带有 Qdrant 的自封矢量数据库: 深海底

Sunday, 23 November 2025

一. 导言导言导言导言导言导言一,导言导言导言导言导言导言

与RAG系列有关: 本文章提供对Qdrant的深度下潜,Qdrant是用于下列用途的矢量数据库:

第4部分:ONNX和Qdrant 执行 - 建立语义搜索
第5部分:混合搜索和自动插入 - 生产一体化

解冻本文包含核心概念、 C# 客户端、性能调控、生产模式等。

什么是 Qdrant ?

A A A 矢量数据库与找到精确匹配点的传统数据库不同, Qdrant 发现 字义相似 项目。

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

密钥 Qdrant 特性 :

HNSW 指数化 - 亚线性搜索时间
过滤过滤 - 将相似搜索与元数据过滤器相结合
gRPC 和REST APPs 税号 - 高绩效准入
分配部署部署 - 水平水平缩放
抓图 - 备份和恢复

核心概念核心概念

实收款

A A A 收藏收藏象一张表格它持有具有固定维度和距离度的矢量。

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
    }
);

远程度量 (医生数):

余弦 - 衡量矢量之间的角度(文字最好)
点点 - 原始内部产品(用于预先正常化的病媒)
欧元 - 几何距离(空间数据)

点点

A A A 点点是包含以下内容的单一记录:

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 });

过滤过滤

过滤过滤运行中之前相似性搜索 -- -- 效率极高。

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" }
        }}
    }
};

过滤器类型 (医生数):

Match.Keyword - 精确的字符串匹配
Match.Text - 全文匹配
Match.Any - 在数组中匹配任意
Range - 数值范围(Gte、Lte、Gt、Lt)
GeoBoundingBox / GeoRadius - 地理过滤

C# 客户端

安装官方 Qdrant. 流利软件包( 软件包)吉特胡布):

dotnet add package Qdrant.Client

连接设置

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"
);

总是使用 gRPC 生产(第6334号港) - 3-5x比废弃能源省快。

Windows HTTP/2 修补

在 Windows 上, 启用未加密 HTTP/2 之前创建客户端 :

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

关键关键业务

搜索搜索

// 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}");
}

批次发件夹

// 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
);

删除删除删除

// 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 指数图

HNSW 香港新南威尔士州 (史无前例的可控小型世界) 是Qdrant的指数算法。

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

索引参数

// 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
);

搜索时间精确度 :

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

计票指南 : 使用 case M EfConstruct HnswEf |----------|---|-------------|--------| 快速,低召回 86432 -=YTET -伊甸园字幕组=- 翻译:

高举回想起 * * * 32 * 200 * 256 * * * 高举回想起 * * 32 * 200 * 256 *

有效载荷指数

创建创建有效有效有效有效有效指数用于经常过滤的字段:

// 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
);

影响: 10-100x 快速过滤大型收藏。

量化

量化减少内存使用量 :

// 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
        }
    }
);

权衡: 减少 4x 内存, ~ 2% 回想损失, 1. 5x 快速搜索。

Doccker 部署

# 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:

安全安全安全安全安全安全安全安全

启用启用 API IPI 键密钥认证:

environment:
  - QDRANT__SERVICE__API_KEY=your-secret-key

监测监测监测

计量数

Qdrant 曝光普罗米修斯指标年月时 /metrics:

curl http://localhost:6333/metrics

关键指标:

qdrant_collections_vector_count - 总矢量
qdrant_rest_responses_duration_seconds - 询问时的延缓
qdrant_memory_usage_bytes - 内存消耗

抓图

创建创建备份:

# 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/)

共犯组织

1. 港口混乱

6333 = STEST API = STEST API = STE = STE API = STE = STE API = STEST API = STE = STEST API = STEST API = STEST API = RET = STEST API = STEST API
6334 GRPC API(使用这个! )

2. 矢量尺寸差

Error: expected dim: 384, got 768

您的嵌入模型和收藏必须匹配 :

all-MiniLM-L6-v2:384维
nomic-embed-text:768个维度
开放国际 text-embedding-3-small: 1536 维

3. 缓慢的第一次查询

HNSW 将懒惰的装入记忆中。启动后暖和起来 :

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

4. 阵列过滤

使用使用 Match.Any 对于数组字段:

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

资源资源资源资源资源资源资源资源资源资源

官方 Qdrant 文档

概览概览概览概览概览概览概览概览 - 开始
概念概念概念 - 核心概念
实收款 - 创建和管理收藏
点点 - 与病媒合作
搜索搜索 - 查询业务
过滤过滤 - 过滤条件
编制索引索引 - 新南威尔士州和有效载荷指数
量化 - 内存优化
安全安全安全安全安全安全安全安全 - 验证 - 验证
监测监测监测 - 计量和遥测
抓图 - 备份和恢复

客户图书馆

Qdrant.net 客户端 - 官方C#SDK
NuGet 软件包 - 最近释放

源代码

所有代码可用于: com/scottgal/ mostlylylucidweb 缩略图/ com/ scottgal/ mostlylylucidweb 缩略图/ 缩略图/ 缩略图/ 缩略图/ 缩略图/ 缩略图/ 缩略图/ 缩略图/ 缩略图/ 缩略图/ 缩略图/ 缩略图/ 缩略图/ 缩略图/ 缩略图/ 缩略图/ 缩略图

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