Add autoloop perf-comparison program for tsb vs pandas benchmarking

[Copilot]

New autoloop program that iteratively benchmarks tsb (TypeScript/Bun) against pandas (Python), one function per iteration, with results displayed on the playground doc site.

Autoloop program

• .autoloop/programs/perf-comparison/program.md — runs every 6h, metric is benchmarked_functions (higher is better). Each iteration picks an unbenchmarked function, writes matching TS + Python benchmarks, runs both, records results.

Benchmark infrastructure

• benchmarks/tsb/ and benchmarks/pandas/ — paired benchmark scripts with identical datasets (100K elements, 5 warm-up + 50 measured iterations)
• benchmarks/run_benchmarks.sh — orchestrator that discovers pairs, runs both, computes ratios, writes results.json
• Seed benchmark: bench_series_creation for both runtimes

Playground page

• playground/benchmarks.html — bar chart + detail table, dynamically loads results.json
• playground/index.html — added Performance section linking to benchmarks
• .github/workflows/pages.yml — copies benchmark results into deployed site

[mrjf]

@copilot fix ci

[Copilot]

@copilot fix ci

Fixed in 05c8606. Two issues:

1. Unused Label import in tests/stats/rank.test.ts (pre-existing noUnusedImports lint error)
2. benchmarks/results.json formatting didn't match biome's expected format

Lint, typecheck, and all 1130 tests pass now.

[github-actions]

---

Warning

The push_to_pull_request_branch operation failed: Failed to apply patch. The code changes were not applied.

Evergreen Fix: Merge Conflict Resolved

I resolved the merge conflict in .github/workflows/pages.yml by merging main into this branch.

Conflict: Both this PR and main modified the same section of pages.yml. The PR added a benchmark results copy step; main (via #93) added Python setup and validation steps.

Resolution: Both step groups are included in sequence:

1. Copy benchmark results to playground (from this PR)
2. Set up Python 3.12 + install pandas/numpy (from main)
3. Validate Python playground examples (from main)

Generated by Evergreen — PR Health Keeper · ● 1.4M · ◷

[mrjf]

@copilot resolve the merge conflicts in this pull request

[Copilot]

@copilot resolve the merge conflicts in this pull request

Resolved in 6ad6a7a. The only conflict was in .github/workflows/pages.yml — main added Python validation steps at the same location where our branch added the benchmark results copy step. Kept both: benchmark copy runs first, then Python setup + validation.