Summary

This PR addresses issue #172 by clarifying the cryptographic security properties of sentence generation methods.

Changes

• Documentation Updates: Added comprehensive documentation to all sentence generation methods (, , , ) explaining that cryptographic security is the caller's responsibility
• Security Guidance: Provided clear guidance on when to use secure RNGs vs deterministic RNGs for different use cases
• API Preservation: Maintained the existing API while adding clarity about security implications

Key Points Clarified

1. Caller Responsibility: Made it explicit that cryptographic security depends on the RNG implementation provided by the caller
2. RNG Choice Guidance: Directed users to use secure RNGs like for cryptographic applications
3. Deterministic Output: Informed users about using seedable RNGs with fixed seeds for testing/reproducibility
4. Future Considerations: Acknowledged the potential migration to mentioned in the original issue

Impact

This change resolves the ambiguity about whether the library intentionally supports cryptographically secure generation by making it clear that:

• The library provides the infrastructure for random sentence generation
• The security properties depend entirely on the RNG choice made by the caller
• Users have full control over the cryptographic properties through their RNG selection

Closes #172