Request: Add link to community BISAC classification tools

## Summary

I've created a set of tools that classify the PG-19 corpus using BISAC subject headings and would appreciate a link from the README to help other researchers discover them.

## The Tools

**Repository:** https://github.com/fredzannarbor/pg19-enhancements

### Features

- **BISAC Classification**: Dual-method (keyword + semantic embedding) classification for all 28,044 books into standard publishing categories
- **Gutenberg ID Mapping**: Maps PG-19 internal IDs to canonical Gutenberg numeric IDs for linking to Project Gutenberg
- **Pre-computed Data**: Ready-to-use JSON files with full classification results

### Example Use Cases

- Filter PG-19 by genre (find all fiction, all history, all science, etc.)
- Link PG-19 books directly to Project Gutenberg for full text
- Analyze genre distribution in the corpus
- Create genre-specific subsets for targeted language modeling

### Classification Statistics

| Category | Books | % |
|----------|-------|---|
| FICTION | 8,247 | 29.4% |
| HISTORY | 4,891 | 17.4% |
| BIOGRAPHY | 2,103 | 7.5% |
| RELIGION | 1,847 | 6.6% |
| SCIENCE | 1,523 | 5.4% |

## Suggested README Addition

```markdown
### Community Tools

- [pg19-enhancements](https://github.com/fredzannarbor/pg19-enhancements) - BISAC subject classification and Gutenberg ID mapping for all 28,044 books
```

## Why This Helps

Many researchers using PG-19 need to:
1. Filter by genre/subject for domain-specific experiments
2. Link back to original sources on Project Gutenberg
3. Understand the composition of the corpus

These tools address all three needs without modifying the core dataset.

Thank you for maintaining this valuable benchmark!