Inleiding

Met betrekking tot de RAG-serie: Dit artikel biedt een diepe duik in Qdrant, de vector database gebruikt in:

Deel 4: ONNX & Qdrant-implementatie - Semantische zoektocht naar gebouwen
Deel 5: Hybrid Search & Auto-Indexing - Productie-integratie

Qdrant (uitgesproken "quadrant") is een open-source vector database gebouwd in Rust. Dit artikel heeft betrekking op kernconcepten, de C# client, performance tuning, en productiepatronen.

Wat is Qdrant?

A vectordatabase slaat hoogdimensionale vectoren (inbeddingen) op en maakt een snelle zoektocht naar gelijkenis mogelijk. In tegenstelling tot traditionele databases die exacte overeenkomsten vinden, vindt Qdrant semantisch vergelijkbaar items.

flowchart LR
    A[Text: 'Docker deployment'] --> B[Embedding Model]
    B --> C["Vector: [0.12, -0.34, 0.56, ...]"]
    C --> D[Qdrant]
    E[Query: 'container setup'] --> F[Embedding Model]
    F --> G["Vector: [0.11, -0.32, 0.58, ...]"]
    G --> H[Similarity Search]
    D --> H
    H --> I[Similar Results]

    style B stroke:#6366f1,stroke-width:3px
    style D stroke:#ef4444,stroke-width:3px
    style F stroke:#6366f1,stroke-width:3px
    style H stroke:#10b981,stroke-width:2px

Belangrijkste Qdrant functies:

HNSW-indexering - Sublineaire zoektijden
Filteren - Combineer overeenkomst zoeken met metadata filters
gRPC & REST API's - High-performance toegang
Verdeelde implementatie - horizontaal schalen
Snapshots - Reservekopie en herstel

Kernbegrippen

Verzamelingen

A collectie is als een tabel - het houdt vectoren met een vaste dimensionaliteit en afstand metriek.

flowchart TB
    subgraph Collection["Collection: blog_posts"]
        A[Vector Size: 384]
        B[Distance: Cosine]
        C[HNSW Index]
    end

    subgraph Points
        D[Point 1: slug=docker-intro]
        E[Point 2: slug=kubernetes-basics]
        F[Point N...]
    end

    Collection --> Points

    style A stroke:#6366f1,stroke-width:2px
    style B stroke:#6366f1,stroke-width:2px
    style C stroke:#f59e0b,stroke-width:2px
    style D stroke:#10b981,stroke-width:2px
    style E stroke:#10b981,stroke-width:2px

// Create collection - see https://qdrant.tech/documentation/concepts/collections/#create-a-collection
await client.CreateCollectionAsync(
    collectionName: "blog_posts",
    vectorsConfig: new VectorParams
    {
        Size = 384,              // Must match your embedding model
        Distance = Distance.Cosine  // Best for text embeddings
    }
);

Afstandsmeters (docsunit synonyms for matching user input):

Cosinus - Meet hoek tussen vectoren (beste voor tekst)
Punt - Ruw innerlijk produkt (voor voorgenormaliseerde vectoren)
Euclid - Geometrische afstand (voor ruimtelijke gegevens)

Punten

A punt een enkel record bevat:

flowchart LR
    subgraph Point
        A[ID: uuid/int]
        B["Vector: float[384]"]
        C[Payload: JSON metadata]
    end

    style A stroke:#8b5cf6,stroke-width:2px
    style B stroke:#f59e0b,stroke-width:2px
    style C stroke:#10b981,stroke-width:2px

// Upsert points - see https://qdrant.tech/documentation/concepts/points/#upload-points
var point = new PointStruct
{
    Id = new PointId { Uuid = Guid.NewGuid().ToString() },
    Vectors = embedding,  // float[384]
    Payload =
    {
        ["slug"] = "my-post",
        ["title"] = "Vector Databases",
        ["language"] = "en",
        ["categories"] = new[] { "AI", "Databases" },
        ["published"] = DateTimeOffset.UtcNow.ToUnixTimeSeconds()
    }
};

await client.UpsertAsync("blog_posts", points: new[] { point });

Filteren

Filteren draait voor gelijkenis zoeken - uiterst efficiënt.

flowchart TB
    A[Search Query] --> B{Apply Filters First}
    B --> C[Language = 'en']
    B --> D[Year >= 2024]
    C --> E[Filtered Subset]
    D --> E
    E --> F[Vector Similarity Search]
    F --> G[Ranked Results]

    style B stroke:#ec4899,stroke-width:3px
    style E stroke:#f59e0b,stroke-width:2px
    style F stroke:#6366f1,stroke-width:2px
    style G stroke:#10b981,stroke-width:2px

