SEE HOW
FAST FEELS

# Create a vector index using the HNSW algorithm, 768 dimension length, and inner product distance metric
> FT.CREATE idx-videos ON HASH PREFIX 1 video: SCHEMA content_vector VECTOR HNSW 6 TYPE FLOAT32 DIM 768 DISTANCE_METRIC IP content TEXT metadata TEXT

# Add a video vector with metadata
> HSET video:0 content_vector “\xa4q\t=\xc1\xdes\xbdZ$<\xbd\xd5\xc1\x99<b\xf0\xf2<x[…\xf8<” content “SUMMARY:\nThe video discusses the limitations of MySQL at scale and introduces Redis Enterprise” metadata “{\”id\”:\”FQzlq91g7mg\”,\”link\”:\”https://www.youtube.com/watch?v=FQzlq91g7mg\”,\”title\”:\”Redis + MySQL in 60 Seconds\”}”
(integer) 3

# Search for videos using a similar vector and the K-nearest neighbors algorithm
> FT.SEARCH idx-videos “* => [KNN 3 @content_vector $vector AS vector_score]” RETURN 3 metadata content vector_score SORTBY vector_score LIMIT 0 3 PARAMS 2 vector “\b[\xb7;\x81\x12\x9c\xbc\xc6!…\xfe<” DIALECT 2

# Create an index. In this example, all JSON documents with the key prefix user: will be indexed.
rs = r.ft(“idx:users”)
rs.create_index(
schema,
definition=IndexDefinition(
prefix=[“user:”], index_type=IndexType.JSON
)
)

# Use JSON.SET to set each user value at the specified path.
r.json().set(“user:1”, Path.root_path(), user1)
r.json().set(“user:2”, Path.root_path(), user2)
r.json().set(“user:3”, Path.root_path(), user3)

# Find the user Paul and filter the results by age.
res = rs.search(
Query(“Paul @age:[30 40]”)
)

# Result{1 total, docs: [Document {‘id’: ‘user:3’, ‘payload’: None, ‘json’: ‘{“name”:”Paul Zamir”,”email”:”paul.zamir@example.com”,”age”:35,”city”:”Tel Aviv”}’}]}

# b’OK’