[Shreyavemana]

This PR enhances the decorator utilities in nitrostack/core/additional_decorators.py to make caching and rate limiting safer and more predictable.

Changes

• @cache
  • Adds max_size parameter (default 1024, None disables cap).
  • Implements expired-first then LRU eviction to prevent unbounded growth.
  • Exposes cache_clear() for tests.
• @rate_limit
  • Introduces RateLimitExceededError (subclass of ValueError) carrying retry_after_seconds.
  • Supports optional key= callable for per-user or per-resource counters.
  • Error messages now include a “Retry after X seconds” hint.
• Core
  • Added RateLimitExceededError to nitrostack/core/errors.py and exported in __init__.py.
• Tests
  • New tests/test_decorators_limits.py covers eviction, expiry, unbounded mode, cache_clear, typed error, retry_after_seconds, per-key limits, and shared default behavior.

Verification

• New tests plus existing decorator/cache tests pass (30 total).
• Full suite passes (aside from one unrelated httpx dependency).

Impact

• Prevents memory leaks from unbounded caches.
• Provides clearer, typed errors for rate limiting with retry hints.
• Enables per-user rate limiting strategies.

[Shreyavemana]

The solution adds an LRU‑bounded cache with a default size of 1024 and a cache_clear() hook, plus a typed RateLimitExceededError that includes retry hints and supports per‑key counters. All decorator tests passed (30 total).

[manish-wekan]

Install the entire nitrostack-py package from output-schema-keep-null branch, either locally or in /tmp/<name_of_dir>/
then import that package
and test it against #42

#42
If the tests pass, add the test report as comments. here.

[Shreyavemana]

Verification of the proposed fix (feat/cache-eviction-rate-limit, bfe29e8)

Changes: Improved caching with size limits and a cache-clearing option, along with better rate limiting and per-key limits.

Installation and test environment

Cloned the branch into a temporary directory, created a fresh virtual environment, and installed the package using pip install <repo>.

The checks were run against the installed package from outside the repository, and the import resolved to venv\Lib\site-packages\nitrostack, confirming that the installed package was being tested rather than the source checkout.

Results:

• 30 tests passed against the installed package, covering decorator limits, caching, and MCP extensions.
• 822 tests passed in the source checkout.
• End-to-end verification results:
  -- Installed branch: 9/9 checks passed, covering LRU eviction, cache size limits, expiration, cache_clear(), and rate-limit behavior.
  -- Upstream main (a06624c): 0/9 checks passed, as the tested features are not supported.

Conclusion: The results support the proposed changes to caching and rate limiting, with all reported tests passing in their respective environments.

[manish-wekan]

LGTM

[manish-wekan]

Rebase and resolve the conflicts in additional_decorators.py.
Reject max=0, which currently crashes with IndexError.

~sent from claude-code