[ENH] add chroma hosted embedding function for qwen

[c-gamble]

Adds the Qwen model family hosted on the Chroma embedding service to the known embedding functions in the Python and JS/TS SDKs.

Tested by pointing chromadb to my local OSS dir + running add operations against staging with the real hosted embedding service.

[c-gamble]

• [ENH] add chroma hosted embedding function for qwen #5585 👈 (View in Graphite)
• [CHORE] add the ability to distinguish between queries and docs for js client embedding #5589
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[github-actions]

Reviewer Checklist

Please leverage this checklist to ensure your code review is thorough before approving

Testing, Bugs, Errors, Logs, Documentation

• Can you think of any use case in which the code does not behave as intended? Have they been tested?
• Can you think of any inputs or external events that could break the code? Is user input validated and safe? Have they been tested?
• If appropriate, are there adequate property based tests?
• If appropriate, are there adequate unit tests?
• Should any logging, debugging, tracing information be added or removed?
• Are error messages user-friendly?
• Have all documentation changes needed been made?
• Have all non-obvious changes been commented?

System Compatibility

• Are there any potential impacts on other parts of the system or backward compatibility?
• Does this change intersect with any items on our roadmap, and if so, is there a plan for fitting them together?

Quality

• Is this code of a unexpectedly high quality (Readability, Modularity, Intuitiveness)

[propel-code-bot]

Add Chroma Hosted Qwen Embedding Function to SDKs

This PR introduces the ChromaCloudQwenEmbeddingFunction, which supports the Qwen model family on the Chroma hosted embedding service. The embedding function is integrated in both Python and JS/TS SDKs, featuring its own configuration schema, runtime validation, and documentation. All ecosystem build/test tooling and registration mechanisms are updated accordingly. The implementation ensures environment-variable API key usage, various configuration options (model, task, instructions), and robust test coverage. The SDKs' registry, package definitions, and shared schema utilities now include and recognize this embedding function.

Key Changes

• Introduced ChromaCloudQwenEmbeddingFunction in chromadb/utils/embedding_functions/chroma_cloud_qwen_embedding_function.py (Python) and clients/new-js/packages/ai-embeddings/chroma-cloud-qwen/src/index.ts (TypeScript).
• Added schema file schemas/embedding_functions/chroma-cloud-qwen.json for config validation.
• Updated Python SDK: registration in chromadb/utils/embedding_functions/__init__.py, tests in chromadb/test/ef/test_ef.py, and schema tests.
• Updated JS/TS SDK: new package folder (chroma-cloud-qwen), registration in shared embedders, code, README, TypeScript config, build scripts, and tests.
• Extended all-in-one @chroma-core/all package and pnpm-lock.yaml to include new embedding function.
• Enhanced schema loader/util tooling to cover the new schema and enable runtime type safety.
• Environment variable-only API key (no config fallback). Enforced via runtime checks in both Python and JS/TS.
• Tested round-trip config serialization, error paths, API call wiring, and instructions mapping.

Affected Areas

• chromadb/utils/embedding_functions/__init__.py
• chromadb/utils/embedding_functions/chroma_cloud_qwen_embedding_function.py
• schemas/embedding_functions/chroma-cloud-qwen.json
• clients/new-js/packages/ai-embeddings/chroma-cloud-qwen/ (all new files, including source, tests, schema utils, README, build/test configs)
• clients/new-js/packages/ai-embeddings/common/src/schema-utils.ts
• clients/new-js/packages/ai-embeddings/all/ (package.json, src/index.ts)
• clients/new-js/pnpm-lock.yaml
• Shared tests for config/schema compliance

This summary was automatically generated by @propel-code-bot

[jairad26]

why not add a query config?

[jairad26]

ref jina embedding function for that

[c-gamble]

it's not needed in this case

[jairad26]

there will only ever be 2 tasks, one for docs one for queries? just trying to future proof it

[c-gamble]

no those are targets. tasks are like NL_TO_CODE vs CODE_TO_CODE vs CODE_TO_NL. tasks will presumably be extended in the future

[kylediaz]

In the PDR, it said that CloudEmbeddingFunction should be able to get its API key from the client using it. It doesn't look like your implementation has this feature.

[c-gamble]

In the PDR, it said that CloudEmbeddingFunction should be able to get its API key from the client using it. It doesn't look like your implementation has this feature.

imo this is quite a bit of magic and we don't have precedence for anything similar. also, the API is only available if the user is using the cloud client so we wouldn't really have a way to resolve it for local client users.

i think we can do it as a fast follow if users request it but it seems fine to rely on a single env var for both your CloudClient and EF since technically the client and EF are separate entities and you don't need a client to create an EF.