Data Lakes

Search a data lake

Semantic (vector) search over one lake: the query is embedded with the model the corpus was indexed with and the top top_k passages at or above min_score are returned. Only members of this lake are searched - never the caller's other files. tags narrows to files carrying any of the listed tags. Files still being indexed or paused are withheld rather than searched stale: partial_results is then true and retrieval_unavailable counts them (poll GET /api/v1/data-lakes/{id}/files/{file_id} for a specific file). The query embedding is billed as credits. Limited to 10 searches per minute per caller, shared with the product UI.

post/api/v1/data-lakes/{id}/search

Request

  • Base URL: https://your-deployment.example.com
  • URL: https://your-deployment.example.com/api/v1/data-lakes/{id}/search
  • Auth: one of:
    • HTTP bearer
    • API key in header x-api-key

Path parameters

idstring required

Request body

querystring required
top_kinteger
min_scorenumber
tagsstring[]

Example request

{
  "query": "What is our refund policy for annual plans?",
  "top_k": 5
}

Response

The ranked passages and what, if anything, was withheld.

embedding_modelstring required
partial_resultsboolean required

Changes

No recorded changes to this operation across all 2 revisions of this API.