{namespace}`¶

Searches within a single named slot rather than against full row vectors. For example, searching for "Einstein" only in the "subject" slot of Triple-encoded data, ignoring other slots entirely.

Request¶

Content-Type: application/json

URL Parameters:

Parameter	Description
`db_name`	Name of the database
`namespace`	Namespace within the database

Body:

Parameter	Type	Required	Default	Description
`query`	string	✅	—	The value to search for within the slot
`slot`	string	✅	—	The slot to search within (e.g. `"subject"`, `"object"`)
`encoding`	string	❌	`null`	Encoding type: `"semantic"`, `"exact"`, or `"numeric"`. Auto-detected from schema if omitted
`top_k`	int	❌	`10`	Number of results to return
`schema_json_data`	string	❌	`null`	JSON string of the schema. Required if schema is not stored in metadata

Behavior¶

The query is encoded using the slot's encoding strategy, then compared only against the target slot's vector component in each row. This isolates the search to a specific structural position in the data.

If encoding is "semantic", the query embedding is loaded into the cache before the Rust search call.

A schema is required — either stored in metadata from ingest, or provided via schema_json_data.

Responses¶

200 OK¶

{
  "status": "success",
  "query_type": "slot_search",
  "query": "Einstein",
  "slot": "subject",
  "count": 3,
  "results": [
    {
      "_id": 14,
      "_score": 0.961,
      "data": {
        "subject": "Einstein",
        "predicate": "developed",
        "object": "Relativity"
      }
    }
  ]
}

Field	Description
`query`	The search query
`slot`	The slot that was searched
`count`	Number of results returned
`results`	Array of matches with `_id`, `_score`, and `data`

Error Responses¶

Status	Condition
`400`	No schema found and none provided, or invalid query
`404`	Collection not found
`500`	Unexpected internal error

Notes¶

Slot search is more precise than full-row search when you know which structural position a value occupies.
For searching across multiple slots simultaneously with different weights, use /compose/search_slots.
Schema must be available — either stored from ingest or passed in schema_json_data.

Example¶

import requests

SERVER_URL = "http://18.220.128.24:8000"
API_KEY    = "yourapitoken"

def search_slot(db_name: str, namespace: str, query: str, slot: str, top_k: int = 10) -> dict:
    response = requests.post(
        f"{SERVER_URL}/compose/search_slot/{db_name}/{namespace}",
        headers={"X-API-Key": API_KEY},
        json={
            "query": query,
            "slot":  slot,
            "top_k": top_k,
        },
    )
    response.raise_for_status()
    return response.json()


result = search_slot(
    db_name="my_db",
    namespace="triples",
    query="Einstein",
    slot="subject",
    top_k=5,
)
print(result)

Expected output:

{
  "status": "success",
  "query_type": "slot_search",
  "query": "Einstein",
  "slot": "subject",
  "count": 2,
  "results": [
    { "_id": 14, "_score": 0.961, "data": { "subject": "Einstein", "predicate": "developed", "object": "Relativity" } },
    { "_id": 17, "_score": 0.903, "data": { "subject": "Einstein", "predicate": "born_in",   "object": "Ulm" } }
  ]
}

POST /compose/search_slot/{db_name}/{namespace}¶