docs: optimization & data layout report

[nclack]

Adds a docs page summarizing the data-loading optimization work: the pipeline
stages and how they work, the storage/read levers and the measurements behind
them, the corpus-wide finding that most datasets are limited by decompression
(not reading) due to chunk layout, the dataset-format categories, and a
damacy-vs-tensorstore comparison.

• New page docs/optimization.md, linked from the mkdocs nav and the index page.
• Cluster-specific absolute paths scrubbed; references point at the in-repo bench
  harness (bench/run.py, bench/report.py, bench/tensorstore_bench.py).
• Numbers are from real benchmark runs (L40, fast networked storage).

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 57.71%. Comparing base (cca3b09) to head (c307cfa).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #157      +/-   ##
==========================================
- Coverage   57.78%   57.71%   -0.08%     
==========================================
  Files          64       64              
  Lines       10046    10055       +9     
  Branches     1748     1750       +2     
==========================================
- Hits         5805     5803       -2     
- Misses       3494     3501       +7     
- Partials      747      751       +4     

Flag	Coverage Δ
unittests	57.71% <ø> (-0.08%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.
see 4 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[nclack]

Moved to docs/ in nclack/damacy-slides-2026-06; keeping these out of this repo.