semAND - demo application

This is a demonstration application for project semANT. It indexes large number of text documents and provides:

• Search - combination of vector and full text search with metadata filters
• Search result summarization and question answering
• User document collection with text tagging
• RAG
• ...

Technology used

• Weaviate vector database
• FastAPI
• Vue.js
• Quasar

Contribution guidelines

• Create an issue and a new branch with the same name
• Issue should have these parts:
  • Descriptive title
  • Short summary (2 sentences - one paragraph) inlcuding the purpose of the work.
  • Checklist of what needs to be done (technical)
  • Checklist of how to verify correct implementation - what should someone do to test the functionality
• Work on your branch
• Write notes, questions, observations, and other as issue comments
• Write basic tests
• Update relevant readme or documentation. Create/update diagrams where appropriate.
• Merge main into your branch and resolve conflicts
• Create a pull request and assign a reviewer
• After successfull review, merge/rebase the pull request and delete the branch

DB schema

Diagram of DB: weaviate schema