langchain4j+milvus实战

最新推荐文章于 2025-09-11 09:57:28 发布

原创

最新推荐文章于 2025-09-11 09:57:28 发布 · 1.2k 阅读

5 ·

CC 4.0 BY-SA版权

文章标签：

#milvus

序

本文主要研究一下如何使用langchain4j来对接向量数据库milvus

步骤

docker运行milvus

docker run -d \
        --name milvus-standalone \
        --security-opt seccomp:unconfined \
        -e ETCD_USE_EMBED=true \
        -e ETCD_DATA_DIR=/var/lib/milvus/etcd \
        -e ETCD_CONFIG_PATH=/milvus/configs/embedEtcd.yaml \
        -e COMMON_STORAGETYPE=local \
        -v $(pwd)/volumes/milvus:/var/lib/milvus \
        -v $(pwd)/embedEtcd.yaml:/milvus/configs/embedEtcd.yaml \
        -v $(pwd)/user.yaml:/milvus/configs/user.yaml \
        -p 19530:19530 \
        -p 9091:9091 \
        -p 2379:2379 \
        --health-cmd="curl -f http://localhost:9091/healthz" \
        --health-interval=30s \
        --health-start-period=90s \
        --health-timeout=20s \
        --health-retries=3 \
        docker.1ms.run/milvusdb/milvus:v2.5.5 \
        milvus run standalone  1> /dev/null

启动之后访问http://127.0.0.1:9091/webui

在这里插入图片描述

这里需要提前创建embedEtcd.yaml

listen-client-urls: http://0.0.0.0:2379
advertise-client-urls: http://0.0.0.0:2379
quota-backend-bytes: 4294967296
auto-compaction-mode: revision
auto-compaction-retention: '1000'

user.yaml内容为空即可

pom.xml

<dependency>
    <groupId>dev.langchain4j</groupId>
    <artifactId>langchain4j-milvus</artifactId>
    <version>1.0.0-beta1</version>
</dependency>

example

public class JlamaMilvusExample {

    public static void main(String[] args) throws InterruptedException {
        EmbeddingModel embeddingModel = JlamaEmbeddingModel.builder()
                .modelName("intfloat/e5-small-v2")
                .build();

        MilvusServiceClient customMilvusClient = new MilvusServiceClient(
                ConnectParam.newBuilder()
                        .withHost("localhost")
                        .withPort(19530)
                        .build()
        );
        MilvusEmbeddingStore embeddingStore = MilvusEmbeddingStore.builder()
                .milvusClient(customMilvusClient)
                .collectionName("example_collection")      // Name of the collection
                .dimension(384)                            // Dimension of vectors
                .indexType(IndexType.FLAT)                 // Index type
                .metricType(MetricType.COSINE)             // Metric type
                .consistencyLevel(ConsistencyLevelEnum.EVENTUALLY)  // Consistency level
                .autoFlushOnInsert(true)                   // Auto flush after insert
                .idFieldName("id")                         // ID field name
                .textFieldName("text")                     // Text field name
                .metadataFieldName("metadata")             // Metadata field name
                .vectorFieldName("vector")                 // Vector field name
                .build();                                  // Build the MilvusEmbeddingStore instance

        TextSegment segment1 = TextSegment.from("I like football.");
        Embedding embedding1 = embeddingModel.embed(segment1).content();
        embeddingStore.add(embedding1, segment1);

        TimeUnit.SECONDS.sleep(60);

        TextSegment segment2 = TextSegment.from("The weather is good today.");
        Embedding embedding2 = embeddingModel.embed(segment2).content();
        embeddingStore.add(embedding2, segment2);

        TimeUnit.SECONDS.sleep(60);

        String userQuery = "What is your favourite sport?";
        Embedding queryEmbedding = embeddingModel.embed(userQuery).content();
        int maxResults = 1;
        List<EmbeddingMatch<TextSegment>> relevant = embeddingStore.findRelevant(queryEmbedding, maxResults);
        EmbeddingMatch<TextSegment> embeddingMatch = relevant.get(0);

        System.out.println("Question: " + userQuery); // What is your favourite sport?
        System.out.println("Response: " + embeddingMatch.embedded().text()); // I like football.
    }
}

最后输出

WARNING: Using incubator modules: jdk.incubator.vector
INFO  c.g.tjake.jlama.model.AbstractModel - Model type = F32, Working memory type = F32, Quantized memory type = F32
WARN  c.g.t.j.t.o.TensorOperationsProvider - Native operations not available. Consider adding 'com.github.tjake:jlama-native' to the classpath
INFO  c.g.t.j.t.o.TensorOperationsProvider - Using Panama Vector Operations (OffHeap)
Question: What is your favourite sport?
Response: I like football.

quotaAndLimits

quotaAndLimits:
  enabled: true # `true` to enable quota and limits, `false` to disable.
  # quotaCenterCollectInterval is the time interval that quotaCenter
  # collects metrics from Proxies, Query cluster and Data cluste