DataEase is a self-hosted, open-source business intelligence (BI) platform that enables users to connect to diverse data sources, build datasets, create interactive charts and dashboards, and share insights via links or embedded widgets. The codebase is a full-stack monolith with a Java Spring Boot backend (JPA, Calcite SQL parser, Quartz scheduling) and a Vue 3 frontend (AntV G2/G2Plot/L7 for charts, UniverJS for spreadsheets). It supports a wide range of database connectors (MySQL, PostgreSQL, Oracle, ClickHouse, StarRocks, Elasticsearch, MongoDB, etc.), dataset types (SQL, union, custom fields), chart types (bar, line, pie, map, table, indicator, etc.), and enterprise features like row/column-level permissions, scheduled sync, report delivery, and multi-factor authentication. The architecture includes an SDK layer for extensions (datasource providers, sync providers, chart plugins) and a separate Xpack module for commercial features. The platform is deployed via Docker Compose with APISIX gateway. The codebase is actively maintained (24k+ stars, recent commits) but exhibits some technical drift due to deep inheritance hierarchies in chart handlers, code duplication across engine providers, and tight coupling to specific library versions (AntV G2, UniverJS). Overall, it is a solid, production-grade BI tool with good separation of concerns but with room for refactoring in certain areas.
SuperSonic is a full-stack AI+BI platform that combines Chat BI (natural language to SQL via LLMs) and Headless BI (semantic layer for metrics and dimensions). The backend is a modular Java Spring Boot application with modules for chat, headless query, authentication, and common utilities. It supports multiple databases (MySQL, PostgreSQL, ClickHouse, etc.) and integrates with various LLM providers (OpenAI, Ollama, Dify) and embedding stores (Milvus, Chroma, PGVector, InMemory). The frontend is a React application built with UmiJS and Ant Design, providing a semantic model editor, agent configuration, plugin management, and an embeddable chat SDK. The core NL2SQL pipeline involves schema mapping (rule-based and embedding-based), semantic parsing (rule and LLM), SQL correction (chain of correctors), translation to physical SQL, and execution. The system also includes chat memory for few-shot learning, plugin recognition, and post-query recommendations. While primarily an application, it contains data research elements such as evaluation scripts (benchmark.py, evaluation.py) and custom HanLP dictionaries for domain-specific entity recognition. The architecture is extensible via SPI-based component factories for parsers, correctors, mappers, and DB adaptors. The codebase is well-structured but shows signs of complexity with many modules, some commented-out code, hardcoded configurations, and a dependency on external LLM services that may introduce drift.
The Pentaho Platform is the core backend for the Pentaho Business Analytics server. In production, it serves as a content repository (JCR/Jackrabbit), authentication/authorization system, scheduler (Quartz), and a plugin-based execution engine for BI actions (reports, OLAP, data integration). The system is a monolithic Java EE application deployed on Tomcat, using Spring for dependency injection, Hibernate for persistence, and GWT for the legacy user console. It provides REST and SOAP APIs for repository management, security administration, and job scheduling. The codebase is organized into modules (api, core, extensions, repository, scheduler, user-console) and relies on external libraries such as Apache Jackrabbit, Quartz, Mondrian, Kettle, and Spring Security. Architecturally, the platform is built around a central static service locator (PentahoSystem) and a large set of interfaces for actions, components, and services. It supports multi-tenant repository paths, pluggable authentication (LDAP, JDBC, memory), and a rule-based authorization system. The platform is functionally comprehensive but carries significant technical debt: deprecated interfaces, mixed JUnit versions, outdated front-end technologies (Dojo, GWT), and over-engineered exception hierarchies. Some modules (e.g., authorization decisions, cache expiration, file content decoration) are well-designed and potentially reusable.