Searching with auto embeddings

[markomilivojevic]

Proposal:

As shown in the documentation, we should be able to search with a query string and find similar KNN documents:

POST /search
{
    "table": "products",
    "knn": {
        "field": "embedding_vector",
        "query": "machine learning",
        "k": 3
    }
}

I wasn't able to build such a query using this library.

Checklist:

^{To be completed by the assignee. Check off tasks that have been completed or are not applicable.}

Details

• Implementation completed
• Tests developed
• Documentation updated
• Documentation reviewed
• OpenAPI YAML updated and issue created to rebuild clients

[Nick-S-2018]

Could you please provide an example of how you try to perform KNN search in the client?
Do you use the same code as here?

[markomilivojevic]

Yes, I use the same code from that link:

$search->knn('some_float_vector_field', 5, 100);

But I should be able to use it like this, to mimic the JSON posted in the first comment:

$search->knn('some_float_vector_field', 'some query string', 100);

[Nick-S-2018]

@markomilivojevic The issue should be fixed in the dev-master version now.

[github-actions]

Thanks for the update! However, the internal checklist is not yet complete. A Manticore team member will close this issue once everything is checked off. 🛠️