// Filter conditions - see https://qdrant.tech/documentation/concepts/filtering/#filtering-conditions
var filter = new Filter
{
    Must =  // AND conditions
    {
        new Condition { Field = new FieldCondition
        {
            Key = "language",
            Match = new Match { Keyword = "en" }
        }},
        new Condition { Field = new FieldCondition
        {
            Key = "published",
            Range = new Range { Gte = 1704067200 }  // 2024-01-01
        }}
    },
    MustNot =  // Exclude conditions
    {
        new Condition { Field = new FieldCondition
        {
            Key = "slug",
            Match = new Match { Keyword = "draft-post" }
        }}
    }
};

Filtertypes (docsunit synonyms for matching user input):

Match.Keyword - Exacte string match
Match.Text - Volledige tekst overeenkomst
Match.Any - Komt overeen met elke array
Range - Numeriek bereik (Gte, Lte, Gt, Lt)
GeoBoundingBox / GeoRadius - Geofiltering

De C#-client

Installeer de ambtenaar Qdrant.Client pakket (GitHub):

dotnet add package Qdrant.Client

Verbindingsinstelling

using Qdrant.Client;
using Qdrant.Client.Grpc;

// gRPC client (recommended) - see https://qdrant.tech/documentation/interfaces/#grpc-interface
var client = new QdrantClient(
    host: "localhost",
    port: 6334,  // gRPC port (6333 is REST)
    https: false
);

// With API key - see https://qdrant.tech/documentation/guides/security/
var secureClient = new QdrantClient(
    host: "your-qdrant.cloud",
    port: 6334,
    https: true,
    apiKey: "your-api-key"
);

Gebruik altijd gRPC (poort 6334) voor productie - 3-5x sneller dan REST.

Windows HTTP/2 Fix

Schakel op Windows ongecodeerde HTTP/2 in voor het aanmaken van de client:

AppContext.SetSwitch("System.Net.Http.SocketsHttpHandler.Http2UnencryptedSupport", true);

Sleutelbewerkingen

Zoeken

// Vector search - see https://qdrant.tech/documentation/concepts/search/
var results = await client.SearchAsync(
    collectionName: "blog_posts",
    vector: queryEmbedding,
    limit: 10,
    filter: filter,
    scoreThreshold: 0.5f,  // Minimum similarity
    searchParams: new SearchParams
    {
        HnswEf = 128,  // Search accuracy (higher = better recall)
        Exact = false  // Use approximate search
    },
    withPayload: true
);

foreach (var result in results)
{
    Console.WriteLine($"{result.Payload["title"].StringValue}: {result.Score}");
}

Batch upsert

// Batch operations - see https://qdrant.tech/documentation/concepts/points/#batch-update
var points = documents.Select(doc => new PointStruct
{
    Id = new PointId { Uuid = doc.Id },
    Vectors = doc.Embedding,
    Payload = { ["slug"] = doc.Slug, ["title"] = doc.Title }
}).ToList();

await client.UpsertAsync(
    collectionName: "blog_posts",
    points: points,
    wait: true  // Wait for indexing
);

Verwijderen

// Delete by filter - see https://qdrant.tech/documentation/concepts/points/#delete-points
await client.DeleteAsync(
    collectionName: "blog_posts",
    filter: new Filter
    {
        Must = { new Condition { Field = new FieldCondition
        {
            Key = "slug",
            Match = new Match { Keyword = "old-post" }
        }}}
    }
);

HNSW Index Tuning

HNSW (Hierarchical Navigable Small World) is Qdrant's index algoritme.

flowchart TB
    subgraph "HNSW Graph Layers"
        L2[Layer 2 - Sparse]
        L1[Layer 1 - Medium]
        L0[Layer 0 - Dense]
    end

    Q[Query] --> L2
    L2 --> L1
    L1 --> L0
    L0 --> R[Nearest Neighbors]

    style L2 stroke:#8b5cf6,stroke-width:2px
    style L1 stroke:#6366f1,stroke-width:2px
    style L0 stroke:#3b82f6,stroke-width:2px
    style Q stroke:#10b981,stroke-width:2px
    style R stroke:#ef4444,stroke-width:2px

Indexparameters

// HNSW config - see https://qdrant.tech/documentation/concepts/indexing/#hnsw-index
var hnswConfig = new HnswConfigDiff
{
    M = 16,              // Edges per node (16-32 recommended)
    EfConstruct = 100,   // Build-time accuracy (100-200)
    FullScanThreshold = 10000  // Brute force threshold
};

await client.UpdateCollectionAsync(
    collectionName: "blog_posts",
    hnswConfig: hnswConfig
);

