Skip to content

Releases: Dcup-dev/dcup

v1.1.3-beta

26 May 22:11

Choose a tag to compare

What's Changed

Full Changelog: v1.1.2-beta...v1.1.3-beta

v1.1.2-beta

18 May 14:54
13f4383

Choose a tag to compare

The goal from this version is fixing bugs and tests
i set up e2e tests and make sure all works fine and find many bugs that need to deal with

What's Changed

New Contributors

Full Changelog: v1.1.0-beta...v1.1.2-beta

v1.1.0-beta

29 Apr 11:32

Choose a tag to compare

New Features & Integrations

1. AWS S3 Integration: Users can now connect an S3 bucket, browse objects, and ingest files with functionality mirroring Google Drive.

  • AWS Authentication: Securely connect using AWS credentials.
  • AWS Folder Picker: Browse and select specific buckets or folders for ingestion.
  • AWS File Processing: Automatically process and index files from S3 buckets.

2. Dropbox Integration: Users can link their Dropbox accounts to Dcup, enjoying the same capabilities as with Google Drive.

  • Dropbox Authentication: Securely connect your Dropbox account.
  • Dropbox Folder Picker: Choose specific folders to ingest.
  • Dropbox File Processing: Automatically process and index Dropbox files.

3. Direct File Upload Option: A native file upload feature now allows users to upload files directly via the Dcup UI.

  • Native Upload Feature: Upload files straight from your device.
  • Connection Set for Direct Upload: Organize and manage uploaded files within a connection.
  • Add & Delete Files: Easily add new files or remove existing ones from the connection.
  • Dashboard Display: Uploaded files are visible and manageable in the Dcup dashboard.

4. Query Page: A new page dedicated to querying data, complete with search and filtering capabilities.

  • Search Input: Enter queries effortlessly.
  • Filtering Options: Refine results using API-supported filters.
  • Results Window: View organized, easy-to-read search results.

5. Comprehensive Ingestion API Endpoints: A full suite of API endpoints to give users complete control over data ingestion.

  • GET Endpoints:

    • Retrieve all connections.
    • Fetch a specific connection by ID.
    • Get all files associated with a connection ID.
    • Access specific file chunk data by chunk_id.
  • POST Endpoints:

    • Create a new direct upload connection.
    • Sync an existing connection by connection_id.
  • PUT Endpoint:

    • Update connection details and add new files to direct upload connections.
  • DELETE Endpoints:

    • Remove specific files from a connection by connection ID and file name.
    • Delete an entire connection by ID.

Improvements

  • Code Refactoring
  • Extensive updates to the codebase for better performance, scalability, and maintainability.
  • Optimized ingestion and retrieval processes to enhance efficiency.

Bug Fixing

  • Addressed multiple bugs to improve stability and ensure a smoother user experience across all features.

Final Thoughts

With Dcup v1.1.0-beta, you’ve added incredible flexibility and power, making it easier for users to manage and process their data from multiple sources. The refactoring and bug fixes further solidify its reliability. Great work—this release is a fantastic milestone! Let me know if you’d like to tweak anything in these notes.

What's Changed

Full Changelog: v1.0.0-beta...v1.1.0-beta

Dcup v1.0.0-beta – The Open-Source RAG Platform is Here!

17 Mar 09:04
24dc637

Choose a tag to compare

I'm thrilled to announce the beta release of Dcup v1.0.0-beta – your new open-source, self-hostable RAG (Retrieval-Augmented Generation) platform built on top of Qdrant! This release marks the first major milestone on my journey to simplify and empower AI-powered retrieval for developers.

What's New in v1.0.0-beta:

Fully Open Source & Self-Hostable:

Enjoy complete control over your data pipelines with an entirely open-source solution.

Connected RAG:

Easily connect your application to popular data sources like Google Drive with our pre-built integrations – more are coming soon!

Future-Ready Features:

Advanced capabilities such as LLM re-ranking, summary indexing, entity extraction, and hybrid search (leveraging OpenAI embeddings and Qdrant vector storage) deliver highly accurate and scalable retrieval.

Intuitive Retrieval API:

Our well-documented POST /api/retrievals endpoint allows you to seamlessly query your indexed data, complete with optional re-ranking for enhanced precision.

Robust Data Processing:

Dcup automatically handles data ingestion, chunking, and indexing to ensure your documents are always ready for efficient retrieval.

Developer-Friendly:

With easy-to-use APIs and clear documentation, you can have your RAG pipeline up and running in minutes. Dcup is designed with modularity and flexibility in mind, so you can customize and scale as your project grows.

Quick Start Guide:

Self-Host with Docker Compose:

  1. Clone the repository.
  2. Update your environment configuration using .env.example.
  3. Create containers with:
docker compose -f docker-compose.prod.yml --env-file .env up

Explore the Retrieval API:

Send POST requests to /api/retrievals to retrieve relevant data chunks. Check out our documentation for full details on request and response formats.

For Developers:

Dcup is modular and open for contributions. Get started by exploring the code, reviewing our contributing guide, and joining the community.