: numerical computing) - array operations
+
+## Visualization
+- **Plotly** (type: charting) - dual-layer interactive charts (scatter + annotations)
+
+## Key Source Files
+| File | Purpose |
+|------|---------|
+| `database.py` | MotionDatabase singleton, DuckDB connection, 9-table schema |
+| `explorer.py` | Explorer page with 4 tabs (Motion, MP, Party, Evolution) |
+| `explorer_helpers.py` | Pure helper functions, Plotly chart builders, coordinate computation |
+| `analysis/` | SVD pipeline, UMAP projection, clustering algorithms |
+| `pipeline/` | Data fetch → transform → store pipeline |
+| `pages/1_🗳️_Stemwijzer.py` | Quiz page (thin wrapper) |
+| `pages/2_🔍_Explorer.py` | Explorer page (thin wrapper) |
+| `config.py` | Dataclass Config pattern |
+
+## Database Tables
+- `motions` - parliamentary motions with id, title, date, category
+- `mp_votes` - individual MP votes on motions (1/0/-1)
+- `svd_vectors` - SVD-computed political positions (entity_id, window, vector_2d)
+- `mp_party_history` - MP-to-party mappings over time
+- `party_centroids` - aggregated party positions
+- `windows` - time period definitions
+- `mp_trajectories` - MP position changes across windows
+- Plus 2 additional tables (exact names vary)