docs: fix type preservation claims in comparison.md

[27Bslash6]

Summary

Update comparison.md to reflect that L1-only mode (backend=None) now preserves Python types after #73/#117.

Three claims corrected:

• Feature comparison note: tuples/sets preserved in L1-only, only converted with distributed backends
• Single-process section: removed incorrect "tradeoff" note about serialization
• Validation evidence: updated to reflect L1-only stores raw objects

Closes #74

Summary by CodeRabbit

• Documentation
  • Clarified type-preservation differences between single-process (L1-only) and distributed cache modes.
  • Emphasized that single-process mode stores raw Python objects and avoids serialization overhead.
  • Noted distributed backends perform serialization that can convert tuples/frozensets to lists by default.
  • Added guidance on configuring serialization to preserve Python types when using distributed backends.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 74df2f72-1d75-45db-a3dd-497b6f8a67dd

📥 Commits

Reviewing files that changed from the base of the PR and between 7d56dc2 and e6f28b7.

📒 Files selected for processing (1)

• docs/comparison.md

✅ Files skipped from review due to trivial changes (1)

• docs/comparison.md

---

📝 Walkthrough

Walkthrough

This PR updates docs/comparison.md to correct claims about type preservation: L1-only (backend=None) preserves raw Python types without serialization, while distributed backends use MessagePack (StandardSerializer) which converts tuples/frozensets to lists by default unless serializer='auto' is configured. Three related sections are updated.

Changes

Type Preservation Documentation

Layer / File(s)

Summary

Type preservation clarification across three sections docs/comparison.md

Updated the "Type preservation" guidance to distinguish L1-only (backend=None) raw-object storage from distributed-backend serialization (MessagePack StandardSerializer converting tuples/frozensets to lists), replaced the single-process "tradeoff" text with a "Type safe" note about L1-only, and adjusted the validation evidence bullet to state L1-only preserves Python types while distributed backends serialize unless serializer='auto' is used.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

Possibly related PRs

• cachekit-io/cachekit-py#117: Implements the code-level change that preserves raw Python types in L1-only mode while distributed backends serialize, aligning docs with behavior.

Poem

🐰 I hopped through lines of docs so neat,

Swapped a claim for clarity, tidy and sweet.
L1 keeps objects whole, no serialize art—
Distant backends pack and convert apart.
A small carrot dance for docfix complete.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	The description covers the key changes and links to the related issue, but is missing several required sections from the template (Motivation, Type of Change, Security/Documentation Checklists, Testing, Backward Compatibility).	Complete the PR description by adding required sections from the template: motivation for the changes, type of change checkbox, documentation validation checklist items, and testing/compatibility sections.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately and concisely summarizes the main change: fixing incorrect type preservation claims in comparison.md documentation.
Linked Issues check	✅ Passed	The PR properly addresses issue #74's three main objectives: correcting the type preservation claims in the feature comparison note, removing the incorrect serialization tradeoff statement, and updating validation evidence.
Out of Scope Changes check	✅ Passed	All changes in comparison.md are directly scoped to fixing the type preservation claims identified in issue #74; no out-of-scope modifications detected.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch docs/fix-comparison-type-claims

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ All tests successful. No failed tests found.

📢 Thoughts on this report? Let us know!