Zoektijdnauwkeurigheid:

var searchParams = new SearchParams
{
    HnswEf = 128  // Higher = better recall, slower (64-256)
};

Richtsnoeren voor het afstellen: Use Case M EfConstruct HnswEf |----------|---|-------------|--------| Fast, low recall 8 64 . . . 32 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Uitgebalanceerd 16 100 128 High recall . 32 . 200 . 256 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Payload-indexen

Aanmaken laadvermogensindexen voor vaak gefilterde velden:

// Keyword index - see https://qdrant.tech/documentation/concepts/indexing/#payload-index
await client.CreatePayloadIndexAsync(
    collectionName: "blog_posts",
    fieldName: "language",
    schemaType: PayloadSchemaType.Keyword
);

// Integer index for ranges
await client.CreatePayloadIndexAsync(
    collectionName: "blog_posts",
    fieldName: "published",
    schemaType: PayloadSchemaType.Integer
);

Impact: 10-100x sneller filteren op grote collecties.

Kwantisering

Kwantisering vermindert het geheugengebruik:

// Scalar quantization - see https://qdrant.tech/documentation/guides/quantization/#scalar-quantization
await client.UpdateCollectionAsync(
    collectionName: "blog_posts",
    quantizationConfig: new ScalarQuantization
    {
        Scalar = new ScalarQuantizationConfig
        {
            Type = ScalarType.Int8,  // float32 -> int8
            Quantile = 0.99f,
            AlwaysRam = true
        }
    }
);

Afhandelen: 4x minder geheugen, ~2% geheugenverlies, 1,5x sneller zoeken.

Docker Implementatie

# docker-compose.yml - see https://qdrant.tech/documentation/guides/installation/
services:
  qdrant:
    image: qdrant/qdrant:v1.12.1  # Pin version!
    ports:
      - "6333:6333"  # REST
      - "6334:6334"  # gRPC
    volumes:
      - qdrant_data:/qdrant/storage
    environment:
      - QDRANT__SERVICE__GRPC_PORT=6334
      - QDRANT__SERVICE__HTTP_PORT=6333
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:6333/health"]
      interval: 30s
      timeout: 10s
      retries: 3

volumes:
  qdrant_data:

Veiligheid

Inschakelen API-sleutelauthenticatie:

environment:
  - QDRANT__SERVICE__API_KEY=your-secret-key

Toezicht

Metrics

Qdrant onthult Prometheus metrics op /metrics:

curl http://localhost:6333/metrics

Sleutelgegevens:

qdrant_collections_vector_count - Totaal vectoren
qdrant_rest_responses_duration_seconds - Query latency
qdrant_memory_usage_bytes - Geheugenverbruik

Snapshots

Aanmaken back-ups:

# Create snapshot
curl -X POST http://localhost:6333/collections/blog_posts/snapshots

# List snapshots
curl http://localhost:6333/collections/blog_posts/snapshots

# Restore (copy snapshot to storage/collections/blog_posts/snapshots/)

Gewone HebbesCity in New Jersey USA

1. Havenverwarring

6333 = REST API
6334 = gRPC API (gebruik dit!)

2. Vector Dimensie Mismatch

Error: expected dim: 384, got 768

Uw inbedding model en collectie moet overeenkomen met:

all-MiniLM-L6-v2: 384 dimensies
nomic-embed-text: 768 afmetingen
OpenAI text-embedding-3-small: 1536 afmetingen

3. Langzame eerste zoekopdracht

HNSW lui-ladingen in het geheugen. Opwarmen na het opstarten:

await client.SearchAsync("blog_posts", new float[384], limit: 1);

4. Array Filtering

Gebruik Match.Any voor arrayvelden:

new Match { Any = new RepeatedStrings { Strings = { "AI", "ML" } } }

Middelen

Officiële Qdrant-documentatie

Overzicht - Beginnen.
Concepten - Kernbegrippen
Verzamelingen - Het maken en beheren van collecties
Punten - Werken met vectoren
Zoeken - Opvragen
Filteren - Filtercondities
Indexering - HNSW- en ladingsindexen
Kwantisering - Geheugenoptimalisatie
Veiligheid - Authenticatie
Toezicht - Metrics en telemetrie
Snapshots - Reservekopie en herstel

Bibliotheken van klanten

Qdrant .NET Client - Officiële C# SDK
Pakket nuGet - Laatste release

Broncode

Alle code beschikbaar op: github.com/scottgal/meestallucidweb

Mostlylucid.SemanticSearch/Services/QdrantVectorStoreService.cs - Qdrant integratie

Self-Hosted Vector Databanken met Qdrant: Een Diepe Duik