PublicDataWorks / verdad-frontend

MIT License
1 stars 0 forks source link

VER-197: [Backend] Evaluate Supabase/pgvector vs. Rememberizer for Related Snippets and Hybrid Search #168

Open nhphong opened 1 day ago

nhphong commented 1 day ago

We want to evaluate the tradeoffs between using Supabase/pgvector and Rememberizer

So that we can determine the most effective solution for implementing full-text search capabilities for related snippets and hybrid search.

Acceptance Criteria:

  1. Research and Analysis:
    • Conduct a thorough analysis of Supabase/pgvector and Rememberizer, focusing on their capabilities for full-text search and hybrid search.
    • Evaluate the integration complexity with our existing system.
    • Assess the scalability and performance of each solution, especially in handling large datasets of disinformation snippets.
    • Analyze the cost implications of implementing and maintaining each solution.
  2. Comparison Metrics:
    • Provide a detailed comparison based on factors such as search accuracy, speed, scalability, ease of use, and cost.
    • Include pros and cons for each solution.
  3. Recommendations:
    • Present a recommendation on which solution to pursue, backed by data and analysis.
    • Suggest next steps for implementation if applicable.

Tasks:

linear[bot] commented 1 day ago

VER-197 [Backend] Evaluate Supabase/pgvector vs. Rememberizer for Related Snippets and Hybrid Search