traceAI-qdrant 0.1.0

OpenTelemetry instrumentation for Qdrant vector database

traceAI-qdrant

OpenTelemetry instrumentation for Qdrant vector database.

Installation

pip install traceAI-qdrant

Quick Start

from fi_instrumentation import register
from traceai_qdrant import QdrantInstrumentor
from qdrant_client import QdrantClient
from qdrant_client.models import VectorParams, Distance, PointStruct

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

client = QdrantClient(":memory:")
client.create_collection("documents", vectors_config=VectorParams(size=384, distance=Distance.COSINE))

client.upsert("documents", points=[
    PointStruct(id=1, vector=[0.1]*384, payload={"text": "Hello"}),
    PointStruct(id=2, vector=[0.2]*384, payload={"text": "World"}),
])

results = client.search("documents", query_vector=[0.15]*384, limit=5)

Instrumented Operations

Operation	Span Name	Description
search	qdrant search	Vector similarity search
upsert	qdrant upsert	Insert or update points
delete	qdrant delete	Delete points
retrieve	qdrant retrieve	Get points by ID
scroll	qdrant scroll	Iterate through points
recommend	qdrant recommend	Recommendation search
count	qdrant count	Count points

License

Apache License 2.0