向量数据库

@Configuration
public class EmbeddingConfig {
    @Bean
    public BatchingStrategy customTokenCountBatchingStrategy() {
        return new TokenCountBatchingStrategy(
            EncodingType.CL100K_BASE,  // 指定编码类型
            8000,                      // 设置最大输入令牌计数
            0.1                        // 设置保留百分比
        );
    }
}

TokenCountEstimator customEstimator = new YourCustomTokenCountEstimator();
TokenCountBatchingStrategy strategy = new TokenCountBatchingStrategy(
		this.customEstimator,
    8000,  // maxInputTokenCount
    0.1,   // reservePercentage
    Document.DEFAULT_CONTENT_FORMATTER,
    MetadataMode.NONE
);

@Configuration
public class EmbeddingConfig {
    @Bean
    public BatchingStrategy customBatchingStrategy() {
        return new CustomBatchingStrategy();
    }
}

FilterExpressionBuilder b = new FilterExpressionBuilder();
Expression expression = this.b.eq("country", "BG").build();

EQUALS: '=='
MINUS : '-'
PLUS: '+'
GT: '>'
GE: '>='
LT: '<'
LE: '<='
NE: '!='

AND: 'AND' | 'and' | '&&';
OR: 'OR' | 'or' | '||';

Expression exp = b.and(b.eq("genre", "drama"), b.gte("year", 2020)).build();

IN: 'IN' | 'in';
NIN: 'NIN' | 'nin';
NOT: 'NOT' | 'not';

Expression exp = b.and(b.in("genre", "drama", "documentary"), b.not(b.lt("year", 2020))).build();

void delete(List<String> idList);

// 创建并添加文档
Document document = new Document("The World is Big",
    Map.of("country", "Netherlands"));
vectorStore.add(List.of(document));

// 按 ID 删除文档
vectorStore.delete(List.of(document.getId()));

void delete(Filter.Expression filterExpression);

// 创建具有不同元数据的测试文档
Document bgDocument = new Document("The World is Big",
    Map.of("country", "Bulgaria"));
Document nlDocument = new Document("The World is Big",
    Map.of("country", "Netherlands"));

// 将文档添加到存储
vectorStore.add(List.of(bgDocument, nlDocument));

// 使用过滤表达式删除来自保加利亚的文档
Filter.Expression filterExpression = new Filter.Expression(
    Filter.ExpressionType.EQ,
    new Filter.Key("country"),
    new Filter.Value("Bulgaria")
);
vectorStore.delete(filterExpression);

// 通过搜索验证删除
SearchRequest request = SearchRequest.builder()
    .query("World")
    .filterExpression("country == 'Bulgaria'")
    .build();
List<Document> results = vectorStore.similaritySearch(request);
// 结果将为空，因为保加利亚文档已删除

void delete(String filterExpression);

// 创建并添加文档
Document bgDocument = new Document("The World is Big",
    Map.of("country", "Bulgaria"));
Document nlDocument = new Document("The World is Big",
    Map.of("country", "Netherlands"));
vectorStore.add(List.of(bgDocument, nlDocument));

// 使用字符串过滤器删除保加利亚文档
vectorStore.delete("country == 'Bulgaria'");

// 验证剩余文档
SearchRequest request = SearchRequest.builder()
    .query("World")
    .topK(5)
    .build();
List<Document> results = vectorStore.similaritySearch(request);
// 结果将只包含荷兰文档

try {
    vectorStore.delete("country == 'Bulgaria'");
}
catch (Exception  e) {
    logger.error("Invalid filter expression", e);
}

// 创建初始文档 (v1) 并带有版本元数据
Document documentV1 = new Document(
    "AI and Machine Learning Best Practices",
    Map.of(
        "docId", "AIML-001",
        "version", "1.0",
        "lastUpdated", "2024-01-01"
    )
);

// 将 v1 添加到向量存储
vectorStore.add(List.of(documentV1));

// 创建相同文档的更新版本 (v2)
Document documentV2 = new Document(
    "AI and Machine Learning Best Practices - Updated",
    Map.of(
        "docId", "AIML-001",
        "version", "2.0",
        "lastUpdated", "2024-02-01"
    )
);

// 首先，使用过滤表达式删除旧版本
Filter.Expression deleteOldVersion = new Filter.Expression(
    Filter.ExpressionType.AND,
    Arrays.asList(
        new Filter.Expression(
            Filter.ExpressionType.EQ,
            new Filter.Key("docId"),
            new Filter.Value("AIML-001")
        ),
        new Filter.Expression(
            Filter.ExpressionType.EQ,
            new Filter.Key("version"),
            new Filter.Value("1.0")
        )
    )
);
vectorStore.delete(deleteOldVersion);

// 添加新版本
vectorStore.add(List.of(documentV2));

// 验证只有 v2 存在
SearchRequest request = SearchRequest.builder()
    .query("AI and Machine Learning")
    .filterExpression("docId == 'AIML-001'")
    .build();
List<Document> results = vectorStore.similaritySearch(request);
// 结果将只包含文档的 v2

// 使用字符串过滤器删除旧版本
vectorStore.delete("docId == 'AIML-001' AND version == '1.0'");

// 添加新版本
vectorStore.add(List.of(documentV2));