Fix jina-code model alias

[Whamp]

Fixes #1025.

Summary

• Point jina-code at jinaai/jina-embeddings-v2-base-code, which exists on Hugging Face.
• Remove the stale README note that this model requires an HF token.
• Add a regression test for the alias.

Verification

• npm test -- tests/search/embedding-strategy.test.ts
• npm run typecheck
• npm run lint (passes with existing warnings)

[github-actions]

Thank you for your submission, we really appreciate it. Like many open-source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution. You can sign the CLA by just posting a Pull Request Comment same as the below format.

---

I have read the CLA Document and I hereby sign the CLA

---

Test seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You can retrigger this bot by commenting recheck in this Pull Request. Posted by the CLA Assistant Lite bot.}

[greptile-apps]

Greptile Summary

This PR fixes a broken model alias by pointing jina-code at the official jinaai/jina-embeddings-v2-base-code namespace on Hugging Face instead of the stale Xenova/ mirror, and removes the outdated HF-token requirement note from the README. A targeted regression test is added to guard against future regressions.

Confidence Score: 4/5

Safe to merge; the fix is minimal and well-tested with no logic changes beyond the model alias string.

Only P2-level observation (sibling jina aliases still use Xenova/ prefix), no blocking issues. Change is a one-liner with a regression test and a matching README cleanup.

src/domain/search/models.ts — verify jina-small and jina-base Xenova mirrors are still functional if they aren't updated.

Important Files Changed

Filename	Overview
src/domain/search/models.ts	Single-line alias fix: Xenova/jina-embeddings-v2-base-code → jinaai/jina-embeddings-v2-base-code. Note that sibling aliases jina-small and jina-base still use Xenova/ prefixes; if those mirrors are also stale, a follow-up may be needed.
tests/search/embedding-strategy.test.ts	Adds a regression test that asserts MODELS['jina-code'].name equals the new canonical path. Straightforward and correct.
README.md	Removes the "(requires HF token)" qualifier from the jina-code row, consistent with the model now being publicly accessible.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A["User selects model: 'jina-code'"] --> B["getModelConfig('jina-code')"]
    B --> C["MODELS['jina-code'].name"]
    C -->|"Before PR"| D["Xenova/jina-embeddings-v2-base-code\n(stale mirror, required HF token)"]
    C -->|"After PR"| E["jinaai/jina-embeddings-v2-base-code\n(official namespace, public)"]
    E --> F["loadModel → @huggingface/transformers pipeline"]
    D --> G["❌ 401 Unauthorized / model not found"]
    F --> H["✅ Embeddings generated"]

Loading

Comments Outside Diff (1)

1. src/domain/search/models.ts, line 30-43 (link)

   Sibling jina aliases still use Xenova/ mirrors

   jina-small and jina-base still point to Xenova/jina-embeddings-v2-small-en and Xenova/jina-embeddings-v2-base-en respectively. If the root cause of this bug was that the Xenova/ mirror for jina-code became stale or gated, it's worth checking whether those mirrors are still publicly accessible — the fix pattern here (jinaai/ namespace) would apply to them too.

   

_{Reviews (1): Last reviewed commit: "fix: point jina-code at published model ..." | Re-trigger Greptile}