Qdrant: Search points (vectors)

qdrant_search_points searches for similar vectors (points) '

Usage

qdrant_search_points(
  conn,
  collection_name = NA_character_,
  vector = list(),
  limit = 5,
  with_payload = TRUE,
  with_vector = FALSE,
  score_threshold = NA_real_
)

Arguments

conn

a connection object created by get_qdrant_connection()

collection_name

the name of the collection you want to use. Defaults to NA.

vector

a vector of embedding values that is used to initiate the search. Defaults to an empty list. Note that this vector is simply a list containing numerical values for your embeddings. If you use an embedding of size 384, this should be a simple list with 384 numerical values in it.

limit

the maximum amount of vectors to return after the search. Defaults to 5.

with_payload

if TRUE, will display the payload of the resulting vectors. Defaults to TRUE.

with_vector

if TRUE, will display the full vector values in the results. Defaults to FALSE.

score_threshold

a threshold for the minimum similarity score to return results. Defaults to NA.

Details

This will run a search for similar vectors on qdrant.

Examples

conn <- get_qdrant_connection()
#> → Connection to Qdrant confirmed
qdrant_create_new_collection(conn, collection_name="test_db",
vectors=list(size=3,distance="Cosine"))
#> → Collection test_db was just created on the qdrant instance.
new_vectors <- list(points=list(
list(id=1, payload=list(text="hi there"),vector=list(0.1,0.5,0.6)),
list(id=2, payload=list(text="well well well"),vector=list(0.6,0.1,0.3))
))
qdrant_upsert_points(conn, points=new_vectors, collection_name="test_db")
#> → Vectors were added in the test_db collection on the qdrant instance.
qdrant_search_points(conn, collection_name="test_db", vector=list(0.2,0.5,0.6))
#> $result
#> $result[[1]]
#> $result[[1]]$id
#> [1] "9b96a1fe-1d54-4cbb-c960-cc6a0286668f"
#> 
#> $result[[1]]$version
#> [1] 0
#> 
#> $result[[1]]$score
#> [1] 0.9924029
#> 
#> $result[[1]]$payload
#> $result[[1]]$payload$text
#> [1] "hi there"
#> 
#> 
#> 
#> $result[[2]]
#> $result[[2]]$id
#> [1] "005ae072-9959-484d-19c5-b26b61200dd0"
#> 
#> $result[[2]]$version
#> [1] 0
#> 
#> $result[[2]]$score
#> [1] 0.6400773
#> 
#> $result[[2]]$payload
#> $result[[2]]$payload$text
#> [1] "well well well"
#> 
#> 
#> 
#> 
#> $status
#> [1] "ok"
#> 
#> $time
#> [1] 0.005961041
#> 
qdrant_delete_collection(conn, "test_db")
#> → Collection test_db was just deleted on the qdrant instance.