langchain4j+Chroma小试牛刀

最新推荐文章于 2025-05-08 15:31:23 发布

hello_ejb3

最新推荐文章于 2025-05-08 15:31:23 发布

阅读量382

点赞数 2

文章标签：人工智能

本文链接：https://blog.csdn.net/hello_ejb3/article/details/146051365

版权

序

本文主要研究一下如何使用langchain4j对接Chroma向量数据库

步骤

安装Chroma

docker run -d \
  --name chromadb \
  -p 8000:8000 \
  -v "$(pwd)/chroma_data:/chroma/chroma" \
  -e IS_PERSISTENT=TRUE \
  -e ANONYMIZED_TELEMETRY=TRUE \
  docker.1ms.run/chromadb/chroma:latest

pom.xml

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

example

public class JlamaChromaExample {

    public static void main(String[] args) {
        String chromaEndpoint = "http://localhost:8000";
        EmbeddingStore<TextSegment> embeddingStore = ChromaEmbeddingStore
                .builder()
                .baseUrl(chromaEndpoint)
                .collectionName("test1_collection")
                .logRequests(true)
                .logResponses(true)
                .build();

        EmbeddingModel embeddingModel = JlamaEmbeddingModel.builder()
                .modelName("intfloat/e5-small-v2")
                .build();

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

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

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

        System.out.println(embeddingMatch.score()); // 0.8144288493114709
        System.out.println(embeddingMatch.embedded().text()); // I like football.
    }
}

这里使用了Jlama提供的JlamaEmbeddingModel，官方示例的AllMiniLmL6V2EmbeddingModel在mac下会报错ai.djl.engine.EngineException: Unexpected flavor: cpu

输出如下

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)
0.8279024262570531
I like football.