4 个仓库
Custom logic implemented via external scripts to extend the database query engine.
Distinct from Custom Module Implementations: Focuses on extending query-time graph analysis logic specifically, whereas Custom Module Implementations is a generic software architecture pattern.
Explore 4 awesome GitHub repositories matching software engineering & architecture · Query Extensions. Refine with filters or upvote what's useful.
RouteLLM 是一个路由框架和流量管理器,旨在将提示词(prompts)在高性能模型和低成本大语言模型之间进行智能分发。它作为一个模拟 OpenAI 规范的 API 网关,能够跨不同的模型提供商路由请求。 该系统通过基于预测胜率和提示词复杂度将流量分配到不同模型层级,从而优化运营成本。它包含一个校准工具,用于分析样本查询并确定流量分配的最佳成本-质量平衡点。 该框架提供了一个使用基准数据集衡量路由策略准确性和成本效率的工具,并通过基于接口的插件和自定义路由逻辑支持扩展性。
Provides an extensible interface for developers to implement custom routing algorithms and win-rate predictors.
Nominatim 是一个自托管的地理空间搜索引擎和地理编码服务器,利用 OpenStreetMap 数据。它为正向地理编码(将地址或地名转换为地理坐标)和反向地理编码(将坐标转换为人类可读的物理地址)提供了完整的基础设施。 该项目具有一个专门的数据导入器,可将原始地图数据解析为 PostgreSQL 地理空间数据库。它的独特之处在于一个可配置的导入管道,该管道使用样式文件过滤地图特征,并使用基于重要性的排名系统根据流行度和特征类型对搜索结果进行优先级排序。 该系统涵盖了广泛的能力,包括空间索引、分层地址计算以及用于模糊查找的基于标记的搜索索引。它支持集成用于门牌号和邮政编码中心点的外部数据集,以及用于使本地数据集与全球源保持同步的增量数据复制。 该服务作为基于 ASGI 的 Web 应用部署,提供支持多种输出格式(包括 JSON、GeoJSON 和 XML)的地理空间搜索 API。
Uses external modules to transform and categorize raw search queries before they enter the tokenization pipeline.
Memgraph is an in-memory, distributed graph database designed for high-performance labeled property graph management. It utilizes a Cypher query engine for declarative data retrieval and manipulation, providing a scalable knowledge graph backend that integrates vector search and graph traversals. The system distinguishes itself as a real-time graph analytics platform, employing native C++ and CUDA implementations to execute complex network analysis and dynamic community detection on streaming data. It provides specialized support for AI integration, including GraphRAG capabilities, the constr
Allows developers to implement specialized graph analysis logic by extending the query engine with external scripts.
pgdog is a PostgreSQL sharding proxy, distributed SQL router, and connection pooler. It is designed to enable horizontal data distribution by splitting tables and indices across multiple independent servers to scale storage and processing capacity. The project distinguishes itself through online resharding capabilities, using logical replication to move data between shards without application downtime. It supports multiple routing strategies, including hash, list, and range-based query routing, and manages distributed atomic transactions using a two-phase commit process to ensure consistency
Provides a plugin architecture for extending SQL routing logic via shared libraries.