Skip to content

Latest commit

 

History

History
44 lines (32 loc) · 2.23 KB

File metadata and controls

44 lines (32 loc) · 2.23 KB

Notice and Attribution

This repository is a public learning, secondary-development and interview-reference version of dream-rag. It may incorporate code, scaffolding, examples, or module structures derived from upstream open-source projects, engineering templates, or tutorial materials.

Upstream Projects and Dependencies

This project builds on the following open-source frameworks and services:

Project License Usage
Spring Boot Apache 2.0 Application framework
Vue 3 MIT Frontend framework
Naive UI MIT UI component library
Apache Kafka Apache 2.0 Message queue / async processing
Elasticsearch Elastic License 2.0 / SSPL Search engine
Redis BSD-3-Clause Caching / session state
MinIO AGPLv3 Object storage
MySQL GPLv2 (server) Relational database
Apache Tika Apache 2.0 Document parsing
Apache PDFBox Apache 2.0 PDF processing

Attribution Policy

  • Do not remove upstream author comments, copyright headers, license files, or notices.
  • If a previously unlisted upstream project is identified, add its name, repository link, and license here.
  • Template, scaffold, and admin-panel code (e.g., Soybean Admin patterns) should not be presented as original work.

Historical Notes

The project was previously developed under internal naming conventions including "pai-smart" and related variants. Some file naming, configuration keys, and package references may retain traces of these earlier stages. These are being progressively cleaned up and do not affect functionality.

Scope of Original Contribution

The core value demonstrated in this repository lies in:

  • RAG document processing pipeline (upload → parse → chunk → vectorize)
  • Kafka-based async processing with retry and dead-letter queue
  • Elasticsearch hybrid search (BM25 + vector)
  • WebSocket streaming chat with WebFlux SSE consumption
  • Redis-based generation state management and reconnection support
  • Multi-tenant authorization with organization tag filtering
  • Docker Compose orchestration with health checks

Third-party infrastructure, generic admin CRUD, billing/token-recharge modules, and UI scaffolding are supporting components and should be evaluated as such.