Add tokenization benchmarks comparing Links Notation with JSON, YAML, XML

[konard]

Summary

This PR implements UTF-8 character count benchmarks comparing Links Notation (lino) against JSON, YAML, and XML formats. The benchmarks are implemented in all six supported languages and include a GitHub Actions workflow for automatic report generation.

Key Changes

• New benchmark implementations in all 6 supported languages:

  • Rust (primary, used in CI/CD for auto-generating BENCHMARK_RESULTS.md)
  • JavaScript
  • Python
  • C#
  • Go
  • Java

• 5 benchmark test cases covering different data structures:

  • employees - Employee records with nested structure
  • simple_doublets - Simple doublet links (2-tuples)
  • triplets - Triplet relations (3-tuples)
  • nested_structure - Deeply nested company structure
  • config - Application configuration

• GitHub Actions workflow (benchmarks.yml) that:

  • Runs the Rust benchmark on push to main
  • Automatically commits updated BENCHMARK_RESULTS.md if results change
  • Validates all language implementations in CI

Benchmark Results

Format	Total Characters	vs Lino
Lino	734	-
JSON	1332	+81.5%
YAML	920	+25.3%
XML	1882	+156.4%

Average savings with Lino:

• vs JSON: 47.9% fewer characters
• vs YAML: 21.5% fewer characters
• vs XML: 61.5% fewer characters

Files Changed

• .github/workflows/benchmarks.yml - New CI workflow for benchmark automation
• benchmarks/ - New benchmark directory with:
  • BENCHMARK_RESULTS.md - Generated results (auto-updated by CI)
  • README.md - Documentation for running benchmarks
  • data/ - Test data files in all formats
  • Language-specific benchmark implementations
• rust/links-notation-benchmark/ - Rust benchmark crate
• rust/Cargo.toml - Updated workspace members

Test Plan

• All 6 benchmark implementations produce consistent results
• Rust benchmark tests pass (cargo test -p links-notation-benchmark)
• Existing Rust library tests still pass
• GitHub Actions CI passes for all benchmark validations

Issue Reference

Closes #209

---

🤖 Generated with Claude Code

[konard]

🤖 Solution Draft Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

• Public pricing estimate: $8.161039 USD
• Calculated by Anthropic: $6.654238 USD
• Difference: $-1.506801 (-18.46%)
  📎 Log file uploaded as Gist (838KB)
  🔗 View complete solution draft log

---

Now working session is ended, feel free to review and add any feedback on the solution draft.