traceAI-redis

OpenTelemetry instrumentation for Redis Vector Search (RediSearch/RedisVL).

Installation

pip install traceAI-redis

Quick Start

Using RedisVL (Recommended)

from fi_instrumentation import register
from traceai_redis import RedisInstrumentor
from redisvl.index import SearchIndex
from redisvl.query import VectorQuery

trace_provider = register(project_name="my-rag-app")
RedisInstrumentor().instrument(tracer_provider=trace_provider)

index = SearchIndex.from_yaml("schema.yaml")

# All operations are traced
query = VectorQuery(
    vector=[0.1, 0.2, ...],
    vector_field_name="embedding",
    return_fields=["content"],
    num_results=5
)
results = index.search(query)

Using redis-py directly

from redis import Redis
from redis.commands.search.query import Query

client = Redis()

# FT.SEARCH with vector queries are traced
query = Query("*=>[KNN 5 @embedding $vec AS score]").dialect(2)
results = client.ft("documents").search(query, query_params={"vec": vector_bytes})

Instrumented Operations

License

Apache License 2.0