fix(debrid): limit torbox magnet checks to 2 per user at any given time

[webstreamr]

this is a simple and stupid workaround to avoid hitting the 5/s limits with the current aiostreams design. it seems to work quite well on my local system and should still leave room for wrapped add-ons doing their own availability checks.

in case this gets merged - something similar will be needed for nzb checks too I suppose. I only have a single indexer and don't experience problems there though.

Summary by CodeRabbit

• Dependencies

  • Added lru-cache dependency to manage cached controls and improve stability.

• Performance Improvements

  • Improved magnet validation by introducing per-account concurrency limits, reducing request spikes and improving overall responsiveness and resource efficiency.

[coderabbitai]

Walkthrough

Adds per-token concurrency limiting for magnet checks in TorboxDebridService by introducing an LRU-backed cache of limiter functions and wiring checkMagnets to run via a per-token limiter.

Changes

Cohort / File(s)	Summary
Repository root package.json	Added lru-cache v11.2.6 to project dependencies.
Core package deps packages/core/package.json	Added lru-cache dependency to the core package.
Concurrency control in Torbox packages/core/src/debrid/torbox.ts	Added private readonly token field, a static checkMagnetsLimiters LRU cache, and changed checkMagnets to retrieve/create a per-token limiter (default capacity 2) and execute stremthru.checkMagnets through it.

Sequence Diagram(s)

sequenceDiagram
  rect rgba(200,200,255,0.5)
    participant Caller
    participant TorboxDebridService
    participant LRUCache
    participant StremThruInterface
  end

  Caller->>TorboxDebridService: checkMagnets(token, magnets)
  TorboxDebridService->>LRUCache: get or create limiter for token
  LRUCache-->>TorboxDebridService: limiter(function)
  TorboxDebridService->>limiter: run(() => stremthru.checkMagnets(magnets))
  limiter->>StremThruInterface: stremthru.checkMagnets(magnets)
  StremThruInterface-->>limiter: results
  limiter-->>TorboxDebridService: results
  TorboxDebridService-->>Caller: results

Loading

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• chore(debrid): only add torrent via URL if it is not cached already #696: Modifies stremthru.checkMagnets internal caching and URL-vs-magnet logic, directly affecting the function call being rate-limited by this PR.

Suggested reviewers

• Viren070

Poem

🐰 I hop from token to token with care,
A tiny LRU keeps limiters fair,
Two checks at a time, no frantic race,
Gentle throttles keep the pace,
Hooray — smooth magnets, tidy place!

🚥 Pre-merge checks | ✅ 4

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately describes the main change: implementing per-user concurrency limits (2 per token) on Torbox magnet checks to avoid rate limiting.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Merge Conflict Detection	✅ Passed	✅ No merge conflicts detected when merging into main

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

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

No actionable comments were generated in the recent review. 🎉

---

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

[coderabbitai]

Actionable comments posted: 1

🤖 Fix all issues with AI agents

In `@package.json`:
- Around line 34-38: The core package is missing a direct dependency on
lru-cache even though packages/core/src/debrid/torbox.ts imports it; add
"lru-cache": "^11.2.6" to the dependencies object in packages/core/package.json
so the core package declares and installs this runtime dependency explicitly,
then run install to update the lockfiles.

🧹 Nitpick comments (1)

packages/core/src/debrid/torbox.ts (1)

154-161: Limiter is process‑local; consider cross‑instance coordination if you scale out.

This cache lives in‑process, so multiple server instances can still exceed 2 concurrent checks per token. If you run more than one Node process, consider a shared semaphore/queue (e.g., Redis) to enforce a global limit.

[webstreamr]

not sure what a good number here is tbh and it's a bit hard to test right now because some systems have intermittent issues. maybe it should just be 1 🤷‍♂️

[webstreamr]

TorBox raised limits to 300/s and ST switched to non-batched post requests. This dirty workaround is not needed for now.