fix: preserve Python types in L1-only mode, allow cache_clear() on async#117
Merged
Conversation
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (2)
🚧 Files skipped from review as they are similar to previous changes (2)
📝 WalkthroughWalkthroughL1-only mode now stores raw Python objects via ObjectCache; sync and async wrappers read/write this cache and track keys; invalidate_cache/ainvalidate_cache remove ObjectCache entries; cache_clear() works synchronously for async wrappers in L1-only mode. Tests updated to reflect these behaviors. ChangesL1-only mode with object type preservation
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (1)
tests/unit/test_cache_clear_async.py (1)
98-128: ⚡ Quick winThis test never exercises the async+backend invalidation path.
The class/docstring says this is the recommended path for async functions with a backend, but
@cache(backend=None)only covers the L1-only branch. Please either rename it to match the scenario under test or back it with a backend double so the async backend invalidation path is actually covered.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@tests/unit/test_cache_clear_async.py` around lines 98 - 128, The test TestAsyncInvalidateCacheStillWorks currently uses `@cache`(backend=None) so it only exercises the L1-only path; either update the test to reflect that by renaming the class/docstring to indicate L1-only, or change the decorator to supply a backend test double (e.g., a simple async-capable backend mock) so async_func and its method ainvalidate_cache actually exercise the async+backend invalidation path; modify the test to instantiate and pass that backend into cache(...) and keep calling async_func.ainvalidate_cache(5) to validate backend invalidation behavior.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@src/cachekit/decorators/wrapper.py`:
- Around line 587-593: The wrapper currently coerces ttl to 300 on cache writes,
violating create_cache_wrapper()'s contract that ttl=None means no expiration;
update the write sites (in the cache wrapper where _object_cache.put(cache_key,
result, ttl=ttl or 300) and the similar block around lines 912-917) to pass ttl
through unchanged (or translate None to the explicit sentinel expected by
ObjectCache.put) instead of using "ttl or 300", ensuring that ttl=None remains
unbounded expiry when calling ObjectCache.put with cache_key, result.
- Around line 579-585: In the L1-only fast-return path inside the wrapper (when
_l1_only_mode is true and _object_cache.get(cache_key) yields found), call
features.clear_correlation_id() before resetting function stats and returning
the cached value so the correlation context is not preserved across requests;
i.e., insert a call to features.clear_correlation_id() just prior to
reset_current_function_stats(token) / return cached_value in the branch that
currently invokes _stats.record_l1_hit().
---
Nitpick comments:
In `@tests/unit/test_cache_clear_async.py`:
- Around line 98-128: The test TestAsyncInvalidateCacheStillWorks currently uses
`@cache`(backend=None) so it only exercises the L1-only path; either update the
test to reflect that by renaming the class/docstring to indicate L1-only, or
change the decorator to supply a backend test double (e.g., a simple
async-capable backend mock) so async_func and its method ainvalidate_cache
actually exercise the async+backend invalidation path; modify the test to
instantiate and pass that backend into cache(...) and keep calling
async_func.ainvalidate_cache(5) to validate backend invalidation behavior.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: a865f557-8877-4d5b-bc91-de7d3ef68c8c
📒 Files selected for processing (2)
src/cachekit/decorators/wrapper.pytests/unit/test_cache_clear_async.py
Codecov Report❌ Patch coverage is
📢 Thoughts on this report? Let us know! |
… async L1-only mode (`backend=None`) now stores raw Python objects via ObjectCache instead of serializing through MessagePack. This fixes: - #73: tuples, sets, frozensets no longer degrade to lists on cache hit - #76: cache_clear() works synchronously on async functions in L1-only mode (no backend I/O needed, so no reason to force async invalidation) The type inconsistency was particularly insidious: first call (miss) returned the original type, second call (hit) returned the deserialized type. Now both return the same object. Closes #73 Closes #76
27Bslash6
force-pushed
the
fix/l1-only-type-preservation
branch
from
7b8ab0e to
ee843fa
Compare
- Fix ttl=None coercion: use 1-year sentinel (31536000s) instead of hardcoded 300s, preserving the contract that ttl=None means no expiry - Clear correlation ID on L1 cache hit (both sync and async paths) to prevent context leaking across requests - Rename TestAsyncInvalidateCacheStillWorks to reflect it only tests L1-only path
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
@cache(backend=None)now stores raw Python objects (viaObjectCache) instead of serializing through MessagePack. Tuples, sets, frozensets are preserved on cache hit.cache_clear()works synchronously on async functions in L1-only mode (no backend I/O needed).Root Cause
The L1-only code path serialized results into MessagePack bytes on cache miss but returned the original object. On cache hit, it deserialized — returning a different type (tuples → lists, sets → lists). This made
type(fn()) != type(fn())across calls.Design
@cache(backend=None)now usesObjectCache(same as@cache.local()) for storage. This means:lru_cache)Users who need serialization consistency with L2 (for upgrade path testing) can use
@cache(backend=None, serializer='standard')— but the default behavior should match user expectations.Test plan
cache_clear()works on async L1-only functionscache_clear()still raises TypeError for async+backend (intentional)Closes #73
Closes #76
Summary by CodeRabbit
Bug Fixes
Improvements
Tests